You're right, I did. Thanks for pointing that out. I must have looked over them when I removed the material at the top, not knowing that some were interwiki links.—Zhaladshar (Talk) 01:16, 29 September 2005 (UTC)


Hi, ThomasV, you are now a sysop.—Zhaladshar (Talk) 17:41, 30 November 2005 (UTC)

thanks. 17:45, 30 November 2005 (UTC)

UN Security Council ResolutionsEdit

Welcome to adminship. As I check Wikisource:Possible copyright violations but I cannot speak French well, I would like to invite you there to see a very hot discussion about whether UN Security Council Resolutions are admissible at Wikisource. I have to stop adding UN resolutions until they are admissible here. This also applies at French and Chinese Wikisource.--Jusjih 08:25, 3 December 2005 (UTC)

yeah, I heard about it. I do not think that you have to stop adding them. what I understood is that some people considered that the legal disclaimer of wikisource should be adapted to include law texts. I really do not know more about the topic. cheers. ThomasV 11:19, 3 December 2005 (UTC)
Not all law texts are in the public domain, so whenever I add Taiwanese laws, I always use copyright tags to specifying that govenmental translations are in the public domain. I have stopped adding UN Resolutions due to unclear discussions as of yet. In the meantime, I add Taiwanese and Macanese laws in the public domain.--Jusjih 04:20, 8 December 2005 (UTC)
I have noticed French Wikisourse with a tab "license" for UN resolutions. As you hate copyright tags right on articles, that is a good idea to tell readers how articles there may be used. I don't know if you have time to proofread English or French texts. As I am an administrator here and at Chinese Wikisource but not at French Wikisource, I can eventually protect English and Chinese UN texts, but for French ones, either you or another administrator protect them.--Jusjih 01:03, 29 December 2005 (UTC)
I will write an explanation notice for copyright tabs. I noticed you wrote "J'ai corrigé cet article" in several discussion pages. I suppose you mean "proofread". The french word for that is "vérifié". "corrigé" is misleading, it means that you found a typo and fixed it, but it does not imply that the text is now error free. ThomasV 06:52, 29 December 2005 (UTC)
I intend to mean "proofread". You mean that I should better say "J'ai vérifié cet article" after proofreading an article, right? Shall we fix typos? Please excuse my basic and poor French, but you may send me messages in French for my reading practice. My French writing and speaking are much poorer, but I can read simple French.
Merci beaucoup pour votre conseil.--Jusjih 10:18, 31 December 2005 (UTC)

I just added a message at Wikisource:Scriptorium#Articles.2C_project_pages.2C_and_United_Nations_resolutions about whether the index page of the United Nations Security Council Resolutions should be an "article" not prefixing "Wikisource:" or a "project page" prefixing "Wikisource:".

I added a French project page at multilingual Wikisource on 23 July 2005, then you moved it to an article namespace on 25 August 2005. After moved to French Wikisource, I moved to back to a project page becuase I would like to make the namespace the same in English, French, and Chinese. However, as I am not very sure whether I should have done it and I plan to add trilingual indexes for the UN General Assembly Resolutions, in addition to scriptorium, I would like to ask you whether fr:Wikisource:Résolutions du Conseil de sécurité de l'ONU that I have gathered from the UN web site with my changes of format for Wikisource and the order of the resolutions should indeed be an "article" or a "project page".--Jusjih 10:50, 18 January 2006 (UTC)

I moved it in the "main" namespace because I thought it was like the table of contents of a book. but if you have good reasons to put it in the wikisource namespace, go ahead. ThomasV 10:54, 18 January 2006 (UTC)
Based on your and others' comments, I have made a reverse page move there while I have no more good reason to put it in the Wikisource namespace. I will move the Chinese index and prepare an index of the UN General Assembly Resolutions as well. Merci beaucoup pour votre conseil.--Jusjih 23:03, 18 January 2006 (UTC)

United Nations resolutions are found to be in the public domainEdit

===United Nations works after 1 March 1989 no longer acceptable at Wikisource for being GFDL-incompactible===

Due to Wikimedia Foundation's requirement that results in the copyright policy prohibiting non-commercial licenses, UN works published after 1 March 1989 are copyrighted for 95 years since publication by default and Template:UNCopyright with non-commercial license is GFDL-incompactible here, so unless available at the US Government Websites (suffixed .gov) without additional copyright restrictions, they have to be deleted from English Wikisource.

Has French Wikisource enforced the ban of non-commercial licenses yet? Chinese Wikisource has yet to enforce it while I have mentioned this matter. Here, I am deleting so many UN resolutions published after 1 March 1989 that are normally copyrighted by default, but Chinese and French Wikisource do not have such a big mess.--Jusjih 13:39, 16 May 2006 (UTC)

Please see the revised Template:PD-UN after WS:COPYVIO and WS:S have found United Nations resolutions to be in the public domain.--Jusjih 15:22, 12 November 2006 (UTC)

Thomas DretartEdit

  • Franch alias : Thomas Dretart : author page of Thomas Dretart, apparently made by the author himself. His works are protected by copyright. I tried to tell him that he has to release his texts under a free licence a few months ago, but he did not seem to understand. if somebody wants to talk to him, his username is Tomahawk something on ThomasV 12:50, 2 December 2005 (UTC)
Thomas Dretart: To make my biography by my self is my liberty. Administrator ThomasV seid:"His works are protected by copyright".Non, authors works of Thomas Dretart are copyright by him self. I prouved.Thomas Dretart is Tomislav Dretar. My works are no protected by copyright for my self. Es gibt keine Wahrheit in dem Konstatation von ThomasV alis ThomasBhot. ThomsBhot a menti tout le monde.

J'ai lui expliqué tout, mais il ne veux pas comprendre. Il a ses propres raisons pour pourchasser Tomislav Dretar de partout où il se trouve. I have deleted my biography im october bat ThomasV remaid the problem in december.!?!? --Tomislav Dretar 06:43, 6 December 2005 (UTC) --Tomislav Dretar 06:43, 6 December 2005 (UTC)

Thomas, vous êtes vraiment un miracle en monde. J'ai tout effacé, tout arrêté, mais vous continuez. Pourqoi bon Dieu? Laissez moi tranquille au Nom de Dieu! Droits d'auteurs sont à moi. J'ai bien prouvé, mais Vous n'écoutez pas! Trouvez, bon Dieu, un autres jouet! Ce que j'ai fait un peu de publicité pour moi même ce n'est pas interdit. Et tout est la vérité et rien que la vérité. En plus je suis dans la bibliographie française. Regardez un peu. Vous pourchassez un écrivain français. Ce sont mes origines croates qui vous font vous affoler, ha? --Tomislav Dretar 06:55, 6 December 2005 (UTC)

you are indeed free to write your own biography. but the Wikisource project is free to accept it or not on its pages. biographies of famous people are appropriate on wikipedia, not on wikisource. why don't you write your biography there?
now concerning your works, they ARE copyrighted. whether you want it or not. That's the law. Wikisource does not accept copyrighted texts. Wikisource is not a free advertising site for young authors.
ThomasV 08:08, 6 December 2005 (UTC)
Hi Thomas, could you tell me what's going on with Thomas Dretart? Are his works legit, or should we delete them?
I talked to him on the phone. Apparently he understood that he has to release his works under a free licence. he used the wiki in order to publish his own texts, and since then he has found an editor. his editor is willing to publish his works, but who does not want his texts to be published here. so you may delete his works, they are copyvio. ThomasV 17:20, 20 December 2005 (UTC)


It's odd that no one's reviewing the extension (aren't there a lot of developers?). If Brion's busy, someone else should be available to review it. Oh well, I'm sure at some point it will be reviewed--we just have to be patient (remember, the Special:Import feature still has yet to be completed).—Zhaladshar (Talk) 17:12, 20 December 2005 (UTC)

Hello Thomas. I already thought about the need of such an extension and think it is very helpful. I don't know much about coding, so I can't say much about the extension. But as far as I can see, the "compare" links are always added. I think there should be an extra wiki link ("[[comp:fr:..]]) because the texts have to be formatted the same and this mostly won't be the case (eg author pages). By the way: Is there an international discussion page for topics concerning all wikisource projects (like this)? If not, I think it should be created (for example at int.wikisource as subpage of Scriptorium) It would be a pity if the projects would completely separate and there are many (software) problems they have in common. --Jofi 00:17, 21 December 2005 (UTC)

I think the scriptorium at should be the place for international discussions. I first posted my message about the extension there. however, I re-posted it to a few other scriptoriums today, because I had the feeling that not a lot of people had read the first message. if you like the extension please vote for it. ThomasV 00:30, 21 December 2005 (UTC)
Ok. I was not active for a longer period of time, so I missed some things. --Jofi 00:37, 21 December 2005 (UTC)
re compare links: they are always added, but only in the main namespace. ThomasV 00:46, 21 December 2005 (UTC)
So they would be added to author pages, too? --Jofi 00:58, 21 December 2005 (UTC)
yes, unless a separate namespace is created for author pages. but I do not think that it is a bad thing to be able to compare author pages in different languages, even if they do not have the same format. it is not necessary for the pages to have the same format. here is an example where two pages are compared even though they do not have the same format. the extension still works well. ThomasV 01:11, 21 December 2005 (UTC)
You are right. I thought it would mess things up. A last question: What about pages in the same subdomain? For example de.wikisource has 2 versions of the Lutherbible, or texts in original old language and translated new versions. It would be nice, if they could also be displayed there. --Jofi
I agree. you are not the first one to ask this. it is not implemented yet, but I think it would be very easy to do. ThomasV 01:44, 21 December 2005 (UTC)


You added the dotabs() function to Mediawiki:monobook.js, and it's generating an error in Firefox 1.5.

Error: a has no properties
Source file:
Line: 123

and it points here:

function dotabs()
   //first tab
'''   var a = document.getElementById("ca-nstab-main");'''
   var s = a.innerHTML; 

Omegatron 21:29, 12 February 2006 (UTC)


This is meant to be an explanation of my note on Scriptorium and not a refutation of your statement.

According to Leslie Lamport (who developed LaTeX)

The LateX document preparation system is a special version of Donald Knuth's TeX program. TeX is a sophisticated program to produce high-quality typesetting, especially for mathematical text. LaTeX adds to TeX a collection of commands that simplify typesetting by letting the user concentrate on structure of the text rather than formatting commands.

I used to use LaTex to write technical papers. It can be thought of as a markup language and is very versatile. I mentioned TeX because I have read that there is talk about using a modified form of TeX to replace HTML. In my own experience, in making contributions to Wikisource, both Wiki markup and HTML are very frustrating because it is so difficult to format text in a way we take for granted in typeset books. PDF is great and has a huge installed base but IMO it might not be the format for the future. Just some thoughts. Like I said I'm not an expert. I'll stop rambling now. --Droll 21:59, 7 March 2006 (UTC)

Hello Thomas V I need your advice on Recent Changes could you help cleaner we need some 15:38, 15 March 2006 (UTC)

please stop posting junk ThomasV 15:42, 15 March 2006 (UTC)

Why you call it Junk? 15:43, 15 March 2006 (UTC) You want Poetry fine... its not just me??? but thanks...

Thomas as I said at noticeboard here is the site... a newbie in training... 15:58, 15 March 2006 (UTC) Firefly see don't say because the title i not sure of... bare with all. 15:58, 15 March 2006 (UTC)


side by side comparisonEdit

Is there away to set-up the side by side comparison for an article in the same language domain. An example would be Poems (Donne)/Elegie I and Poems of John Donne (1896)/Elegy I. Thanks for looking at this--BirgitteSB 03:09, 26 March 2006 (UTC)

you are not the first one to ask about this. it would require a modification of the software. unfortunately, I have no time for it atm. If you know php, you can write a patch that does that. but this is the easy part; once you have written your patch, you 'll need to have developers accept your patch, and that is really difficult to go though :-). I took me months of nagging them on irc, almost every day. ThomasV 08:14, 26 March 2006 (UTC)

Lorem ipsumEdit

I restored this page as it in not latin. It actually gibberish meant to resemble latin that is used by publishers as a placeholder to test out layout before they enter actual information.--BirgitteSB 05:58, 6 April 2006 (UTC)

it is not English, is it? ThomasV 11:57, 6 April 2006 (UTC)
It is now nominated for deletion. I always learned it was gibberish. But Pathoschild says it is based on a latin text. I really don't mind it being deleted, but I thought it was a misunderstanding. I was just thinking it was very common text used by English speakers and it was late if that is any excuse :) --BirgitteSB 12:03, 6 April 2006 (UTC)


Hi. There has been a request to create a .hu subdomain for wikisource. here I don't know if you are the right person to ask to do it, but I would like to ask you to help if you can. --Dubaduba 10:01, 10 April 2006 (UTC)

ThomasBor for el:Edit

Is it possible use the bot in the Greek Wikisource? For example to split articles like following into smaller articles? el:Κατά Ιωάννην

-- w:en:User:Ανδρέας

Author talk pages included in Category:Author-PD-oldEdit

Hi, is there a reason for some author talk pages such as Author talk:Charles Dickens having copyright tags and being included in the relevant copyright category? If not, I'd like to delete the tags thus removing the author talk pages from the categories.--GrafZahl 12:50, 12 June 2006 (UTC)

Since you haven't answered for one week, I'm simply doing it now ;)--GrafZahl 12:24, 20 June 2006 (UTC)
sure go ahead . ThomasV 10:13, 22 June 2006 (UTC)

Labeled section transclusionEdit

Thomas, thanks so much for sharing your solution. Please help improve the paragraph describing it here.

What became clear yesterday was that three separate Wikisource languages (French, English, and Hebrew) all found themselves facing the same problem and the same need, and all of them have tried to implement various technical solutions to the problem this year. To me, this fact points more than ever to the need for a solution that is both flexible, powerful and convenient to use.

I posted a note that effect at the Wikitech mailin list.

Hope you will stay involved and give more feedback! Dovi 07:39, 10 September 2006 (UTC)

ProofreadPages and page qualityEdit

Thanks for ProofreadPages. If this works out, the whole proofreading process should become much simpler.

Now, here's my first problem: I've created Template:pagequality as a copy of fr:Modèle:PageQuality and added a call to Page:A Treatise on Electricity and Magnetism Volume 1 005.jpg, but on A_Treatise_on_Electricity_and_Magnetism/Preface, where it is transcluded, the page quality indicator is still at 0%. Maybe some JavaScript or CSS has not been copied over from fr: to en: yet?

Thanks for your help.--GrafZahl 13:20, 14 September 2006 (UTC)

yes, some javascript is missing. I did not include this part because I was not sure if the English wikisource was still using these quality indicators. I'll take care of that. ThomasV 13:41, 14 September 2006 (UTC)
Thanks!--GrafZahl 07:54, 15 September 2006 (UTC)

side by side image and text implementationEdit

Hi Thomas, I am a sysop at te wiksource. I am trying to implement side by side image and text thing like you have here in Page: namespace. I am assuming I copied all the necessary code but I could not get it to work. What are the steps I need to go through to make this work in Telugu wikisource. Your help is appreciated very much as we are going to start a big project in te wikisource --Vyzasatya 19:02, 12 October 2006 (UTC)

you need to ask a developer to enable the ProofreadPage extension. ThomasV 09:18, 20 October 2006 (UTC)

Page NamespaceEdit

Hello Thomas,

thanks a lot for your advice and your help! That was exactly the feature I was looking for, when I asked in the Scriptorium for the correct way to do page references. Best regards, --Linda.floren 07:58, 15 April 2007 (UTC)

2 categoriesEdit


Il y a 2 catégories concurrentes :

Yann 18:36, 5 May 2007 (UTC)

non, la première contient SRW qui n'utilise pas le mode Page. et puis la 2e regroupe des pages d'index, ce qui n'est pas la même chose que de placer les pages dans des catégories. ThomasV 18:44, 5 May 2007 (UTC)


Hi, thanks for the ProofreadPage overhaul, bugzilla:7957 in particular. I'm somewhat overloaded with work right now, so it'll take some time until I clean up my ProofreadPage projects.--GrafZahl (talk) 15:36, 7 May 2007 (UTC)

Changes in Proofread pageEdit

Have you recently made changes to this? I hadn't work with it for a few days and now the image isn't enlarged like it used to be. Any ideas?--BirgitteSB 18:25, 18 May 2007 (UTC)

I do not know what you mean.. can you show me an example? ThomasV 15:16, 20 May 2007 (UTC)
The image used to appear larger at that link. [1]--BirgitteSB 13:33, 21 May 2007 (UTC)
I do not know. what navigator are you using ? ThomasV 19:10, 22 May 2007 (UTC)
Internet Explorer. The strange thing is I have used it before with the setup and had a larger readable version of the image. Also the edit box used to narrower to allow the image more room.--BirgitteSB 13:34, 23 May 2007 (UTC)
I think the way it used to be rendered in IE (with this large image) was buggy, and is now fixed. the problem with IE is that the zoom does not work, so you cannt see the image precisely. although I agree that this is not a valid answer, I can only suggest to use firefox. ThomasV 09:44, 24 May 2007 (UTC)
Thanks. At least it is not me going crazy.--BirgitteSB 14:07, 24 May 2007 (UTC)

Proof Read page extension on te wikisourceEdit

Hi ThomasV,

We have created Page namespace and also alias namespace with localname in Telugu wikisource and had the extension activated. For some reason it is not working. can you please look into it. Here is the link to request I made in bugzilla

Thanks --Rvyz 08:55, 3 June 2007 (UTC)

in fact, creating the namespaces is not sufficient; you actually need to set two system messages : Proofreadpage_namespace (its default value is Page) and Proofreadpage_index_namespace (its default value is Index) ThomasV 21:31, 3 June 2007 (UTC)
Thanks I'll have it set by programmers --Rvyz 04:43, 9 June 2007 (UTC)


I am a newbie on WikiSource, ive had some experience on English Wikipedia (as User:Twenty Years - 8 months) and Commons (as User:Five Years - 1 month). I was wondering if i would be able to upload a list of Youth Sports Champions for a single association, called the Public Schools Association, would that be appropriate for wiki source. Thanks alot. Five Years 11:28, 17 June 2007 (UTC)

1001 NightsEdit

Thanks for the comment. I was wondering who among the old-timers would be first to notice.:-)

I don't know if page scans are available for The Nights. There is a site which already has the entire Burton nights on line in a text version ( and this seems to have the basis for the megabyte volume 1 text that was already provided to us. (Both show a consistent OCR error of putting "shine" for "thine") I have no idea what edition was used for the upload. I have been proofreading the text as I go along by comparison with my own Burton Club edition of the text. I have also been restoring Greek text in the footnotes where the scanned source only gives "Greek letters". If I get really ambitious, I may upload the illustrations in the edition. This includes the Letchford illustrations, and some by unidentified illustrators.

One interesting question: Where a reference to the Qur'an is made in the notes I am linking it the indicated chapter of the Qur'an, but with such things there remains the question of what the appropriate version is for such links, Eclecticology 17:18, 5 July 2007 (UTC)

Proofread Page ExtensionEdit

Hi - I'm trying to install the Proofread Page extension on my locally hosted version of MediaWiki. Are there any tricks to installing it? I have done the usual for extension installation (placed it in the extension folder and made a require_once line in LocalSettings). I also created MediaWiki:Proofreadpage index namespace and MediaWiki:Proofreadpage namespace as well as Page and Index namespaces. However, every time I uncomment the require_once line, my wiki produces blank pages, indicating some type of php problem. Are there any more steps for installation? Thanks! -SCEhardT 19:59, 9 July 2007 (UTC)

check the php log to know about the error.
for images that are larger than the half window, the extension uses thumbs; this requires that you install a 404 handler, such as WebStore.php. but I guess your problem does not come from there, because apparently all pages are blank; so, check the log
ThomasV 21:05, 9 July 2007 (UTC)
Thanks - turned out to be a problem with the name of the extension directory (must be ProofreadPage) -SCEhardT 20:58, 10 July 2007 (UTC)

A french suggestionEdit

Copied from my talk page: C'est plutôt l'inverse. C'est Wikisource qui devrait utiliser WebCodex pour saisir les versets en bases de données. Cordialement MLL 21:14, 9 July 2007 (UTC)

I think this is implying we could automatically take verse from a webcodex database (rather than copying and pasting). I shared this information with User:Sanbeg, but he cannot see how. Since you speak both french and computer(!), I was wondering if you could take a look. (More information at User talk:Sanbeg#Bible and Talk:Bible/Obadiah/1/1).--BirgitteSB 14:06, 12 July 2007 (UTC)
sorry but I do not know anything about webcodex. ThomasV 09:26, 13 July 2007 (UTC)

Bug in ProofreadPageEdit

Hi ThomasV, it seems the following two lines of code:

diffButton = document.getElementById("wpDiff")

need to be added to oldwikisource:MediaWiki:Page.js, function pageURL(), after

saveButton = document.getElementById("wpSave"); 
previewButton = document.getElementById("wpPreview"); 

Thank you!--GrafZahl (talk) 22:15, 24 July 2007 (UTC)

thanks for pointing that. I'll try to fix it today ThomasV 06:18, 26 July 2007 (UTC)
it is fixed now, and I made many improvements. we just have to wait until servers are updated... ThomasV 12:46, 26 July 2007 (UTC)

Proofread Page and Unused mediaEdit

Hi, while playing around with Special:Unusedimages, I noticed that it includes media that is "used" on a "Page:". e.g. Page:SomeRemarksOnCorset column1.png is the first "unused" png listed on the second page of Special:Unusedimages; and Page:Literary_Digest_1928-01-07_Henry_Ford_Interview_3.jpg is a file on commons that checkusage says is unused. Are there any plans to resolve this? John Vandenberg 05:53, 4 September 2007 (UTC)

Indeed it would be great to have the pages listed as "used". I do not have time to write a patch now, but feel free. ThomasV 15:16, 4 September 2007 (UTC)

More PageQuality problemsEdit

Hi, ThomasV. Perhaps it is related with Wikisource:Scriptorium#site-independent_JavaScript.2Farticle_tab_breakage (or not), or with the disappeared footer and header of the Page's. "My" problem now is that at Catalan WS, the pages with the PageQuality template, like ca:Pàgina:Tenditinardua.jpg shows the little image of 75%, but in the secure servers don't: . Please, could you have a look at :ca:dontnowwhere (don't hurry, just if you have time; i'm not very keen on mediawiki...). Thanks ;) -Aleator 19:26, 4 September 2007 (UTC)

yes, it is the same bug as reported on en:ws. the working code is here : . You will need to replace "Page" with "Pagina" I guess.
It has nothing to do with the header and footer fields. if you want to see them click on the "+" button.
ThomasV 21:18, 4 September 2007 (UTC)
More things:
  • What is the "+" button?
I found the button! Javascript had to be activated. -Aleator 18:05, 8 September 2007 (UTC)

indent levelEdit

Hi ThomasV, I think the indent level of your comment needs to be adjusted. It is a reply to Eclecticology? John Vandenberg 23:16, 9 October 2007 (UTC)

it is, indeed ThomasV 06:25, 10 October 2007 (UTC)

further to 1001 NightsEdit

Further to your comments, wrestling with a scanner to take picture of a 400-page book does not strike me as a useful way to spend my time, not to mention the risk of such wear and tear to the book itself. If someone else likes scanning though, I would not stand in their way. And 1001 Nights has 17 volumes. The page numbering is not there with a view to dividing the book into separate files for each page, but to allow for more detailed cross-referencing from other places. Dividing the book by tales makes it much more readable for the visitor. In the head page, The Book of the Thousand Nights and a Night, I have already added links to bring a person to when a particular night begins, and I plan to use a similar system for dealing with the index at the end of the volume.

I do admit that at some point I should learn to use bots, and I did look at the User:ThomasBot/Help:Bots page. That page seems to give an adequate explanation of the "Requete" template's structure, but what seems to be missing is where to put the template so that it will work.

While I can see how having page scans can help in terms of preventing vandalism, putting this as a primary goal is exactly what I saw as paranoid in my earlier responses to Pathoschild. It is also based on a general assumption of bad faith on the part of people who just come to add in their favorite material. Most people do not attempt to add wilfully deceptive materials; assuming so would be contrary to the fundamental assumptions that have made wikis so successful. That being said, sure, your proposal will do everything that you said it will. The risk lies in discouraging those people with very good content who may appear clueless in technical expertise. That state of things will not encourage the growth which is absolutely necessary if we are to keep up with the competition. Eclecticology 16:48, 25 October 2007 (UTC)

the question is not whether we shall assume godd or bad faith from contributors, but rather what external readers assume from us. IMO, scans provide a guarantee to external readers, who do not necessarily assume good faith (or sufficient expertise) from us. There is a difference between "anyone can improve our texts" and "our texts are backed by scans". ThomasV 10:48, 26 October 2007 (UTC)

odd bot behavior?Edit

Have a look at Page:H.R. Rep. No. 94-1476 (1976) Page 154.djvu. The page quality for this page previously stood at 50%, then User:ThomasBot came along and moved the {{PageQuality}} template inside the <noinclude> tag. Now, when I edited the page and incremented the page quality counter to 75%, the result was that both the “50%” and “75%” icons now appear in the “page” tab at the top of the screen. Tarmstro99 18:27, 5 December 2007 (UTC)

sorry, this is because I have not updated the javascript. I'll try to do this tonight. ThomasV 23:07, 5 December 2007 (UTC)
not a problem, and no real hurry. Tarmstro99 00:11, 6 December 2007 (UTC)
it is done. please reload any page to update your javascript ThomasV 08:18, 6 December 2007 (UTC)
Thanks for working on this. At the moment, the problem seems to still be present. Following this edit to Page:H.R. Rep. No. 94-1476 (1976) Page 153.djvu, the page once again shows both the (preexisting) “50%” icon and the (new) “75%” icon in the Page tab. Tarmstro99 14:37, 6 December 2007 (UTC)
as I said, you need to update your javascript. click reload in your browser ThomasV 15:10, 6 December 2007 (UTC)
clicked reload repeatedly, even tried a different browser. Still seeing “[[Image:50%.svg]][[Image:75%.svg]]” in the Page tab at Page:H.R. Rep. No. 94-1476 (1976) Page 153.djvu. Tarmstro99 15:21, 6 December 2007 (UTC)
oh sure, that page is corrupted. updating your js ensures that you will not corrupt further pages :-) ThomasV 15:22, 6 December 2007 (UTC)
ah, got it. Thanks. Tarmstro99 15:29, 6 December 2007 (UTC)

Talk pageEdit

At first, I'm like, why was my talk page whacked, but then I checked the log and saw a notorious vandal/sock tried to get back at me via wikisource. Thanks for helping. RlevseTalk 15:23, 8 December 2007 (UTC)

a "bug" reportEdit

Hi, a bug report in case want to either fix it or explain it. John Vandenberg 05:26, 11 December 2007 (UTC)

OCR bot; two columnsEdit

Hi, following up on this, can the bot handle two columns? John Vandenberg 09:02, 30 December 2007 (UTC)

not currently. it is based on Tesseract, which does not handle this. ThomasV 15:32, 5 January 2008 (UTC)

History of West AustraliaEdit

Thanks Thomas but I've rem'd pages9-11 transclusions as they haven't been proofread yet and the OCR quality is so awful they are virtually unintelligible (esp as the two column thig isn't working). I'm going to have to rescan these and try another tack.

My little desktop scanner at home seems to manage the OCR better, except it struggles with the page size. But I will overcome! Moondyne 07:35, 15 January 2008 (UTC)

Thanks, I suspected that was the wrong thing to do but couldn't think of an alternative. How do you transclude a section? - thats news to me. Can you do it so I have a working example? Also, for your information, I don't intend to do any more saving of page images and creating [[page:]]'s. It takes too long and so I'll go straight into text with my little OCR scanner. Moondyne 16:46, 16 January 2008 (UTC)

tiff supportEdit

Any chance you could add tiff support to proofread.js proofreadpage_default_setup(). When the pagename is a tif/tiff, this is the HTML to create and add:

 <object classid="CLSID:106E49CF-797A-11D2-81A2-00E02C015623">
  <param name="src" value="..." />
  <embed type="image/tiff" src="..." />

That will allows us to at least set up pages like [[Page:EB1911-Vol15-EC5A664.TIF]] as an interim solution.

Thinking further ahead, maybe we can improve mw:Extension:ScanSet to act like a DJVU file. i.e. if we create a dummy [[Image:EB1911_tiff.scanset]] and it understands [[Image:EB1911_tiff.scanset/Vol15/EC5A664]] (and converts the TIFF to PNG). John Vandenberg 01:35, 16 January 2008 (UTC)

i would prefer the second option, because tiff suport is not widespread. also, a scanset is naturally ordered, which means that no index page is needed to create navigation buttons (like in pdf/djvu files).
unfortunately I do not have much time to do this myself atm. feel free to send me a patch, though.
ThomasV 08:17, 16 January 2008 (UTC)

[[:File:Major_General_H._P._Babbage.png]] is refusing to work for the Page namespace: [[Page:Major_General_H._P._Babbage.png]]. Any ideas? John Vandenberg (chat) 00:51, 30 January 2008 (UTC)

it works now. there was a bug yesterday, perhaps you encountered it ThomasV 08:00, 30 January 2008 (UTC)

PageQuality -> TextQualityEdit

I noticed you recently made some changes to pages that are part of Index:History of Iowa From the Earliest Times to the Beginning of the Twentieth Century. I was going to get around to asking for someone to change the templates, but you did it before I even asked :). However, only pages i through xviii were changed. I'd be grateful if you could change pages 1 through 325 as well. Also, any pages that don't have the template on them should. They were all proofread by me.

On a side note, you have a minor error on your bot user page. The description says: "The consequences of an erro are under your responsibility!". You could replace that with "The consequences of an error are under your responsibility!".

Thanks! Psychless 03:56, 2 February 2008 (UTC)

hello psychless,
I used a robot in order to update all the pages where the template pagequality was present, because its syntax changed.
All the other pages remained unchanged, so i guess this is why pages 1 to 325 were unchanged.
however, you're not using this extension in an expected way : there are no scans with your pages.
this results in a javascript error, and the quality buttons are not visible (at least not for me)
so I suppose that you added the pagequality template manually ?
why don't you just upload scans ?
ThomasV 11:49, 2 February 2008 (UTC)
I realize I've misunderstood the use of those templates. I've been using Google's page scans. I think it would be best if all of the page/text quality templates were removed. If you can do this with your bot notify me; otherwise, I'll just manually remove them. Psychless 00:20, 4 February 2008 (UTC)

hello psychless!

It is possible to use this extension with google books. For that you have to figure out the url of each image. to show you an example, I modified Page:History of Iowa From the Earliest Times to the Beginning of the Twentieth Century/4/2. (the image is wrong, it is just meant to show you how it works) when the image is present, you can see that there are also 3 'page status' buttons in the edit window.

However, I do not recommend this solution, because google books is not a safe source; they sometimes emove books from public availability, with no apparent reason. you scans will be safer if you create a djvu file and upload it here.

ThomasV 07:43, 4 February 2008 (UTC)

Mediawikis for proofreadEdit

Hi ThomasV,

Do you know if local wikisources can change at any Mediawiki (I didn't find them!) the words "toggle noinclude sections visibility" (tooltip of "+" toggle button), "Page status" (in a page being edited), "problematic, not proofread, proofread" (autotext when checkbox pressed), "{{PageQuality|3|<User>}}" (auto-noinclude-text; local wikisources can have built PageQuality template before this autotext, so now it must be adapted; it would be a good idea to announce at wikisource-list or local scriptoriums these advancements [wikisource news by Pathoschild is not updated anymore :o( ])?. Thanks, again! -Aleator 02:02, 3 February 2008 (UTC)

they are in Special:Allmessages.
However, the PageQuality template is hardcoded, you need to define it on your wiki. This template must also put the page in the corresponding category.
are you on ? it would be good to request a 'page' namespace.
ThomasV 07:50, 4 February 2008 (UTC)

Wikisource:Scan parties questionEdit

Hi Thomas, I noticed that you are the author of List of authors about to enter the public domain and I have a question about w:Ernst Ludwig Kirchner, who died in 1938.

Is this artist's work now in the w:public domain in countries where the rules are artist life + seventy years? --Jahsonic (talk) 14:47, 5 February 2008 (UTC)

i do not know; not an expert about legal questions ThomasV 17:16, 5 February 2008 (UTC)
His works are copyright throughout 2008, which is the 70th year of copyright protection. In some countries, his works will be public domain on Jan 1, 2009. John Vandenberg (chat) 21:03, 5 February 2008 (UTC)

French translationEdit

Hey, sorry to bother - but would you be able to try your hand at A Farewell to Love? The language is too archaic for my basic skills, and I'd appreciate if you could give it a try :) Sherurcij Collaboration of the Week: Author:Sabine Baring-Gould 18:28, 7 February 2008 (UTC)

sorry but I am not interested in translation projects. meanwhile, I think translations should not be hosted on wikisource but on wikibooks. ThomasV 10:07, 8 February 2008 (UTC)

Index Pages Registering Changes to PageQualityEdit

Out of curiosity, how long does it (or should it) take for changes in the PageQuality setting of a page in the Page namespace to "migrate upward" to the corresponding Index page? I ask because I've been proofreading pages under Index:H.R. Rep. No. 94-1476 and incrementing the PageQuality settings, but the changes are not being reflected by a change in the colored background boxes on the Index page. Is that updating process automated, or must the Index page be updated by hand whenever the PageQuality setting of one of the underlying pages is altered? Thanks for any guidance. Tarmstro99 16:35, 13 March 2008 (UTC)

you have to 'purge' the cache in order to update the colours. For this, click on the 'pages' link, that has 'action=purge' in the URL ThomasV 10:32, 16 March 2008 (UTC)
Ah, I understand. Thanks very much for your help. Tarmstro99 13:19, 17 March 2008 (UTC)

pagelist syntaxEdit

A while ago I noticed that you changed one of the "Index:" files to present prettier page numbers. i.e. it displayed page numbers like "i, ii, iii" .. was I imagining this? I am reminded of this as a result of a discussion on WS:S#DjVu_Already_OCRed...

Also, while I am here, we have a proposal to move "Index:" pages into their own namespace; see WS:S#"Index" namespace. Sorry I didnt notify of this earlier. John Vandenberg (chat) 11:11, 19 March 2008 (UTC)

Proofreading in Page:Edit

I don't know if you have a better place to report bugs, but I have been having problems getting the Proofreading to be saved when using the toggle button in the Page: namespace. Here is an example of a problem page. I was wondering if the problem had to do with SCEhardt's syntax which already existed on the page. I am really out my depth in trying to troubleshoot this, but I thought you would like to know about the issue.--BirgitteSB 18:57, 20 March 2008 (UTC)

DoubleWiki broken?Edit

Hello ThomasV. DoubleWiki seems to have broken recently; see "Is DoubleWiki broken?" on the Scriptorium. —{admin} Pathoschild 08:44:55, 30 March 2008 (UTC)

Account renamingEdit

Please can you rename an account for me on The account over there is mentioned there, in a posting from Feb 16 2008 Thanks, AP aka -- 19:47, 22 April 2008 (UTC)

Bot decategorizedEdit

Hello ThomasV. Did you mean to remove the bot category from User:ThomasBot? —{admin} Pathoschild 22:58:51, 07 May 2008 (UTC)

no, it is a mistake. ThomasV 23:02, 7 May 2008 (UTC)


You were absolutely correct. Thanks for notifying me before I messed up all the pages. Psychless 19:18, 8 June 2008 (UTC)

Page statusEdit

Hi Thomas,

Would you mind updating the extension to give edit summaries of the form

"/* Not proofread */ "

instead of

"/* Not proofread */"

That is, a space at the end of the summary. That way you don't have to type a space before adding a personalised edit summary. This is also the way section editing works on Wikipedia. (I had a look through the MediaWiki namespace, and I couldn't find any way that I could do this myself, so I am assuming it is part of the extension implementation.)

Obviously I won't fret if this can't be done, or if it can but doesn't warrant updating the installation.

Cheers, Hesperian 12:57, 21 June 2008 (UTC)

done, that was easy. wait for r36530. ThomasV 18:01, 21 June 2008 (UTC)
Thanks. Hesperian 01:42, 22 June 2008 (UTC)


Sorry to bother you again. I was wondering if the value of the span id's in Template:Page are particularly important. I would like to replace <span id="pr_position"> with <span id="{{{num|pr_position}}}">, in the hope/expectation that this will make it possible to hashlink to a page.

My reason for wanting to do so is because I have been transcribing works of systematic botany, and these tend to be strongly crosslinked. For example, Flora Australiensis/Volume V/CIV. Proteaceae/28. Banksia#Banksia ericifolia lists a long series of previous treatments of that species, beginning with "R. Br. in Trans. Linn. Soc. x. 203". In this case "R. Br." should be linked to Author:Robert Brown; "Trans. Linn. Soc." should be linked to Transactions of the Linnean Society of London; "x." should be linked to Transactions of the Linnean Society of London/Volume 10; and 203 should be linked to page 203 of that volume. The last of these is problematic: I could link directly to Page:Transactions of the Linnean Society of London, Volume X - p. 203.png, but I think it is unwise to do so. It would be much better if I could link to Transactions of the Linnean Society of London/Volume 10/On the Proteaceae of Jussieu/Banksia#203, and have that link take me to the right spot. I could make this work by manually enclosing that particular {{page}} call with a <div id="203">, but I hope you'll agree it would be better if this were automated.

Actually, the ability to link to a particular page number seems to me like a fundamental operation. Presumably the need for it is far more prevalent than my rather obscure example.

Hesperian 13:44, 23 June 2008 (UTC)

I suppose it is not important. There used to be some javascript associated with this template, but now it is no longer active. So what you describe should, in principle, be feasible. I encourage you to try and test (perhaps first with another template). You might also want to ask User:Sanberg, he probably knows since he refactored that template. ThomasV 08:25, 25 June 2008 (UTC)
Okay, thanks, I'll have a fiddle in my sandbox. It should be pretty straightforward I think. Hesperian 11:50, 25 June 2008 (UTC)

Help with ProofreadPageEdit

Hi, ThomasV. Can you tell me what I did wrong here. All I did was edit the page (which had been proofread) and clicked the "validate" button on the page and then save it. Afterwards, the text quality strip goes away (which had been yellow).—Zhaladshar (Talk) 14:11, 24 June 2008 (UTC)

I do not know. it works for me.
could you (1) update your javascript (load any page and click on relaod) and (2) test if that happens again ? thanks
ThomasV 14:19, 24 June 2008 (UTC)

It works when I use User:ZSPuppet. There must be something about my monobook.js that messes with the proofreading.—Zhaladshar (Talk) 14:38, 24 June 2008 (UTC)

margin troubleEdit

Hi again,

I have been transcribing a document that makes use of a mix of indents, hanging indents and no indents; see, for example, Page:Transactions of the Linnean Society of London, Volume 10 - p. 367.jpg. Whereas I usually ignore these when transcribing text, they contain information in this case, so I have used the text-indent and margin-left properties to capture them. This works fine until I come to transclude all the pages into Transactions of the Linnean Society of London/Volume 10/An Account of a new Genus of New Holland Plants named Brunonia, at which point the text looks fine but the page numbers in the margins are all wiggly. Any suggestions on how to handle this?

Hesperian 01:05, 30 June 2008 (UTC)

sorry, I do not know. I guess this has to do with the 'page template.... perhaps you should ask jayvdb ThomasV 13:07, 30 June 2008 (UTC)
Okay, will do; thanks. Hesperian 13:13, 30 June 2008 (UTC)

About A Treatise on Electricity and MagnetismEdit

I'm proofreading (at my best) some pages of the book A Treatise on Electricity and Magnetism , just to learn some more trick. I see that you're the first editor... there's some to verify by the original. When there are two variables   and   mentioned, I presume that the first one is alwais   (alpha)... How can this be verified? The scanning quality is far from perfect. --Alex brollo 13:57, 1 July 2008 (UTC)

I assume that the djvu files on are better quality. I suggest that they are uploaded to commons, and then the current pages can be moved across to those images. John Vandenberg (chat) 14:13, 2 July 2008 (UTC)
The page move should be accomplished easily by stripping step 3 from User:TalBot/ and User:TalBot/ and changing the base names. Just contact me once you've uploaded the new DjVu.--GrafZahl (talk) 08:03, 3 July 2008 (UTC)
djvu files uploaded. John Vandenberg (chat) 11:07, 3 July 2008 (UTC)
do not forget to delete the redirects after the move ThomasV 08:57, 3 July 2008 (UTC)
What have I to do? I guess it's better I stop my edits. --Alex brollo 15:41, 4 July 2008 (UTC)
I saw that A Treatise... has been moved successfully! The scan quality is really good.
John, don't matter to load text into empty pages. I feel very comfortable to load texts from my text processor, after some pre-elaboration (i.e, introducing h3 tags to sub-headings and introduce empty templates where math codes are to be written). There is much work that can be done effectively offline with a rather good text application.
In the meantime, I suggest you again to use widely the Talk page of Index to give some good suggestion about general style of the work. --Alex brollo 19:19, 5 July 2008 (UTC)

error in template:optionEdit

The javascript for template:option is causing an error (I verified in FireFox), which was discussed in scriptorium#Bottom-headers, cont. I've commented that code out until a fix can be found. thanks Steve Sanbeg 18:39, 3 July 2008 (UTC)

move fixEdit

Thanks for that move fix. I must have hit Ctrl-V twice. :-( Had so many to do I didn't notice my error. Hesperian 14:38, 10 July 2008 (UTC)

Side by side image view for proofreadingEdit

RE: Diff, where is the place for the topic? Jeepday (talk) 16:06, 11 July 2008 (UTC)

I do not think that links to wikipedia should be added to source texts, but that is a separate issue. As far as i know, this site's policy is to accept such links. Thus, I suppose that there is a page describing this policy, and you probably want to add it there. this help page is about how to use the ProofreadPage extension; filling it with irrelevant information will only make it look more complicated than it is.
ThomasV 18:01, 11 July 2008 (UTC)
How about like this Diff? Provides direction to the the article guidelines without adding complication. Jeepday (talk) 20:24, 11 July 2008 (UTC)
ok that's great ThomasV 06:06, 12 July 2008 (UTC)

Proofread Page extension: missing licence and improving magnificationEdit

HI! First of all thank you for this great extension. It's very useful for all the Wikisource project, I use it on it.wikisource.

Only two point:

  • In the template in the extension page there is a red "No license specified", it should be fixed.
  • I don't know if it's possible, but the magnification seem to be improved. At the moment the extension give a sort of on/off magnification. I think should be usefull to have a pair of +/- like in this system. Do you think is a good idea improve the extension in this direction?

Thanks for your work, --Accurimbono 12:50, 14 July 2008 (UTC)

that zoom is in Flash. I do not think that you can smoothly zoom like that in javascript ThomasV 13:45, 14 July 2008 (UTC)

OK, I just suggest, I'm not a technician! :) Hi! --Accurimbono 20:42, 14 July 2008 (UTC)

Extension:Proofread pageEdit

As you are credited with the creation of this MediaWiki extension, I would like to ask what it would take for this script to be site-engine agnostic. What I mean is, I want the DjVu-to-PNG converter plus text layer displaying feature to be not tied into MediaWiki, but basically a giant function where you put in a .djvu file and you get out the png and text. You may respond by emailing me. Messedrocker 04:57, 24 July 2008 (UTC)

this question is not about the proofreadpage extension, but about djvu support in Mediawiki. ask Tim, or look in the code ThomasV 08:58, 24 July 2008 (UTC)

Pages validated by botsEdit

Hello Thomas,

Look at the history of Page:MKGandhi patriot.djvu/2. Your bot validated the page. I think bots shouldn't be able to do that. Best regards, Yann 11:37, 27 July 2008 (UTC)

you validated the page with 100%. ThomasV 15:14, 27 July 2008 (UTC)
Oh, the value of the notation has changed since I validated that page. Before, it was "proofread = 100%", now it is "proofread = 75%". So never mind. Yann 15:27, 27 July 2008 (UTC)


As both the note on the file, and the comment above yours on my talk page noted, the file was for translation purposes. But I think I will simply find another place to contribute texts. Libertatia 04:55, 1 August 2008 (UTC)

Large Djvu uploadEdit

Could you upload this djvu file of EB1911 Volume 1 to the Commons? It's about 80MB, and we want to keep it together. Psychless 20:02, 15 August 2008 (UTC)

no i cannot. I suppose you have to ask someone who has shell access ThomasV 08:05, 16 August 2008 (UTC)
Do you know someone with shell access who would be willing to do it? Psychless 13:58, 18 August 2008 (UTC)
perhaps Tim Starling. ask him on irc. ThomasV 16:32, 18 August 2008 (UTC)

proposal to encode a paragraph start at top of transcription pageEdit

Does this look ok to you? [2] Are there better solutions in use on and/or --John Vandenberg (chat) 13:58, 19 August 2008 (UTC)

Weird annotationEdit

Hi ThomasV,

I don't understand why this note is placed vertically in page mode here and here. Is the display vertical only on my screen or on other ones as well? It is normal on this page and on this one, though. Is there something that I have forgotten? Thanks if you can help.- --Zyephyrus 22:44, 30 August 2008 (UTC)

Final push for the Proofread of the Month...Edit

This month's Proofread of the Month, Index:The Pilgrim Cookbook.djvu, is still a ways away from being fully validated. However, we're within striking distance.

If all ten members proofread just two (but preferably three) pages a day, we'll be able to finish the book before the end of the month.

We can do it. :) EVula // talk // 01:20, 24 September 2008 (UTC)

Basic index edit pageEdit


I would like to be able to copy Index: pages. For example, I would like to be able to copy the code from Index:The_Perth_Gazette_and_Western_Australian_Journal_1(3).djvu, and paste it to Index:The_Perth_Gazette_and_Western_Australian_Journal_1(4).djvu, Index:The_Perth_Gazette_and_Western_Australian_Journal_1(5).djvu, Index:The_Perth_Gazette_and_Western_Australian_Journal_1(6).djvu, Index:The_Perth_Gazette_and_Western_Australian_Journal_1(7).djvu, etc; making only tiny changes each time. Unfortunately the form-based interface makes it impossible to do so. What I need is something akin to the Commons basic upload form, which allows one to bypass the form-based standard upload form. What do you think?

Hesperian 04:54, 30 September 2008 (UTC)

you may temporarily disable javascript ThomasV (talk) 21:34, 30 September 2008 (UTC)
Will do; thanks for the tip. Hesperian 23:10, 30 September 2008 (UTC)

A includeonly problemEdit

Hi ThomasV,

A includeonly part of this page ought to appear in this one but it doesn't. Can you understand why? I don't know how to obtain the two parts of the poem in one another continuity. It doesn't work with two pages transcluded. Can you find a solution for that without breaking the layout? Thanks if you can help, it is our Project of the month proofreading.---Zyephyrus (talk) 17:21, 26 November 2008 (UTC)

Calling OCR by a botEdit

I'm boldly testing my bot Alebot into it.source; I know some python and I really enjoy such an opportunity. Recently I created some pages into a Page: spacename with a {{OCR}} only content, and all ran marvellously. Then I created some Page: pages with a different content, then I tried to replace such content with {{OCR}}, but nothing happened. I suppose from this mistake by me, that OCR is called only when you create a new page and put the template as the first content. Have I to delete wrong pages, and to put {{OCR}} into them when I create them again? Can you give me a link to learn more about OCR & bots? Thanks! --Alex brollo (talk) 10:56, 9 December 2008 (UTC)

no, if nothing happens if is because my bot crashed ThomasV (talk) 13:07, 9 December 2008 (UTC)

Merry Christmas by Alebot :-)Edit

Hi Thomas, and Merry Christmas by Alebot ... and by me, its driver.

Just to let you know that I'm working about python bot "post-elaboration" of OCR output with encouraging results, since I found that there are many typical and recurrent OCR mistakes that can be found and fixed by a script. I'm working on it:Indice:Manzoni.djvu; the current pre-pre-beta script is ocr procedure in the collection of procedures here: it:Utente:Alebot/Scripts. --Alex brollo (talk) 13:30, 24 December 2008 (UTC)

Bot not runningEdit


Your is bot is not running now. Could you start it please? Thanks, Yann (talk) 16:58, 28 December 2008 (UTC)

The bot is running, so thank you, but there is a problem with the OCR; it is completely useless: Page:MKGandhi patriot.djvu/124 :((. Even in French, I got better results with scans of this quality. Yann (talk) 15:28, 31 December 2008 (UTC)

On Page: Proofread status lost when adding category with HotCatEdit

Gday Thomas. To a proofread page when I add Category:Works needing image cleanup using the HotCat (preferences) gadget, the proofread status disappears, so it has to be reinstated with another edit. I am presuming that this is due to the HotCat, though I just wanted to float the query past you first. Thanks. -- billinghurst (talk) 06:48, 31 March 2009 (UTC)

I guess too. ThomasV (talk) 18:07, 31 March 2009 (UTC)

pagelist forgets romanEdit

Hi Thomas,

There seems to be a problem with <pagelist>. Before,

<pagelist 1="1;roman" 6="1"/>

would yield

i ii iii iv v 1 ...

Now, it yields

i 2 3 4 5 1 ...

Hesperian 14:21, 31 March 2009 (UTC)

hi. the syntax changed (see mailing list). ThomasV (talk) 18:06, 31 March 2009 (UTC)
There is other secret information out there, somewhere else? Cool. Which mailing list, and where are the archives? If it is not too rude to ask. :-) Thx. -- billinghurst (talk) 23:23, 31 March 2009 (UTC)
Thanks Thomas. Is this the usual Wikisource mailing list? If so, I guess I'll get your messing in the next digest. Hesperian 23:45, 31 March 2009 (UTC)
the secret information is at Wikisource:Mailing list. there's even a secret link to the archives ThomasV (talk) 06:05, 1 April 2009 (UTC)
Is there a way to make the roman numbers start at one? I'm looking at Index:Debates in the Several State Conventions, v4.djvu. Page 5 should start with roman numeral i, not v. The old way (semicolon) doesn't seem to be working. --Spangineerwp (háblame) 06:09, 1 April 2009 (UTC)
Thomas was so kind as to demonstrate the new format for me here. I think you will find your answer there. Hesperian 06:36, 1 April 2009 (UTC)
Got it now—thanks! --Spangineerwp (háblame) 06:48, 1 April 2009 (UTC)

Problem with Javascript on Index: pagesEdit

There seems to be a problem with the Javascript on Index: pages. I have twice edited Index:Charles von Hügel (1903 memoir).djvu (which I didn't create), entered an edit summary and hit save, only for the Javascript to erase my edit summary and replace it with "/* new index */". Hesperian 12:14, 5 April 2009 (UTC)

yes, I know about that. thank youThomasV (talk) 11:50, 6 April 2009 (UTC)

Strange OCR cutoffEdit

I notice that Index:The Fight at Dame Europa's School.djvu produced strange OCR results, cutting off the lines long before the margin, resulting in only the left half of each page being transcribed, and hundreds of "half-sentences" being OCRed alone.

The problem doesn't seem to be with the file, since the OCR attempt seen here was much clearer and didn't have the same problems a page like Page:The Fight at Dame Europa's School.djvu/25 has. Sherurcij Collaboration of the Week: Author:Carl Jung. 06:36, 20 April 2009 (UTC)

why don't you use the text then ? ThomasV (talk) 07:59, 20 April 2009 (UTC)
I am going back and slowly replacing the OCR with the text; just alerting you to the issue in case it would help troubleshoot in the future. Sherurcij Collaboration of the Week: Author:Carl Jung. 05:16, 21 April 2009 (UTC)

proofreadpage rtlEdit

Hi Thomas, replied to you earlier here. Dovi (talk) 07:34, 28 April 2009 (UTC)

Hi, I'm obviously doing something wrong, because I get no zoom when I click either on the he: example or on examples here, whether on the Page: page or in edit mode. I found the "Zoom" discussion at Wikisource:Scriptorium but don't understand what horizontal versus vertical zoom is or how to do it. Dovi (talk) 11:39, 28 April 2009 (UTC)
the current zoom does not work by clicking in the image, there are buttons in the toolbar for that. in the future I plan to add mouse wheel zooming ThomasV (talk) 11:43, 28 April 2009 (UTC)

Thanks for Proofread extensionEdit

We don't often stop and take the opportunity to reflect and to say thank you. The work that you have been doing for the proofread extension is both useful and valuable. Thanks for the time that you take with its development, the work and time is appreciated.   -- billinghurst (talk) 01:18, 4 May 2009 (UTC)

I'll second that...   Dovi (talk) 02:09, 4 May 2009 (UTC)
thanks for the support :-) ThomasV (talk) 16:45, 4 May 2009 (UTC)

a suggestionEdit

Hi Thomas,

A minor ergonomic improvement suggestion for the ProofreadPage extension:

I often navigate through transcription pages by clicking on the left and right arrow buttons. Mostly, so long as I don't move the mouse, the same button remains under my mouse cursor the whole time, so I only have to aim my mouse at the forward button once, and thereafter I can move through pages by repeatedly clicking the mouse button.

This is true except for the first and last pages. On the first page there is no left arrow button, so the right arrow button is moved to the left. This catches me out often. I will click on the right arrow from page 1, and end up on page 2. Then, without moving my mouse, I will click again, intending to go to page 3, but the button under my mouse cursor is now the left-arrow button, so I will end up back at page 1 again.

This minor annoyance would be removed if you altered the code to disable the left arrow button on page 1, rather than removing it; and similar for the right arrow button on the last page.

Hesperian 00:53, 8 May 2009 (UTC)

Javascript problem: not filtering quotesEdit

Look at the title entry at Index:Illustrationes florae Novae Hollandiae.djvu. Now edit that page and look at the title entry. Everything after the first quote has been consumed by proofreadpage_index.js. It occurred to me that this bug could well make us vulnerable to code injection, but I haven't been able to confirm that. But it is a bug all the same.

(In future, would you prefer me to raise these bugs in bugzilla?)

Hesperian 02:00, 14 May 2009 (UTC)

thanks for the report. that bug does not expose to code injection. I've set up a cross-domain page oldwikisource:Wikisource:ProofreadPage for reporting bugs. ThomasV (talk) 06:07, 14 May 2009 (UTC)

Proofread bug with IE?Edit

Hello Thomas. Could you review User talk:Hesperian#Page:A Specimen of the Botany of New Holland.djvu/27 please. When I attempt to edit that page in IE the edit box vanishes after a couple of seconds. It works fine in FF. Here's a screen dump and here's the error screen. Djanga (talk) 07:49, 15 June 2009 (UTC)

Missing scroll bar in side by side edit viewEdit

You mentioned on WS:S that an update was made to the code for side by side proofreading; did this update happen to remove the scroll bar from the page image in the side by side edit view? It seems to have disappeared recently, at least for me in FF 3. --Spangineerwp (háblame) 19:09, 15 June 2009 (UTC)

yes, in vertical mode scrolling now zooms in/out. ThomasV (talk) 22:10, 15 June 2009 (UTC)


Cheers for that, and for everthing you do! Is there a simple way that I can "purge" a work? Cygnis insignis (talk) 13:45, 19 June 2009 (UTC)

visit the file page, add ?action=purge to the url. if the djvu file does not contain a text layer, you can even download it, ocr it, add the text layer, and upload it as a new version of the file. ThomasV (talk) 13:47, 19 June 2009 (UTC)

Thanks! Is this something anyone can easily do? (I tried modifying the url as you say but there must be more to it, as it isn't working for me) --Spangineerwp (háblame) 14:42, 19 June 2009 (UTC)

if it does not work it means that the file you purged has no djvu layer ThomasV (talk) 15:02, 19 June 2009 (UTC)
Got it now—I didn't realize it had to be done on the file in Wikisource's file namespace. Thanks! --Spangineerwp (háblame) 15:36, 19 June 2009 (UTC)
I messed about with this, without success, and managed to break something at Index:Voyage in search of La Perouse (Stockdale).djvu. If you can help, I promise not to try this sort of thing again :) Cygnis insignis (talk) 17:59, 21 June 2009 (UTC)
don't worry, you did not break it. it is a bug in my code. if it is what I think it is, then I already fixed it a few days ago, but the fix is not yet active. I will purge again once the fix goes live ThomasV (talk) 20:01, 21 June 2009 (UTC)
Is there a problem with PageNum? Cygnis insignis (talk) 17:10, 27 June 2009 (UTC)
yes, it is so heavy that it quickly reaches the transclusion limit, and it has been deemed insane by developers. use 'pages' instead, it does the same and is easier to use. ThomasV (talk) 17:49, 27 June 2009 (UTC)
If you mean <pages index />, it is an improvement that has a couple of bugs: a broken table at Description of the Natives and the previously given link Hermione/How Suffering Purifies One! I'll convert to {{Page}} and number manually if this is still being developed. Cygnis insignis (talk) 18:20, 27 June 2009 (UTC)
hmm, sorry I did not know about that bug (the broken table). is there another bug ? if yes, I do not see it ThomasV (talk) 18:25, 27 June 2009 (UTC)
No links to Page: space, and therefore no numbering, at Hermione. It worked at the other example, curiously. Cygnis insignis (talk) 18:38, 27 June 2009 (UTC)
the links are there. they are optional, see the 'links to scanned pages' link displayed in the toolbox ThomasV (talk) 18:40, 27 June 2009 (UTC)
Thanks for the replies. Cygnis insignis (talk) 18:43, 27 June 2009 (UTC)
ok, I just fixed the bug in svn [3]. thanks for the report. please be patient, the fix will not go live until the next update. ThomasV (talk) 18:56, 27 June 2009 (UTC)
Ta. Is it possible to the have links displayed on unfinished pages, perhaps defaulting to the current arrangement (off) when it is protected? The downside of this would be that users might not become familiar with this new option. I might use it myself, if I was reading a longer work ... why do we have Page:links in the left margin? Cygnis insignis (talk) 20:08, 27 June 2009 (UTC)
this is currently not possible. links are in the left margin because of the settings of the Page template. this can be configured differently, though. ThomasV (talk) 20:14, 27 June 2009 (UTC)
Thanks, I switched that on again. Is there a fix available for this, it still reads error. Cygnis insignis (talk) 18:23, 8 July 2009 (UTC)
not now ThomasV (talk) 08:54, 9 July 2009 (UTC)

{{PageNum}} and <pages index>Edit

I am guessing you removed the information about the PageNum template from H:Side for some reason... and I see Cygnis insignis has a similar question to mine: basically, is there something wrong with PageNum or is there some reason why we should not use it? I looked at the code and it looked awfully ungainly. It seems much easier to use than producing multiple instances of {{tl|Page}. Any suggestions and comments regarding this are appreciated. --Mkoyle (talk) 17:53, 27 June 2009 (UTC)

see above : you should use the 'pages' command ThomasV (talk) 17:56, 27 June 2009 (UTC)
Would it be acceptable to undo your deletion of the PageNum information at H:Side using <pages>? or do we need to wait a while to be certain pages works as expected? --Mkoyle (talk) 18:17, 27 June 2009 (UTC)
ok, undo me if it does not work as expected. ThomasV (talk) 18:27, 27 June 2009 (UTC)
I'll put pages in there and start using it and if there is any problem, I will let you know. Thanks, Thomas --Mkoyle (talk) 18:31, 27 June 2009 (UTC)

When editing a new Page: pageEdit

The text when I pull up a new Page namespace page is

There is currently no text in this page. You can search for this page title in other pages, search the related logs, or edit this page.

which doesn't seem to cover DjVu pages where there may be a new text layer. Is that text one of your creations, or is it a Mediawiki file? Thanks. -- billinghurst (talk) 04:37, 14 July 2009 (UTC)

it is a mediawiki message. on I configured it to be different for the 'page' namespace, using some javascript ThomasV (talk) 14:08, 14 July 2009 (UTC)

here is the relevant bit :

function add_msg(){
	if((!wgCurRevisionId) && self.proofreadPageIsEdit && self.proofreadPageViewURL) {
		var msg  = document.getElementById("msgNewarticletext"); 
		if(msg) msg.innerHTML = "<div style=\"border:1px solid #CCC; padding:0.5em;\">Vous avez suivi un lien vers une page en mode fac-similé, pour laquelle le texte n'est pas encore disponible. Vous pouvez compléter cette page en entrant le texte dans la boîte ci-dessous. Vous pouvez demander une reconnaissance automatique de caractères, en utilisant le service d'OCR de Wikisource.</div>";

Create namespace for vi.wsEdit

Hello Thomas, I have filed a request to create some namespaces for, but it seems that the assignee cannot do it right now. I'm in some kind of hurry to create Author pages and create djVu in Could you take a look at that bug and implement it for me because I know you're a developer. Thanks in advance. Vinhtantran (talk) 16:50, 24 August 2009 (UTC)

no, I cannot do that. you need to ask someone who's got shell access ThomasV (talk) 19:39, 24 August 2009 (UTC)

Problems with new updateEdit

Hi Thomas,

Please see Wikisource:Scriptorium#Problems with new update. Hesperian 05:44, 25 September 2009 (UTC)

Another problem?Edit

Hi Thomas,

Index:Mammals of Australia (Gould), introduction.djvu is showing seven pages as "Not proofread" when in fact these pages are "Validated". I have tried purging both the individual pages and the index. Any idea what is going on here?

Hesperian 12:46, 3 October 2009 (UTC)

I have seen that elsewhere. I do not know what causes it. null-edits fix the pages. ThomasV (talk) 16:26, 3 October 2009 (UTC)


Sorry, I don't mean to harass you. I guess you want to be told about these little issues....

If you click on The Botanical Magazine/Volume II, which doesn't transclude any pages, you'll be redirected to The Botanical Magazine/Volume 2. Just below the title it says "(Redirected from The Botanical Magazine/Volume II)"; this serves two important functions: it informs me that I was redirected, and it gives me access to the redirect page.

If you click on The Botanical Magazine/Volume I, which does transclude pages, you'll be redirected to The Botanical Magazine/Volume 1, but this time there is no "(Redirected from..." message. The green progress bar that the new update puts beneath the title sits in the precise location where the "(Redirected from..." message should have been. Therefore I assume that the progress bar has clobbered the redirect message.

Hesperian 05:21, 5 October 2009 (UTC)

yes I know about that one. it is already fixed in the code. thanks for the report ThomasV (talk) 19:59, 5 October 2009 (UTC)

And anotherEdit

Pages that are transcluded into the Table of Contents section of an index page are not colour-coded in the Pages section; see Index:Life of William Blake, Pictor ignotus (Volume 2).djvu. Hesperian 02:34, 6 October 2009 (UTC)

Hard-Coding Image Parameters in Index TemplateEdit

Would you mind taking a quick look at this issue? I am worried that, due to my ignorance, I may break things if I try to fix the problem. Thanks very much! Tarmstro99 (talk) 15:16, 8 October 2009 (UTC)

go ahead; if you break something you'll just need to revert your changes. you may want to look at the template, I optimized it to save screen space; check a few examples at (the same link is in the toolbar) ThomasV (talk) 16:15, 8 October 2009 (UTC)
Thanks, I will do so. I have been impressed with some of the examples on Tarmstro99 (talk) 20:38, 8 October 2009 (UTC)

In case it is helpful <pages> with #tagEdit

After some struggling, I was able pull together some tags with parameters that may be useful if we ever want to template <pages>.

This specific example is for DNB, and I have it configured to use |section, presuming that naming is consistent, and we do wish to subst: it.

{{subst:#tag:pages||index="Dictionary of National Biography volume {{{volume|}}}.djvu"

If of no use, it is fine. Waste not, want not. -- billinghurst (talk) 12:03, 12 October 2009 (UTC)

Polish WikisourceEdit

Hi, when you have time, could you see what is wrong with this: pl:Indeks:Opis ziem zamieszkanych przez Polaków 1.djvu. It isnt urgent, as the page transcription can begin; see pl:Strona:Opis ziem zamieszkanych przez Polaków 1.djvu/6. --John Vandenberg (chat) 03:28, 13 October 2009 (UTC)

Don't worry ... I have fixed it ;-)
John Vandenberg (chat) 03:32, 13 October 2009 (UTC)

proofread_index.js edit form construction fixEdit

Hi; this is a response from Wikisource:Scriptorium#Index page editing bug: double quote escaping.

I took a stab at fixing the problem described there, and came up with the following patch to construct the index page edit form using DOM methods instead of innerHTML. It's unrun and untested, but hopefully you can use it as the basis of a fix.

--- proofread_index.js.orig	2009-07-19 18:28:19.000000000 +0200
+++ proofread_index.js	2009-10-15 03:20:08.000000000 +0200
@@ -11,6 +11,17 @@
 	return '';
+function makeElement(tagName, attrs, children) {
+	var elem = document.createElement(tagName);
+	if (attrs !== undefined)
+		for (var name in attrs)
+			elem.setAttribute(name, attrs[name]);
+	if (children !== undefined)
+		for (var i = 0; i < children.length; i++)
+			elem.appendChild(children[i]);
+	return elem
 function proofreadpage_index_init() {
 	var toolbar = document.getElementById("toolbar"); 
@@ -33,14 +44,17 @@
 	var container = document.createElement("div");
 	var index_attributes = self.prp_index_attributes.split('\n');
-	var str = '<div style="display:none;"><textarea id="wpTextbox1" name="wpTextbox1">'+new_text.value+'</textarea></div>';
-	str = str + '<table>';
+	container.appendChild(
+		makeElement('div', {style: 'display:none;'}, [
+			makeElement('textarea', {id: 'wpTextbox1', name: 'wpTextbox1', value: new_text.value})])
+	);
+	var rows = [];
 		m = index_attributes[i].split('|');
 		param_name = m[0];
 		if(m[1]) param_label=m[1]; else param_label=param_name;
-		str = str + '<tr><td>'+param_label+': </td>';
 		value = findparam(params,param_name);
 		value = value.replace(/\{\{!\}\}/g,'|');
@@ -49,15 +63,16 @@
 		if( (m[0]=="Footer") && (value=="") ) value = prp_default_footer;
 		if(m[2]) size=m[2]; else size="1";
-		if(size=="1") {
-			str = str + '<td><input	name="'+param_name+'" size=60 value="'+value+'"/></td></tr>'; 
-		}
-		else{
-			str = str +'<td><textarea name="'+param_name+'" cols=60 rows='+size+'>'+value+'</textarea></td></tr>';
-		}
+		rows.push(makeElement('tr', {}, [
+			makeElement('td', {}, [document.createTextNode(param_label)]),
+			makeElement('td', {}, [size == "1"
+				? makeElement('input', {name: param_name, size: '60', value: value})
+				: makeElement('textarea', {name: param_name, cols: '60', rows: size}, [document.createTextNode(value)])
+			)]),
+		]);
-	str = str +'</table>';
-	container.innerHTML = str;
+	container.appendChild(makeElement('table', {}, rows));
 	var saveButton = document.getElementById("wpSave");
 	var previewButton = document.getElementById("wpPreview");

--Piet Delport (talk) 02:57, 15 October 2009 (UTC)

thanks, I 'll take a look when I have time ThomasV (talk) 13:00, 15 October 2009 (UTC)


Hi ThomasV

This request was put when ThomasBot was not connected. Do I simply wait now or do i put the request again to activate it? --Zyephyrus (talk) 13:24, 17 October 2009 (UTC)

It was perfect till the page 107 but now there is a problem: the command Match doesn't want to work past the page 108 here. I have tried again with a shorter text here from this page but it doesn't want to work either. What can I do? --Zyephyrus (talk) 16:35, 21 October 2009 (UTC)


—> Template talk:Page#Missing page numbers? Hesperian 23:52, 27 October 2009 (UTC)


See Wikisource talk:WikiProject OCR -> why is the OCR button not showing on Polish Wikisource? Sample scanned page - no text, no OCR button. --Piotrus (talk) 23:31, 29 October 2009 (UTC)

it is not part of the extension; you need to copy the relevant bits from Common.js ThomasV (talk) 07:57, 31 October 2009 (UTC)

Index:History of West AustraliaEdit

Hi Thomas. A while back you created that for me. Its now redundant to Index:History of West Australia.djvu. Can you zap it plse? Moondyne (talk) 15:38, 6 November 2009 (UTC)

Czech languageEdit

Some time have passed and I am comeing back to Wikisource with a question if you are able to read Czech language with your bot?--Juan de Vojníkov (talk) 09:25, 17 November 2009 (UTC)

it depends on what kind of task you want to do with the robot. Are you talking about the OCR task ? if yes, the ocr uses tesseract, and there is no czech version of tesseract. ThomasV (talk) 13:28, 17 November 2009 (UTC)

OK, thanks. I was talking about OCR.--Juan de Vojníkov (talk) 18:28, 17 November 2009 (UTC)

you can try with the english ocr; maybe it gives good results ThomasV (talk) 18:45, 17 November 2009 (UTC)

Proofread Page in HebrewEdit

Hi Thomas.

Recently at Hebrew Wikisource there has been much discussion and experimentation with the Proofread Page extension, which you have developed most successfully. It has been put to use with several texts, but two issues remain.

The first issue is the magnification. When we discussed this some time ago you mentioned that this had been fixed for RTL. It could be that we at he.wikisource are misunderstanding how it is to be used, but we still seem to see the same problem: Only a portion of the page image (mostly the right side) can be magnified by clicking and rolling. For an example, see any page of this book. It seems to work differently, by the way, in Firefox/Chrome than in Explorer.

The second issue is OCR. On the one hand, is there any way to implement Hebrew OCR into the extension? On the other hand, we work extensively with texts from a website called, which contains PDF scans of over 40,000 out-of-copyright Hebrew books, and has high-quality Hebrew OCR as well. The problem of course is that when these are converted to DjVu the OCR is lost. Is there any way to use the Proofread Page extension with PDF documents? Or is there any way to transfer the OCR when converting a PDF to a DjVu file?

(Also notice that obviously for RTL reasons the "next page" and "previous page" arrows appear wrong in at the top of individual pages.)

It would be great if you could offer any helpful ideas regarding either of these problems. Dovi (talk) 14:17, 8 December 2009 (UTC)

Hi Dovi
The magnification issue was fixed at some point, because I discarded the current zoom; I replaced it by mouse scroll zoom in edit mode, and no zoom in view mode. However some users complained about no zoom in view mode, so I put it back; and this zoom does not work at It is really difficult for me to fix it for rtl languages, because I do not really understand what is going on... In edit mode too, the scroll zoom behaves strangely. However, you can use the horizontal edition layout; it seems to work normally at You can set it up for your whole site, using
var proofreadpage_default_layout="horizontal";
The OCR button is not built into the extension. It is a javascript/ajax gadget that calls a server running on the toolserver. But this OCR uses Tesseract, and I don't think it works in hebrew. In any case, this button method is completely deprecated; it is much much better to work a djvu file that has an ocr text layer. When it is edited for the first time, a page from a djvu file is preloaded with the OCR text layer of the djvu. This allows you to do the ocr with whatever tool you have. So, the procedure is 1. do the ocr of your book. 2 create a djvu file from the pdf, with the ocr text inside 3 upload it to commons. 4. create index page at
The next/previous page issue seems easy to fix. I will do that.
ThomasV (talk) 15:17, 8 December 2009 (UTC)
here's a patch for the next/prev buttons :
it seems to work
ThomasV (talk) 15:38, 8 December 2009 (UTC)
Hi Thomas. Getting back to this after quite a long time... Thanks for the next/previous patch.
The real fundamental issue standing in the way of adoption is OCR. You write in 2: "create a djvu file from the pdf, with the ocr text inside". But how in the world does can one do that? If there is a program that can accomplish this it would be amazing.
I'll ask people to experiment more with the zoom. Dovi (talk) 10:38, 2 February 2010 (UTC)

Category: and Match and SplitEdit

Hi TweekGuru. Noticed with Match and Split that if there is a category at the end of a work, that it is collected by the last split page. Is it possible for the next rendition of Split to ignore Category:... when it is pasting to the Page: namespace? Thx billinghurst (talk) 06:22, 13 December 2009 (UTC)

no, you have to move at the top of the page what you dont want to be moved in the page namespace ThomasV (talk) 09:08, 13 December 2009 (UTC)

match & split with url containing " "Edit

Just to make a note that when a work url/name has " " in the title, that it just complains and doesn't inform of anything. When work moved to without " " all worked fine. billinghurst (talk) 05:01, 15 December 2009 (UTC)

transcluding tables into index pageEdit

Hi Thomas,

I have seen a lot of problems lately with transcluding tables into index pages. This is not an uncommon situation, because often contents pages get transcluded to the index page, and of course contents pages may be laid out using tables. See Index:Popular Science Monthly Volume 2.djvu for an example of this problem. Do you know how to fix this? Since this is not an uncommon problem, is it possible to fix it in the index template, so that these problems do not keep occurring?

Hesperian 23:57, 20 December 2009 (UTC)

A <br /> with the table starting a new line fixes the problem. Undertaken in above link.
Note that this is more than tables where this occurs and it seems to be transcluding of wiki code, rather than other code, and seems more related to abutting <noinclude></noinclude> than the Page: environment per se. Another example is when the start of the file ==Chapter 2== where it shows it as is, rather than wikify it. Same thing happens if you abutt the start of the Notes field of {{header}} with similar wiki code. billinghurst (talk) 01:27, 21 December 2009 (UTC)
Strange that <br /> works but neither {{blank line}} nor {{nop}} do... but as you say they are "wiki code, rather than other code". Hesperian 01:29, 21 December 2009 (UTC)
Doesn't {{blank line}} only work in main namespace, not Index: Maybe we should invert its function so that it only does nothing in Page: namespace. billinghurst (talk) 03:04, 21 December 2009 (UTC)
I wrapped this section inside a div. it should fix it. ThomasV (talk) 01:56, 21 December 2009 (UTC)
Thanks. Hesperian 02:36, 21 December 2009 (UTC)

Feature requestEdit


Often when validating pages, I make some changes, add a comment to the edit summary box, then push the validated radio button. This deletes my edit summary and I have to write it again. It would be good if the Javascript could be tweaked to preserve whatever is already in the edit summary; but of course you don't want to preserve the old status. Maybe the best way to implement it would be to remove anything between "/*" and "*/" at the front of the edit summary, then prepend whatever is left with the new status, e.g. "/* Validated */".

Hesperian 00:33, 24 December 2009 (UTC)

Match and Split discussion (parking for later)Edit

Feature request for Match. Wondering whether Match would be able to either have a stop/skip phrase; where the purpose is to match to a certain point and then stop, or be able to get it to skip for a stretch of marked pages. This would enable batches of blank/image/... pages to be excluded and then allow match to continue through. Often easier to block out sections at the beginning, than to repeatedly undertake the work

Similarly, if one was to mark the pages as without text, or possibly something unique, so that Match & Split would know not to touch those pages when it was strolling through.

Just thoughts to record for whenever you get back to M&S. Thanks. billinghurst (talk) 06:15, 28 December 2009 (UTC)

i guess i'm going to instruct it not to touch pages that are not uncorrected ThomasV (talk) 13:28, 28 December 2009 (UTC)

Forgetting to stopEdit

Occasionally, I see that M&S forgets to stop at the end of the document and continues off into its own fairyland. Here is my latest example. -- billinghurst (talk) 06:35, 28 December 2009 (UTC)

Index: page - document typeEdit

Hi Thomas. With the document type in the Index: namespace, I would like us to consider where historical documents would fit. From the sources that you have been referencing, have you an indication of what they should be called? Here I am thinking of court documents, FBI documents, Wills, etc. Some of the stuff that does not fall under the traditional publication criteria, though still falls within our scope of WS:IO. Thanks. billinghurst (talk) 07:01, 6 January 2010 (UTC)

the different document types have been introduced in order to be able to properly cite our works, using automatic generation of headers. I do not know if historical documents require an extra type for that.
for the moment I have had difficulties to install this feature at, because of the page layout used here.
ThomasV (talk) 02:55, 10 January 2010 (UTC)

The SOURCE tab link, its creationEdit

Hi TV. I thought that the use of <page> om a page was the trigger for the appearance of the sources tab, however at Mandragora it doesn't appear. Is there a way that I can push it to occur? billinghurst (talk) 07:05, 8 January 2010 (UTC)

The adoption of 'running header' for the TOC may be a clue, I noticed some strange behaviour when it is wrapped in another formatting template. eg. {{smallcaps|{{rh|[[Demeter Consolatrix]]||42}}}}
Hmm, I transcluded just page 11 (blanked) and no change. Transclude other pages, and it is produced, undo changes, and it disappears. I try to remember where else I have seen it, and look for similarities. billinghurst (talk) 11:06, 8 January 2010 (UTC)
Put a running header on a page and transcluded, no problems, so it is not a simple issue of {{rh}}. FWIW, wrapping RH is a bad idea, as it is a series of stacked DIV. I learnt that lesson early. billinghurst (talk) 11:11, 8 January 2010 (UTC)
All been replaced by a table (which is nicer anyway as you can set the overarching style); no change. billinghurst (talk)
it's a bug; I guess it has to do with the fact that these pages are transcluded on the index page ThomasV (talk) 18:14, 9 January 2010 (UTC)
Undertook test, and when the pages are not transcluded, and you are correct, in that the SOURCE tab reappears. For the hell of it, I transcluded using the <pages> nomenclature, which I know doesn't work in Index: ns, and that also broke the SOURCE tab. billinghurst (talk) 23:34, 9 January 2010 (UTC)
Found that IF another page from the same work is transcluded that this is enough to force SOURCE to appear. So the solution is to not just transclude the ToC. billinghurst sDrewth 07:41, 23 January 2010 (UTC)

Remove categories from match and split sourceEdit

Hi Thomas,

I guess this page ended up with a category because the original source had a trailing category? Since we generally don't categorise Page: pages, would it be possible to strip categories off the source before performing the match and split?

Hesperian 02:00, 10 January 2010 (UTC)

the person who performs the split has to take care of that ThomasV (talk) 02:49, 10 January 2010 (UTC)
/me pleads that was an early one, and I have learnt more since! I am pretty sure that I added that information to the help pages. I will check to ensure that it is there, as well as my other learnings. billinghurst (talk) 03:13, 10 January 2010 (UTC)
Okays all round. Hesperian 03:15, 10 January 2010 (UTC)
[Technical note:] There is a step between the match and the split where the user cleans up the sections of text for the Page namespace.
In some fashion they are the one, the MATCH includes the check, as a no match due to more than one blank page, requires you to manually match, and then __MATCH__ again until the page is complete. There is also a user clean component AFTER a split too.
Would this be fixed by calling the first section MATCH (and Tidy). Followed by SPLIT (and Confirm)?

Thomas, I did something with this cool procedure that you might be interested to know. I had two Indexes which reprinted an old text, one was already proofread so I used subst: to create a new page. I then ran a match and split to the second Index, creating formatted and linked Pages that only needed checking and headers. Cheers, Cygnis insignis (talk) 03:38, 10 January 2010 (UTC)

Addendum. I have talked to Thomas about a number of these sorts of tweaks that I could see by helpful, eg. a STOP or SKIP function. He has kindly told me that his priority is working on the update to better integrate Proofread Page into the usability feature. I have backed off and left him to those more important bits. billinghurst sDrewth 06:15, 10 January 2010 (UTC)

Gif images in Page:Edit

What can we do with something like Index:Littell's Living Age where I find that it is gif files, and they don't display in Page: ns. I feel that if we cannot do anything with them, that they should be ripped from Index:/Page: and converted to namespace with images. Thoughts? billinghurst sDrewth 23:54, 25 January 2010 (UTC)

Page and Pages links in other skinsEdit

Gday Thomas. Saw comment at User talk:Wknight94 about the links when using {{Page}} or <pages> not working well in other skins. I had a look, and can confirm that the said skin disappears into the links of the sidebar (which obviously isn't a sidebar). I am not sure whether it is an artefact of the transclusion process, or how we work our templates. billinghurst sDrewth 10:35, 26 January 2010 (UTC)

Defaultsort issue from Index ...Edit

Hi ThomasV, I believe that this has been addressed, I just need to double check.

Looking at A Descent into the Maelström, I have applied a DEFAULTSORT: after doing the match and split. You will see that it is now obviously seeing the DEFSORT from the Index: ns. You did something for the next release of the software after we found a gremlin from the DNB side, and just wondering that same fix will override this DEFSORT issue. Thanks. billinghurst sDrewth 08:31, 31 January 2010 (UTC)

Weird <ref> issuesEdit

/me again (he says timidly). I have a ref in a ToC which is transcluded to an Index page

and it also shows on

Phe undid the transclusion on the Index page and it disappeared. We both hoped that it would be covered by previously identified fixes, however, just in case it isn't. billinghurst sDrewth 12:49, 31 January 2010 (UTC)

after the next update I'll do a round tour of all theses issues to see if they are still there. ThomasV (talk) 17:34, 1 February 2010 (UTC)
sounds like a plan. billinghurst sDrewth 21:32, 1 February 2010 (UTC)

OCR bot needs a restart (offline)Edit

Hi ThomasV, when you are next on toolserver, would you mind kicking the daemon for Thomasbot's OCR. Thanks. Nothing urgent. billinghurst sDrewth 12:36, 2 February 2010 (UTC)

doneThomasV (talk) 12:47, 2 February 2010 (UTC)

Dictionary of Music and MusiciansEdit

Hi, I see you've deleted all the mainspace pages I had created for the C's with the note that these are redundant. Actually, they aren't redundant. Most of them contain transclusions from the Appendix in Vol 4 as well as Vol 1 and your bot has only run over Vol 1.

I'm not happy about their disappearance at this point. Aside from it representing over 50 hours work on my part that's just evaporated, until I can get this laptop updated I can't see the version that has replaced them and therefore I can't see how the articles look in their transcluded state.

For the usefulness of the Dictionary I was not intending to only have four pages - one for each Vol - as this will make it impractical for on-line use. My intention is to have 26 pages, each one combining the articles beginning with the same letter regardless of which volume the article is contained in. Is this possible in the proposed "dictionary" transclusion method? Beeswaxcandle (talk) 10:09, 5 February 2010 (UTC)

oh sorry, I thought it would help you. do I need to undelete some of the articles I deleted, or would it be sufficient to run the robot on the other volumes?
do not worry about the index by letters. It will be very very easy to create 26 pages, it does not have to be one page per volume.
ThomasV (talk) 11:03, 5 February 2010 (UTC)

I fixed the code so that it works with IE6. it should work on your laptop now. please let me know. to update your browser's javascript with IE, you have to press control-f5 ThomasV (talk) 12:56, 5 February 2010 (UTC)
Thanks Thomas, yes it now works on my laptop. I've had to ctrl-F5 on both of the pages. Beeswaxcandle (talk) 18:04, 5 February 2010 (UTC)
oh, I just realized that most articles in vol 4 are addendums to other articles. I now understand your point. It should be possible to transclude an article with its addendum. I'll change my code to make it possible. ThomasV (talk) 13:41, 5 February 2010 (UTC)
I'll work at getting the A-section from Vol 1 and 4 sorted, then we can have a look at how it works. It will take me a couple of weeks, so don't rush on this code change. Cheers, Beeswaxcandle (talk) 18:04, 5 February 2010 (UTC)
Are we going to end up with anchors to article? I am wondering how we would like to identify and extract bits from the viewpoint of the contributors at the end of each of the articles. I am hoping to get all the authors' pages completed soon, and now wondering if there will be a ready means to extract the author data from the respective sections. Was looking at something queryable to build the author pages, or a ready extract that we might be able to transclude. billinghurst sDrewth 10:28, 6 February 2010 (UTC)
Don't think you'll need to do any extracting here. At the end of Vol. 4 there is a "Catalogue of Articles Contributed by Each Writer in the Dictionary". If we just sectionalise these 8 pages, the sections can easily be transcluded to the Author pages. Beeswaxcandle (talk) 00:05, 7 February 2010 (UTC)
Very nice. I will try to get to have a look-see. billinghurst sDrewth 01:14, 7 February 2010 (UTC)
The couple of weeks has turned into a couple of months, but the A-section from Vol 1 and 4 are now sorted. When you've a moment can you wave your wand and see what the results are? Thanks, Beeswaxcandle (talk) 09:07, 9 April 2010 (UTC)

<pages> tag is populating the links table with links instead of transclusionsEdit

Hi Thomas,

I reckon there's a reasonably serious bug there. Can you see the discussion at User talk:Hesperian/V#Transcluded pages appearing please.

Hesperian 14:43, 17 March 2010 (UTC)

I don't think I should look at it before the current code is updated. Please remind me after the next code update, if that bug is still there. ThomasV (talk) 07:24, 18 March 2010 (UTC)
Okay. Hesperian 07:25, 18 March 2010 (UTC)
Seems to be fixed; thanks. Hesperian 02:25, 13 April 2010 (UTC)


And thanks also for all your efforts with the ProofreadPage extension. It bothers me a little that your reward for all this seems to be strident complaints about every minor glitch. Hesperian 02:25, 13 April 2010 (UTC)

We are grateful for the excellent interface which we now have thanks to Thomas. I hope he doesn't consider every concern or criticism a strident complaint. Moondyne (talk) 02:38, 13 April 2010 (UTC)
I could understand if he did. Before this gripe there was another gripe. And another before that. And before that there was a huge blow-up on the mailing list. And so it goes. (I wasn't accusing any individual of griping, nor am I complaining about any particular thread; what bothers me is the pattern—a pattern that spans several years by now.) Hesperian 03:02, 13 April 2010 (UTC)
Amen. — billinghurst sDrewth 03:26, 13 April 2010 (UTC)
Thanks for your kind messages. It's normal to hear people complain about glitches. ThomasV (talk) 08:37, 13 April 2010 (UTC)

A belated thanks and a questionEdit

Please add my thanks to the previous list for your efforts, and everything is working fine (for me). A question of secondary importance is having to display the header/footer to access my custom toolbar (in either Monobook or Vector). Is this going to be permanent? — Ineuw (talk) 17:03, 20 April 2010 (UTC)

I don't know why everything seems too work fine for you now. there has been no update in the meantime, afaik. ThomasV (talk) 05:07, 21 April 2010 (UTC)
I have everything needed for my working environment with the exception of the custom toolbar. Instead of the mouse wheel magnification, I use the toolbar button to enlarge. It just took some time to get used to the changes.— Ineuw (talk) 13:36, 21 April 2010 (UTC)

Small suggestion for ProofreadPagesEdit

Hi ThomasV. First things first, I would like to say thank you for providing the PP extension, it has improved Wikisource vastly since my firsts visits some years ago!

I would like to make one small usability suggestion to save a couple of mouse clicks. If you are on a Page: namespace page, and you wish to go to the File: page, you have to to go via the Index: page. This is a little clunky for me (my connection is not very fast, so it takes some time to go via another page), as I make this trip often when looking for the source of the file. Would it be possible to make a direct link on the top bar that goes directly to the File page?

Secondly, I had a thought a while back about about a further "reverse transclusion" link. Very often I need to check how a page looks when transcluded into the main namespace, but to see that you have to go to "What links here" and continue from there. Since most pages are transcluded just once, I thought it might be easy (but I may be wrong) for the PP extension to check the "What links here" links, and if there is only one transclusion, put that link somewhere on the Page: page, so you can link from main to page using the page number link, and from page to main using that link. Just a thought, ignore it if is technical not feasible.

Thanks again for this great tool! − Inductiveloadtalk/contribs 02:20, 30 April 2010 (UTC)

thank you for the suggestions. I already thought about "reverse transclusion", but the non-uniqueness makes it a bit difficult. ThomasV (talk) 05:46, 1 May 2010 (UTC)

Bot flagsEdit

I am contacting everyone who operates an active bot on en.WS. When granting bot-flags I am heavily reliant on the consensus of people like you to ensure that prospective bots approriate and should be flagged. I am not at all competant to evaluate bots independantly and this leads to people waiting a rather long time to find out if their bots will be santioned. Could you please help me out and take the time to examine one of the two current requests at the Scriptorium?[4] [5] Thank you.--BirgitteSB 20:32, 11 May 2010 (UTC)

Speeding up custom toolbar loadEdit

Hi. Is there a way to speed up loading my Vector.js custom toolbar when I editing a page? I still have to use the header/footer option to access the toolbar and the process slows editing. - Ineuw (talk) 23:13, 11 May 2010 (UTC)

Hi. Would you have any suggestions to access the custom toolbar without the header/footer open and speed up the process as well? - Ineuw (talk) 16:28, 23 May 2010 (UTC)

Hi. Sorry, but I have no suggestion except to wait. The thing is, I made some corrections to the code shortly after the last code update, in order to fix a few issues. However, nobody was kind enough to review and deploy my updates, and I got tired of asking. ThomasV (talk) 17:09, 23 May 2010 (UTC)

Thanks for your quick response. I would gladly help in testing anything, (which is a welcome break for me), if only you would direct me to what you wish to be done. I did see the earlier message you mentioned, but I am clueless (so far), what to test. - Ineuw (talk) 20:02, 23 May 2010 (UTC)

only a few developers are allowed to review and deploy software updates; this is why it is so slow. if you want to test things, you need to install the svn version of mediawiki on your machine. ThomasV (talk) 20:33, 23 May 2010 (UTC)

That's no problem. I have the following standalone Wiki software with numerous extensions and installed. It's not accessible from the web and is just for my curiosity and learning.

MediaWiki 1.16wmf4 (r66620), PHP 5.2.4-2ubuntu5.7wm1 (apache2handler), MySQL 4.0.40-wikimedia-log

I can upgrade and add any extensions needed. I can only test as a user and report to you. It serves at the same speed as the Wikisource, which means the comparison is realistic. - Ineuw (talk) 23:25, 23 May 2010 (UTC)

you'll need to upgrate to the svn version of mediawiki, and to install Webstore and ProofreadPage ThomasV (talk) 17:51, 25 May 2010 (UTC)

Will do so and let you know if I have any problems. :-) - Ineuw (talk) 01:16, 26 May 2010 (UTC)

To me it has not been specifically the custom part of the toolbar as the Page: tools load really quickly, it is the standard and monobook.js components that have been slow to load, now not sure whether that is an order of presentation or not, however, it would also seem that they are not readily being pulled from a cache. — billinghurst sDrewth 03:34, 26 May 2010 (UTC)

Hi Thomas. I admit (a temporary) defeat. I made a completely new install of 1.15.4, TortoiseSVN and the Wiki works fine. Think that I must install v1.16 for the features mentioned. I don't want to bother you about what's missing, and I would rather spend the time improving my work here, even if the custom tools load slow. I apologize for wasting your time. :-) - Ineuw (talk) 16:25, 3 June 2010 (UTC)

Helpful Firefox 3.6 speed tweaks foundEdit

Hi Thomas, I came across this website and I thought that it may be of some use to those who have a similar problem and of interest to you. [6]. Being cautious, I've been using it for 3 days just to be sure. The pages load much quicker and there is a discernible speed increase of my vector.js toolbar but still not like it used to be. - Ineuw (talk) 15:49, 11 June 2010 (UTC)

Mediawiki SVN installation errorsEdit

Hi Thomas. I didn't forget about the SVN install, but realized that attending to the PSM authors' list was crucial and it took precedence over everything. I made the first attempt of the SVN version of the Wikimedia software, which generated a bunch of errors which I posted here User:Ineuw/Error messages. Please don't spend much time on the issue because the errors will be resolved as they are all caused by my ignorance. FYI, I installed: TortoiseSVN- - mediawiki-1.15.4 - extensions-nightly-r67036, and modified the LocalSettings.php to include your instructions. Have a nice day. - Ineuw (talk) 14:17, 31 May 2010 (UTC)

problem with tables spanning multiple pages tagsEdit

Hi Thomas,

There seems to be a problem with rendering of tables that span pages in different <pages> tags. An example can be seen at Page:William Blake, a critical essay (Swinburne).djvu/158 and Page:William Blake, a critical essay (Swinburne).djvu/159. The first of these starts a table, but noincludes the closing brace. The second finishes the table, but noincludes the opening brace. When the pages are transcluded together, the two tables ought to merge into one. That is what happens if they are transcluded as

{{page|William Blake, a critical essay (Swinburne).djvu/158|num=142|section=text}}
{{page|William Blake, a critical essay (Swinburne).djvu/159|num=143}}

and it is what happens if they are transcluded as

<pages index="William Blake, a critical essay (Swinburne).djvu" from=158 to=159 fromsection="text" />

but rendering is broken when the pages are transcluded as

<pages index="William Blake, a critical essay (Swinburne).djvu" from=158 to=158 fromsection="text" tosection="text" />
<pages index="William Blake, a critical essay (Swinburne).djvu" from=159 to=159/>

I think I've managed to work around the problem in this case, but the problem was very difficult to diagnose, and I'm concerned that there may be no solution in cases where a table (or tables) spans multiple pages all of which use labelled section transclusion.

(I suspect you're going to tell me that you have no control over this, as your extension cannot force table rendering to be deferred until after the extension has done its bit.)

Hesperian 01:44, 24 June 2010 (UTC)

Test to be seen at sDrewth 03:09, 24 June 2010 (UTC)
fwiw... I mentioned this problem in one of Billinghurst's works where it seemed the over-abundance of opening and closing div tags in use seemed to cause swaths of text to "disappear" (got cut off after the offending div tag) or cause phantom spaces/linebreaks upon transclusion. For me, usually, placing the page-split table within div tags or around the specific table-row fixes the extra space(s) and line break problem(s) when a table spans 2 or more consecutive pages but when I tried my brand of fix in the far above given example - it just cut off a section of text instead of including it. George Orwell III (talk) 03:46, 24 June 2010 (UTC)
That is a different problem. Hesperian 03:51, 24 June 2010 (UTC)
why would one want to call multiple instances of <pages/> like that ? ThomasV (talk) 05:16, 24 June 2010 (UTC)
There are multiple overflowing notes. Cygnis insignis (talk) 05:37, 24 June 2010 (UTC)
Sorry, I assumed you were aware that overflowing footnote solution. I have used it quite a bit, this is the only problematic implementation. Thanks for having a look at the problem. Regards, Cygnis insignis (talk) 06:59, 24 June 2010 (UTC)
[edit conflict] Because the only way to invoke labelled section transclusion of a page is to use the fromsection or tosection attributes, which requires that the page be the first or last in the sequence. So if you want to invoke labelled section transclusion on an intermediate page, you have to split the sequence across multiple <pages> tags.

I suppose your next question will be: why would one want to invoke labelled section transclusion on an intermediate page? The answer is that the best known solution to the problem of how to mark up footnotes that overflow onto the next page uses labelled section transclusion on the overflow page. This particular work has numerous long footnotes that overflow onto following pages, and therefore William Blake, a critical essay/Lyrical poems contains no fewer than nine consecutive <pages> tags. Hesperian 05:45, 24 June 2010 (UTC)

I see. I understand your problem. I have no solution for the moment. However :

  • You should not try to combine multiple instances of <pages/>. It is not supposed to work that way. The page where the text is transcluded is not supposed to know that there are multiple sections, in which order, etc. And it should not worry about that. In object oriented programming, this is called "encapsulation". This is very important, because in the future we might want to display our texts elsewhere that on our wiki pages.
  • The technical solution you are deploying is very complicated. That's bad. It is too complicated for the average user. I am against such a level of complexity, because the problems you are creating are worse than the problem you are trying to solve. Some users will just imitate what you are doing without understanding it.
  • In the absence of a satisfying solution, I would suggest not to be anal about overflowing footnotes. It will hurt nobody if the whole footnote is on a single page. The scan will still be one click away ; I guess we can live with that.
  • Tables are evil. You don't need a table here. You could achieve the same result with the poem tag and style attributes.(I hope you will not use this advice to solve your problem and keep using that overflowing footnote.)

ThomasV (talk) 07:19, 24 June 2010 (UTC)

I note the NO DUPLICATE <PAGES>, so if that is the case, then we need the ability to exclude pages. Where we have images, often in these older books on the reverse there is a blank page or one of those tissue overlays, with that we usually wish to not transclude as it overlays the page number or symbol for the blank page. — billinghurst sDrewth 07:57, 24 June 2010 (UTC)
but this is something you can put in a noinclude section, isn't it ? ThomasV (talk) 08:00, 24 June 2010 (UTC)
No, I don't believe so. How can we noinclude a whole page, or more to the point, the page indicator on the left into it? — billinghurst sDrewth 09:23, 24 June 2010 (UTC)
you can leave the whole page blank, or put everything inside a noinclude section. Of course you'll still have the page indicator on the left. If you really do not want this indicator, then use multiple "page" tags; it's not a problem as long as you do not combine them. But I agree that it would probably be better to have the ability to exclude pages. ThomasV (talk) 09:47, 24 June 2010 (UTC)

Notwithstanding your opposition to this particular use of intermediate page transclusion, there are other compelling examples. Like some magazines and newspapers that will start an article on an early page, but kick the rest of it off to some much later block of pages. For example, it might start on page 5, but pick up halfway down page 20, cover all of page 21 and finish on page 22. I envisage a pages tag with a syntax much like the pagelist tag:

<pages from=5 to=22 5=examplearticle 6to19=exclude 20=examplearticle 22=examplearticle/>

Hesperian 00:56, 25 June 2010 (UTC)

Example: Hesperian 01:17, 25 June 2010 (UTC)
In these cases, I see that this is still the strength of {{Page}} and adaptive use of #lst, as there is still capabilities of #lst that are not within <pages>. They form <<1% of our cases, and can and should be used in those (unusual) circumstances. I don't have the expectation of over-engineering <pages> to be a one stop shop. — billinghurst sDrewth 01:41, 25 June 2010 (UTC)
maybe I wasn't clear: I have no opposition to using multiple "page" tags if they are independent from each other. What we want to avoid is situations where you open a div in the first call and close it in the second one. ThomasV (talk) 05:39, 25 June 2010 (UTC)
How then do you propose to handle the case of a block level format that spans a non-continuous page jump? Suppose an article starts on page 5, continues on page 20, and the page break happens in the middle of a long quotation that is rendered in smaller text. It seems like the only option is not to use the pages tag at all. Hesperian 06:35, 25 June 2010 (UTC)
I already said above that I have no simple solution for this, at least for the moment ; however, for a footnote I think that a complicated solution based on sections is not worth it. For an article it seems that we have no options but section transclusion. ThomasV (talk) 06:55, 25 June 2010 (UTC)
Okay, thanks. I'll leave it with you. Hesperian 07:06, 25 June 2010 (UTC)

In reply to your comment about tables, I agree that tables are evil, but currently there is no way to centre a left-aligned block of text using CSS. This is a well-known and off-complained-about weakness of CSS. We are taking our dedication to the "tables are evil" mantra too far if it forces us to overlook the only means of duplicating a given layout. Content comes first; adhering to coding principles second. (Where possible I would use the {{float center}} template for this, but templates cannot span page boundaries. Hesperian 15:43, 24 June 2010 (UTC)

How about this :

'Knowest thou not, fairies' lord,
How much by us contemned, abhorred,
Whatever hides the female form
That cannot bear the mortal storm?

ThomasV (talk) 05:39, 25 June 2010 (UTC)
You cheated: you hard-coded the width. If the width is too small the text will wrap unnecessarily; if the width is too big it will not be centred. (And the width depends on the font in which it is rendered, which is browser-dependent.) Hesperian 06:35, 25 June 2010 (UTC)
and if the width is not fixed your text might wrap as well, because you do not control the size of the container div, soverses might be split on 2 lines. But if you do not care about that, you can use "width:50%" in the above : it will work too. ThomasV (talk) 06:55, 25 June 2010 (UTC)

Oh, and in response to your edit, you need to use {{#tag:ref}} magic word instead of the <ref> tag, in order to force processing of the <noinclude> tag. I learned/adapted this from Wikipedia:WP:REFNEST. Hesperian 15:47, 24 June 2010 (UTC)

Excerpt or mixtureEdit

Gday ThomasV. Trying to understand why the PROGRESS field has the choice of excerpt/mixture, as I would have thought that this would be an option in the TYPE field. Where we do have excerpts (individual or sequential articles), it makes it difficult to move the progress marker to validated. Would you mind explaining your thinking behind the existing design. Thanks. — billinghurst sDrewth 04:20, 2 July 2010 (UTC)

Hi sDrewth,
About PROGRESS : if a djvu file is an excerpt of a book, it is much better to create a djvu for the whole book (or journal issue), and to move the pages there. Even if you are not interested in the rest of the book, someone else, in the future, might want to publish the whole book, or another part of it. At that point they will need to move the pages you created to the new djvu index, and to update various links and transclusions. So, it creates less work for the community if we can anticipate and do this right from the beginning. Of course, this is not an obligation : if the whole book or journal issue is not available to you, then do not care about it and go ahead with what you have. The purpose of this classification is to minimize the global amount of work, not to bother you.
Concerning the TYPE field, it means "type of book". It is used to manage metadata and to generate citation information in headers, and it works like LaTeX book types. Unfortunately, Proofreadpage headers are not very much used at en.wikisource, so you cannot really see what I mean unless you have a look at an example. Compare those two pages fr:La Bécasse and fr:Les dieux ont soif/Chapitre I, and have a look at the corresponding index pages. In the first case, the title in the header is the title of the novel, because the book is a "recueil" (which means "collection" in English), that contains many short novels with different titles. The third line (citation information) indicates the title of the "recueil". In the second example, the book is a long novel with many chapters. The title in the header is the title of the book, and not "Chapter I". The third line does not duplicate the title, and starts with the publisher.
I am sorry but I did not find time to adapt this to the main problem is the layout problem that I told you about last year.
ThomasV (talk) 09:21, 2 July 2010 (UTC)
Thanks for that, I do work on the premise of bring the whole work, rather than snippets where that is possible. The words don't clarify the aspects of snippets, and to this point my reflections haven't identified a better means, and I need to go and dig deeper, and think more. I don't remember more than the basic basic detail of our earlier conversation. :-/

Bugzilla 21526 — ") breaking pagesEdit

Hi ThomasV,

The conversation at WS:S about bugzilla:21526 took me to #wikimedia-tech where TimStarling said that the code was annotated as being written by you, so I am presuming that it is ProofreadPage. Anyway, he asked that I assign the bug to you, which I did, and subsequently I have added this to ProofreadPage talk page. I know that I ask a lot, though I would like to see this one addressed as it is having an effect through a number of works for which are stalled due to the issue. Thanks. — billinghurst sDrewth 01:53, 6 July 2010 (UTC)

Thanks for your work to coordinate the fix. I presume that we now have to wait for a code update. — billinghurst sDrewth 13:44, 7 July 2010 (UTC)
yes, we need to wait for a code update. there are a couple of other fixes waiting to be deployed. ThomasV (talk) 13:58, 7 July 2010 (UTC)

Page edit speedEdit

Hi ThomasV. Thanks for the improvement. I've been trying it for the past couple of hours and it really makes a difference. Two questions please:

  1. In your Scriptorium post, what did you mean by slow speed? I do have high speed internet (~5MB). Although, I do work in 1024 x 768 x 85MHz mode on a CRT. Is this what you were referring to?
  2. Is the Header/Footer must open when editing? I am interested because it wasn't so before and it reduces my editing box area and it seems to be connected with the slowing down of the custom toolbar's appearance. Just curious. - Ineuw (talk) 02:22, 20 July 2010 (UTC)
  • there are multiple factors that can effect load speed; I don't really know about your system.
  • keep in mind that there is still a bugfix waiting to be deployed ; it will allow you to start editing right away, before the full image is loaded

ThomasV (talk) 04:51, 20 July 2010 (UTC)

I am content with the current changes and await the bugfix deployment. Thanks for the info. - Ineuw (talk) 13:09, 20 July 2010 (UTC)

Page numbering renderingEdit

Looking at The Elizabethan People/Chapter 1 at the bottom, the javascript is interpreting alt-0150 from the Index page numbering as the utf-8 value, or something like it. It would be great if you would be able to cast your eye over it. Thanks. — billinghurst sDrewth 14:54, 8 August 2010 (UTC)

hello sDrewth
thanks for reporting it; I fixed the rendering of "-". Another issue is that there are two page numbers overlapping there, but that was also the case with the old rendering technique.
ThomasV (talk) 20:05, 8 August 2010 (UTC)

I had started using parentheses in the Index form to indicate full page illustrations, (25) for 'illustration facing page 25', but this now displays as .2825.29 I fiddled a bit, but I wonder whether improvements have not incorporated an earlier function that resolved the rendering of non-alpha characters. Perhaps it is now affected by that geturl variation you helped me with. 01:22, 8 August 2010 (UTC)

I guess this is the same issue as User talk:ThomasV#Page numbering rendering. If this is still broken for you, you should take it there. Hesperian 23:26, 8 August 2010 (UTC)

So I did. :-) Cygnis insignis (talk) 04:39, 9 August 2010 (UTC)

if only you could provide a link... ThomasV (talk) 05:06, 9 August 2010 (UTC)
An example is Indian_Fairy_Tales_%28Jacobs%29/Boy_with_Moon_on_Forehead#.28165.29, which is a link on the title page of Indian Fairy Tales (Jacobs). Cygnis insignis (talk) 05:35, 9 August 2010 (UTC)
it works for me ; I guess it was fixed along sdrewth's problem; please check that your javascript is up-to-date ThomasV (talk) 06:09, 9 August 2010 (UTC)
Thanks for the fix ThomasV, works fine. I was aware of the blank page, it was next on the fix. Cygnis, I don't see a problem with your pages (at this time) and I checked some of your contributions. — billinghurst sDrewth 06:19, 9 August 2010 (UTC)
Thank you both, the label displays correctly now. Cygnis insignis (talk) 07:28, 9 August 2010 (UTC)

Sidenotes broken by recent <page> fixEdit

Could you take a look at Wikisource:Scriptorium#Transclusion_breaks_sidenotes? It appears that the fix you put in place for the "searching across page breaks" problem is causing another issue that affects many of our works. —Spangineerwp (háblame) 18:26, 9 August 2010 (UTC)


Sounds like a great idea. (I saw it on the mailing list, but prefer to lurk there without contributing, so will comment here.)

I had a look at the patch. It looks good to me.

  1. The documentation on line 39 (of the patch) probably should be updated.
  2. Overflowing refs could be concatenated with a space or a carriage return. I see you've chosen to use a space. I suppose the choice could have implications for tables. Have you thought about this?
  3. $key either takes a value or null, and is later tested with 'if($key === null)'. $follow either takes a value or false, and is later tested with 'if($follow)'. Is there any reason for the inconsistency? For consistency's sake, I would have set $follow to null rather than false, and tested for 'if($follow !== null)'.
  4. I recently encountered a reference that overflowed onto more than one subsequent page. Have you tested against this case?
  5. Ages ago I encountered a page that contained its own reference, a reference overflowing from a previous page, and a reference overflowing from two previous pages. Have you tested for that? I suspect your patch would put the first overflow at the beginning, and then the second overflow at the beginning, resulting in the second overflow wrongly preceding the first.
  6. It is a pity these names have to be global. It would be tempting to write <ref name="overflow"> and <ref follow="overflow"> but that will be a disaster when transcluding chapters containing many overflowing refs. People are going to have to come up with a naming scheme like <ref name="overflow_page_14"> and <ref follow="overflow_page_14"> in order to protect their overflowing references from each other.

Hesperian 23:50, 11 August 2010 (UTC)

Thanks for the comments. Actually, posting your comments to wikitech-l would not hurt ; they are not offtopic, and it would demonstrate that there is some community support for the proposed change. Answers :
  1. yes, you are right. actually, in a previous version of the patch I did this, and I modified the return statement so that it returns array(false, false, false) ; then I preferred to add the FIXME statement to the code, because it should be array(false, false); this is how I reverted to the unchanged documentation
  2. no, I have not thought about tables. can you give me an example ?
  3. you are probably right ; I'll have to check.
  4. of course I did :-)
  5. no I have not ; it sounds like it would be interesting to test. do you remember the url of that page ?
  6. yes, although names like "p243" are short enough to be usable. The other option I considered was to rely on the parser's processing order, and to append a ref with "follow" to the previous declaration of <ref> that had been encountered ; but I it seems a bad idea to make such assumptions on the parser.
ThomasV (talk) 05:34, 12 August 2010 (UTC)

I saw it on the wikisource mailing list, not wikitech. I see your point about expressing community support; I will have a think about posting this there. Meanwhile feel free to claim my support, and a tentative patch review.
Can't be bothered subscribing and unsubscribing just to post a single message. Hesperian 06:21, 12 August 2010 (UTC)
2. I don't think so. The most common case is where a table has been used to center a block of text, e.g. Page:William Blake, a critical essay (Swinburne).djvu/191 and Page:William Blake, a critical essay (Swinburne).djvu/192. But, leaving aside objections to using a table at all in this context, this is not very interesting because a new row doesn't begin at the page break. What I'm interested in is (a) whether concatenating the refs with a space will break table syntax by causing the "|-" no longer to occur at the start of a line; (b) whether concatenating them with a carriage return will fix that problem; and (c) whether concatenating them with a carriage return might cause other, unanticipated, problems.
5. No but I will have a look around and see if I can find it.
6. Yes, I agree.
Hesperian 05:54, 12 August 2010 (UTC)
Re: 5., I can't find my real example, and I think it is an important test case, so I made you a test: Index:Overflow test case.djvu. Hesperian 06:36, 12 August 2010 (UTC)
Thanks for the example, I now understand. I think it would handle it correctly, except for one rendering detail : on the 3rd page, the two overflowing footnotes are rendered using spans instead of "li" elements (in order to avoid numbering) ; this implies that there will be no carriage return to separate them. I guess this can be fixed, I'll have a look. ThomasV (talk) 06:48, 12 August 2010 (UTC)
I see. My concern was that on the last page the first overflow would be processed and put at the top, and then the second overflow would be processed and put at the top, in which case the second would end up above the thirdfirst. Hesperian 07:01, 12 August 2010 (UTC)
that too ; I just tested. ThomasV (talk) 08:47, 12 August 2010 (UTC)
Not understanding the supertech bits, however, re case (6). As for our situation we are working within a djvu file where the name doesn't change and we have them all as subpages; plus I would suspect that there is only going to be one continuing footnote per page (are there examples of this not being the case?); for engineering purposes, do we just need to note the page of the continuing text? As long as the page(s) is known, and the tagging is appropriate to identify the continuing text can we look to simplify? — billinghurst sDrewth
It is almost impossible to have two overflowing footnotes originating on the same page, because the footnote area will grow, encroaching upwards into the main text area, until the footnote citation is on the last line, and only then, when it cannot encroach any further upwards without pushing its own citation onto the next page, will it overflow. Thus the only way to have two overflowing footnotes originating on the same page is if two footnote citations occur on the same line, and the first is big enough to trigger overflow. For much the same reason I think it is also almost always true that it is only the last footnote on a page that has the potential to overflow. So I would agree that providing the originating page number unambiguously identifies the footnote to follow on from. But still I share Thomas's concern about making too many assumptions. It think it better to rely on named parameters, even if means we have to be a bit careful. Hesperian 12:23, 12 August 2010 (UTC)

Quirky page numberingEdit

Gday. Something quirky happens with with <div class=lefttext> and {{float center}} (a centred-table) as per this edit. Notice that the first page number seems to have disappeared (or maybe that is that I didn't locate it) and the second page number disappearing off to the left. Presumably it would be fixed by an inner/outer divs, however, I just thought it weird that the page number disappeared and worth mentioning. 02:44, 16 August 2010 (UTC)

yes, but this template is evil ; why use a table to center things ? (btw page numbers are there, but in the margin) ThomasV (talk) 06:10, 16 August 2010 (UTC)
more generally, I think that the choice of centering the text, of the margin, etc, should remain with the user's preferences, and not be hard coded in the pages. you can write a gadget that changes the proprties of a class ThomasV (talk) 06:13, 16 August 2010 (UTC)
I reckon people wanted the prose class because they saw the float center format on the page. This is what most quotes use when mixed with the regular formatting - the 'necessary evil' should be in the Page: ns, not a simple way of gathering a jumble of formatting instructions in mainspace. I think users wanted something like the current display, I changed it in Page and here, note the previous diff too. I didn't need a class until the recent changes (by others too) if the pages contain only centred text, it seems to need indented page now. Not really a problem, all the other changes seem pretty good. cygnis insignis 07:39, 16 August 2010 (UTC)
(ec)I wasn't championing the template, as I know that it has weaknesses, I was being lazy and not wishing to rejig all the pages on the work. I only saw the page number for the second page, not the first page [57]. A reason to use the table to centre is that it allows the left margin to align, and still allow the block of text to centre on the page (rather than the lines to centre). Mate, you might be able to write gadgets to do those things, I know that it isn't my skill set, and asking for that in a more general sense, well, I have given up on that task. — billinghurst sDrewth 07:46, 16 August 2010 (UTC)
It is fairly easy to write CSS gadgets, but for them to work you need to standardize the name of the class you use. At we have deprecated the "lefttext" and "centertext" classes in favor on one single class called "text". The two other classes are still there, but they are now identical to "text". Users can decide in their preferences whether the text should be centered or on the left, using this gadget : fr:MediaWiki:Gadget-LeftText.css. As you can see, things would have been easier if we had a single class right from the beginning. ThomasV (talk) 07:59, 16 August 2010 (UTC)
Er ... yes! Indented class was needed, the others are 1. User's applying their own preference over others, or 2. attempting to replicate float center, or 3. reproducing a meaningless feature of a printed page, or 4. creating a new page within a wikimedia page. All this is evil, from the pov of community harmony; deprecating it is a good idea. Wikipedias get along fine with out this, a pointless distraction that it seems you are expected to fix when it all goes wrong. cygnis insignis 08:26, 16 August 2010 (UTC)
Actually, I have no expectations of a fix, and I had addressed the matter by a different means. I simply saw a situation where the page numbering went weird, and brought it to the attention of the developer as per the original statement. To lefttext, yes it has been over-used for many pages, though at the same time, there are situations where it can be deemed appropriate for use. To all the stuff about text css, etc., I might fiddle at the edges, or create minor cases, I am not and have never seen myself as a css expert and I simply don't play in that space. I am of the wrong generation for hacker & html school. — billinghurst sDrewth 12:02, 16 August 2010 (UTC)

[in reply to "this template is evil"] It shouldn't be called "float center"; it should be called something like "centre block", as its purpose is to centre a block of text without affecting text alignment. "float center" implies that we wish inline text would flow around it, which is not the case.

I'm as frustrated as anyone that we have to resort to am embarrassing table hack to implement this. But I think it is a mistake to pretend that we don't need something just because we can't implement it elegantly. Treat the template as a black box: ignore what you know of the implementation, and judge it solely on what it does. In my experience, the ability to centre a block of left-aligned text is absolutely essential.

Hesperian 11:25, 16 August 2010 (UTC)

I am not sure if we are really talking about the same thing. I am not saying it is bad to centre a block of text. I am saying that you should not be doing it with a table.

<div style="margin:0px auto;width:40em;"> <pages index=Mandragora.djvu from=73 to=74 /> </div>

It this what you want to achieve ? ThomasV (talk) 12:46, 16 August 2010 (UTC)

No. You've fixed the width at 40em. Long lines will wrap unnecessarily. If 40em is more than you need, it won't be accurately centred. A crappy implementation that works is better than an elegant implementation that doesn't. Hesperian 13:12, 16 August 2010 (UTC)
And by the way the pages you're using as an example are already centred using a table. It looks the same without that div. Hesperian 13:15, 16 August 2010 (UTC)
Please read the topic of the discussion : we are talking about page numbers. Without the div, the page numbers are not at the right place. ThomasV (talk) 13:29, 16 August 2010 (UTC)
On my browser the page numbers appear in exactly the same place whether the div block is there or not. Hesperian 13:33, 16 August 2010 (UTC)
could you be constructive enough to tell us what your browser is, and where the numbers appear ? ThomasV (talk) 13:38, 16 August 2010 (UTC)
Just to clarify what i get, on mac osx/ camino, safari, firefox: On my browsers page numbers appear in the sidebar, since two days ago, when there is no class. The poem above has the numbers away from the edge of the page, the same as prose class and so on. Indented-page produces the standard display, as always. cygnis insignis 13:42, 16 August 2010 (UTC)
to me too. it looks fine. I think it's ok to use the table centering template inside a page ; the problem is that it was called twice, once inside and once outside. ThomasV (talk) 14:08, 16 August 2010 (UTC)

page numbers not appearingEdit

An odd thing: page numbers are rendered correctly at Makers of British botany/Introduction, but are missing at Makers of British botany/Robert Morison 1620—1683 and John Ray 1627—1705. Both pages use the {{page}} template for transclusion. I can't make sense of it. Firefox 3.6.8 on Windows XP. Hesperian 00:07, 20 August 2010 (UTC)

ok, it should be fixed now ThomasV (talk) 01:11, 20 August 2010 (UTC)
Yep. Thanks. Hesperian 01:33, 20 August 2010 (UTC)

<pages> and unwanted line breaksEdit

Sorry if you've already addressed how to handle this, but could you look at Kelo v. New London/Dissent Thomas, at the joining of pages 506 and 507? This is rendering improperly; the text at the top of page 507 is a continuation of the quote on page 506. If I use the page template, I can avoid this—see pages 6 and 7 of A Short History of Standing Armies in England. Same code in the page namespace in each case to create the indent (just a colon at the beginning of a new paragraph). Is there an easy way to address this (like a template)? Or is there a way to find out what html the colon creates (p? div? span? and how much indentation?) and use that instead? Thanks! —Spangineerwp (háblame) 16:25, 26 August 2010 (UTC)

not easily. the colon creates a combination of dl and dd tags (look at the code), sou you could combine them with noincludes in order to reproduce the indentation. ThomasV (talk) 05:19, 27 August 2010 (UTC)
You could look at using {{hanging indent inherit}} and its partner {{div end}}, rather than the colon to indent. One the first page, the second template goes into the footer, and on the subsequent page add the former template to the header. The values would be something like {{hii|2|0}}. I have used it in a few places to manage that stuff neatly. — billinghurst sDrewth 08:07, 27 August 2010 (UTC)
Thanks; got it working =). —Spangineerwp (háblame) 14:49, 29 August 2010 (UTC)

Use javascript to turn annotations on and off?Edit

Have you given any thought to using javascript to turn wikisource user-created annotations on and off in the main namespace? I've experimented with dividing "real" (straight from the text) references from user-created ones (see here) using the "group" function of the references tag. If javascript could be used to hide by default what I call the "supplemental" notes, this would allow our pages to conform more truly to the text and provide additional information if the reader desires it. Does this sound feasible? —Spangineerwp (háblame) 15:00, 29 August 2010 (UTC)

User-created annotation are a pain, and I think they should not be allowed. How are we supposed to validate pages if users are allowed to add annotations ? ThomasV (talk) 15:15, 29 August 2010 (UTC)
I disagree that user annotations should be disallowed, with some books with obvious errors, or with later agreeing or disputing evidence they can be quite useful. They should very clearly be as <ref> and separated from footnotes. I predominantly encourage use of {{user annotation}} to clearly demonstrate what they are and this should allow validation of a page. — billinghurst sDrewth 15:23, 29 August 2010 (UTC)
(ec) What Billinghurst said. For example, in the work I linked above, the author quotes the 12th article of the Bill of Rights but calls it the 10th, and he describes how Senators are selected—which could be confusing, because what he describes hasn't been true for almost 100 years. If javascript was in place in both the page and main namespaces, the desires to be precisely true to the text and yet helpful to readers could be combined. —Spangineerwp (háblame) 15:45, 29 August 2010 (UTC)
Pardon me, but I think this is being looked at inside-out. It hasn't been true for 100 years because it is a old book?! A copy of a book should be the same as those in the holdings of other libraries. This is a key principal, they would not change their copy. The text is what it is, preserving that is critical, of paramount importance to us, improvements and enhancements should happen elsewhere. Wikisource contributors are not authorities, or editors, that happens elsewhere. I've posed a number of questions on this matter, the answer generally being 'it's useful', but Thomas' is unanswerable. A validated text is often locked, the authority for 'obvious' and 'useful' annotations, links, sic templates and so on could only be changed by an admin, or lobbying one for change. For the record, admins here frequently make erroneous 'corrections'. If our readers are so inept, illiterate, or incapable of processing a typo, do we want them adding what they reckon to the documents. Let's get new works here instead, completely unobjectionable. Or create a text here and link, update and correct it at the other place, if we are wrong someone will catch that. cygnis insignis 20:43, 29 August 2010 (UTC)
The reason I brought this up in the first place is because I know that the use of annotations is a longstanding debate. With javascript, I'm envisioning that it's possible to add a link in the sidebar that allows users, if they desire, to look at annotations. By default, they wouldn't. For you (Thomas and Cygnis), this would be an improvement, because by default readers would no longer see annotations on any works created by me or people who agree with me (that there is a place for annotations on WS). From your point of view your long-term goal should obviously be to convince everyone that annotations don't belong on WS at all, but in the meantime, this idea may enable both sides to be happier than they are now.
The point about validated pages being locked is a good one, but lately locking pages has been rare, and usually the page namespace hasn't been locked above the autoconfirmed level. —Spangineerwp (háblame) 00:04, 30 August 2010 (UTC)
it all depends on where you set the limit. I agree that it makes sense to fix obvious errors, and to do this in a way that we can keep track of the corrections that have been done. However, it looks like annotations at are currently going well beyond the correction of obvious errors : some users add notes and insert wikilinks in texts, with no respect for the work of those who have faithfully corrected it. Is there a consensus that all types of annotation should be allowed ? A clear policy should be defined first. ThomasV (talk) 14:19, 1 September 2010 (UTC)
I've undone too many 'obvious errors' that were not errors, judgement is pretty poor in this area. Users changing words to ones they recognise, changing every page number because they were looking at a different edition. An actual error is sometimes significant, and no one can anticipate whether that is important - we shouldn't be judging that, not our role here.

The only explicit guideline is at Wikisource:Annotations, it requires the User is to make a separate version, leaving a 'clean text', but these changes are made to the Page:namespace now - difficult. This is the solution: The corrected Pages and their Index here could be duplicated at Wikibooks, they can review the annotations then publish them in their main-space. Readers who want this will have it one click away, producing the same outcome as this proposal at what I think is the appropriate sister. cygnis insignis 15:25, 1 September 2010 (UTC)

Highlighting and multi-column textEdit

Hi Thomas,

Multi-column formatting breaks page highlighting; see History of botany (1530–1860)/Book 1/Chapter 3 (pages 119 and 120) for an example. Highlighting is only broken within the multi-column block; it doesn't break it for the rest of the page. So I'm quite happy for you to declare this a WONTFIX. But I thought it couldn't hurt to mention it anyhow.

Cheers, Hesperian 13:08, 5 September 2010 (UTC)


Please confirm that the current state of the page is what you intended. cygnis insignis 17:49, 14 September 2010 (UTC)

no, it was in common.js too ThomasV (talk) 18:40, 14 September 2010 (UTC)
Er, yes, my bad. Done now, cygnis insignis 18:53, 14 September 2010 (UTC)

Without textEdit

I saw your message on the mailing list. I've started going through the "without text" pages, finding those that really ought to be transcluded (for example because they contain an image), and promoting them to some other status. Have done the first 1000 (out of 16800). Leave it with me. Hesperian 02:28, 17 September 2010 (UTC)

did you really read my message ? I also proposed to transclude page content without page number and without new line, which would allow us to float images without breaking the text flow… ThomasV (talk) 06:32, 17 September 2010 (UTC)
Yes I read it. I didn't really understand what you meant by that, and I still don't. I am under the impression, though, that the new definition of "without text" implies that pages will not be transcluded if they are marked "without text". Hesperian 09:53, 17 September 2010 (UTC)
this was only the first of two options. its only advantage seems to be simplicity; but I think that we should go for option 2 (transclusion w/o trailing newline), precisely because of images. ThomasV (talk) 10:46, 17 September 2010 (UTC)

wrappers around pagesEdit

Thanks, I won't do it again. I just copied it from somewhere else and modified it "Monkey sees monkey does". -- Philip Baird Shearer (talk) 06:12, 21 September 2010 (UTC)

I have a problem with 1911 Encyclopædia Britannica/Fives do you have any idea why the section at the top of page 451 is not displaying? -- Philip Baird Shearer (talk) 11:53, 21 September 2010 (UTC)
the section name was not correct. I fixed it. I also replaced transclusions with the pages tag ThomasV (talk) 11:56, 21 September 2010 (UTC)
Thanks again. -- Philip Baird Shearer (talk) 20:17, 21 September 2010 (UTC)

Issue with USSC templatesEdit

If you get a chance, could you take a look at {{USSCcase}} and {{USSCcase2}}, as they appear on Kelo v. New London and Kelo v. New London/Opinion of the Court? The templates, when combined with page transclusions, are not functional—while they appear, the links don't work. USSC cases are an area where the proofread page extension has not penetrated, and I'm reluctant to push it further with such browser issues. —Spangineerwp (háblame) 14:37, 6 October 2010 (UTC)

fixed here. the problem was not the template itself, it was the fact that you were wrapping the "pages" command in an extra div. see Help:Layout ThomasV (talk) 17:46, 6 October 2010 (UTC)
Thanks! Wasn't aware that that div was causing issues now; I guess it should be removed everywhere? —Spangineerwp (háblame) 18:13, 6 October 2010 (UTC)

deeplinks in dictionaryEdit

Re: the A Dictionary of Music and Musicians/Vol 1 model. Have you thought about how we can link entries. I'm not sure I described the problem very well, but there is some musing here. One use would be generating 'static' lists by contributor, on author pages, I would like to deeplink references at the sisters and here. Is this possible?

BTW: the new LST trick seems much better, it 'looks and feels right', and I haven't found any problems so far. cygnis insignis 11:23, 7 October 2010 (UTC)

like this ? ThomasV (talk) 10:29, 12 October 2010 (UTC) Ta. cygnis insignis 14:15, 12 October 2010 (UTC)

Page numbers alignmentEdit

Hi Thomas!

Could you help me with the problem I mentioned at Wikisource:Scriptorium#Page numbers alignment? Any suggestion? Helder (talk) 19:28, 8 October 2010 (UTC)


Any idea why the page numbers at A Portrait of the Artist as a Young Man are rendering as "[.E2.80.94]"? Hesperian 11:08, 18 October 2010 (UTC)

this is fixed ThomasV (talk) 11:57, 21 October 2010 (UTC)

Transclusion HelpEdit

Ran into difficulties when attempting to transclude [see: 18:40 edit on 20th.] from a table framework. Encountered two undesirable consequences when attempting to transclude these two sections to different articles. All passages from this page transcluded even when outside the labeled section. The second unintended situation involved the section identified as ## Goodwin, Harvey ## would shift down one line upon selection of "show preview" thereby breaking the table and misaligning that entry into the wrong column. JamAKiska (talk) 11:37, 21 October 2010 (UTC)
I simplified the table a bit ; try now ThomasV (talk) 12:17, 21 October 2010 (UTC)

This Revision as of 13:11, 21 October 2010 shows both errata sections visible with the sections clearly labelled. JamAKiska (talk) 13:20, 21 October 2010 (UTC)

I fixed this page for you. try to avoid constructs with "onlyinclude", and do not use "div class=indented-text". ThomasV (talk) 13:23, 21 October 2010 (UTC)

Thanks…already transferred to Hardinge article…JamAKiska (talk) 13:30, 21 October 2010 (UTC)

I also modified the javascript, so that the ## can be written right after the pipe symbol "|". This is needed for vertical alignment. Please reload your javascript ThomasV (talk) 13:35, 21 October 2010 (UTC)

Thanks…great refresher too…the format of that text is a bit of a challenge, good to see tables are flexible as well. I’ll reinsert the transcluded components. Wow! JamAKiska (talk) 14:40, 21 October 2010 (UTC)

Revision as of 17:41, 21 October 2010 This is a multi-line Errata entry with an iteresting interaction with the wikipedia entry. Time permitting… Thanks…JamAKiska (talk) 17:47, 21 October 2010 (UTC) Was able to use "includeonly" to relocate the challenging material of the 2nd line…and that adjustment provided a much better presentation. Thank-you…JamAKiska (talk) 18:30, 21 October 2010 (UTC)

Problems with the most popular browserEdit

I recently accessed a few works with IE 7 and IE 8, and was surprised to find out that page numbers do not appear on the left side of the screen for these browsers—meaning that potentially 50% of our readers don't even know that many of our texts are proofed against scanned texts. The only way an IE user can know this is if they click the "Source" tab at the top of the screen and attempt to decipher the Index page, both things a casual reader is unlikely to do.

I know WP/WS editors rarely use IE, and I know too that IE is despised by the developers at the software company I work for. But that doesn't change the fact that a large percentage of our readers use it. I hope there's something we can do to increase our browser compatibility, so that we can increase our readership and our impact. —Spangineerwp (háblame) 14:27, 21 October 2010 (UTC)

It's been like that for weeks here too. Same behavior. 10 days to a mid-term election and I'm slowly losing the inroads made to the online senior demographic and the ability to keep proving their whacky internet tube theories wrong. George Orwell III (talk) 18:50, 21 October 2010 (UTC)
I do not think this is due to the PageNumbers script. Page numbers are rendered well under IE at the french Wikisource. this is probably caused by a bug in this site’s local javascript. ThomasV (talk) 14:55, 21 October 2010 (UTC)
I found one bug in Common.js that causes IE to fail ; after fixing it, page numbers are sometimes displayed ; they show up after I reload the page with F5, but not if I access the page directly. This is not the case at : they are always rendered. This makes me think that there is another bug, but I did not find it.
In the future I plan to move this script into the core of the extension ; that way, it will not longer be vulnerable to bugs in a site’s javascript. ThomasV (talk) 15:43, 21 October 2010 (UTC)
Thanks for looking into this. Your fix seems to have addressed the problem in many places, but, for example, I'm looking at State_Documents_on_Federal_Relations and The Rise and Fall of Society in IE 7 and IE 8 and not seeing page numbers in either browser, even after refreshing. I'm not sure what the difference might be between these works and the ones that are showing up correctly. —Spangineerwp (háblame) 18:42, 21 October 2010 (UTC)
I think I found the problem: it's a purging issue; if the page has been saved/purged recently, the numbers appear in IE just fine. Thanks for your help! —Spangineerwp (háblame) 19:20, 21 October 2010 (UTC)
The problem seems to still occur in Monobook and IE 8 (no page numbers at all). I doubt there are many readers on IE and Monobook, but if it's an easy fix it may be worth it... —Spangineerwp (háblame) 19:29, 21 October 2010 (UTC)
Sure, the time when the change to fix broken search terms over page breaks in transclusion was implemented - the same thing happened. The page numbers came back, then they only came back with a [ctrl+] F5 refresh(es) and eventually stopped appearing altogether again no matter which ver. of IE or what comnination of steps under each I took, so I stopped caring. George Orwell III (talk) 19:34, 21 October 2010 (UTC)
nice to see that I am not the only one to observe this strange behaviour. Do you also confirm that it is not the case for page numbers viewed with IE at ? ThomasV (talk) 19:59, 21 October 2010 (UTC)
IE 8 + Monobook + = no problems, I can see the page numbers just fine. Still not here though. —Spangineer (háblame) 20:06, 21 October 2010 (UTC)
so, there is something in the site’s javascript that causes this. I cannot hunt it now, I have no windows machine at home. ThomasV (talk) 20:19, 21 October 2010 (UTC)
I'm going to take back my previous too; something has most certainly improved though things are still far from normal here.

After systematically working back from IE8 with the "latest" WS skin (which I assume equates to vector.css ??) applied, the degrees and instances where articles that were all a result of transclusions from the Page: namespace using the current <pages index /> method that managed to display the embedded page links along the left side became progressively worse and worse as I went through my anal procedures of clean caches, mirrored-settings, Windows Update, etc. to lower versions of IE until they stopped completely at IE6 w/monobook.css.

My current combination of IE6 and vector.css (or current skin) CAN consistently manage to display the embedded page links but never on first click into the page and only after certain operations relating to what I can only call "refreshing" or "purging" (depending on what was last, combinations of [F5] and/or [CTRL + F5].
Although everything I viewed over on the fr.wikisource was only using about half of the available window space, everything I looked at that had embedded page links displayed & functioned fine under IE6 floating vector. The hover-to-highlight feature on long articles transcluded from many Pages: occasionally displayed the same behaviour here prior to it not working at all where the highlighted section would freeze over and nothing could move the highlight's focus until doing the refresh thing above. Even with the highlighted freeze - scrolling, clicking, etc. worked as expected. George Orwell III (talk) 05:23, 22 October 2010 (UTC)
The previous may have held true just a few hours ago but that may no longer the case. I suspect the reason IE did not fart and sputter for me earlier on fr.wikisource was because the content was being displayed at some pre-determined(?) setting -- at something akin to 20% left-margin, 40% center content and 20% right-margin (most similar to dynamic-layout setting #2 here I think). Cycling through the layouts and some more purging & refreshing eventually brought it back to near original. It seems the embedded page links loathe to disappear compared to what I'm seeing here today... but I'm able to view far more, and with a better degree of alignment, here in a single glance. I'll just avoid this routine whenever I can I guess. George Orwell III (talk) 11:58, 22 October 2010 (UTC)

Executive Order 13252Edit

New weirdness noticed today concerning the caching-javascript thing above and a short, 2 line transclusion within the linked EO. After doing the combinations of [F5] and/or [CTRL + F5] to toggle the embedded link on , the [custom] header "breaks". Repeating until the embedded links disappear again restores the header template to as it always was.

I know this is probably due to something we have in the custom header and whatever that is only revealed itself for the first time after the most recent of changes, but I figure I should mention it in case it can help point to something that narrows the search for the solution. 05:23, 22 October 2010 (UTC)

The bug seems to be fixed now. In addition, I have moved all Proofreadpage related scripts to the beginning of Common.js. This will ensure that they are executed first ; a bug in the rest of the javascript will not affect them anymore. ThomasV (talk) 09:19, 22 October 2010 (UTC)
Thank you for your attention. Unfortunately I can still reproduce this pretty much at will now. Is there anyway to turn off the the focused highlighting thing? I say screw 'searchable terms that break' when across page-breaks or by side-notes if it results even in a marginal improvement in a reader's ability to easily access the index's pages from the main content. George Orwell III (talk) 11:58, 22 October 2010 (UTC)
you are right ; it looked fine because I was not logged in. When I am logged in, it still does not work. In the meantime, I added an option to disable highlighting : try "self.proofreadpage_disable_highlighting = true;" in your .js ThomasV (talk) 14:20, 22 October 2010 (UTC)

Dictionary of Music and MusiciansEdit

Hi Thomas, Well, it's been considerably more than the "couple of weeks" I thought it would take me, but the A's, B's and C's are all proofread and sectioned from Index:A Dictionary of Music and Musicians vol 1.djvu & Index:A Dictionary of Music and Musicians vol 4.djvu. When you've got time can you please run your indexing bot over them to create A Dictionary of Music and Musicians/A, A Dictionary of Music and Musicians/B and A Dictionary of Musicians/C? The two sequences (vol. 1 and vol. 4) should be merged into a single alphabetic sequence for each "chapter". There is text in vol. 4 that has deliberately not been sectioned and it should not appear in these transclusions. Thanks, Beeswaxcandle (talk) 04:10, 25 October 2010 (UTC)

What code recreates a colon?Edit

I'm trying to get a line to indent across multiple pages, and I want to recreate what the colon does normally. What code do I use? Right now I've got <div style="margin-left:2em;">, but the amount of space between subsequent paragraphs isn't quite right. See the first three paragraphs on Page:No Treason, v6.djvu/12—space between paragraphs 1 and 2 doesn't match the space between 2 and 3. —Spangineer (háblame) 22:05, 28 October 2010 (UTC)

a colon translates to <dl><dd> text </dd></dl> ... HTML-wise. So for your situation I applied...


<dd><p> Start of Para. 1
</p></dd></dl> <--- in non-transcluded footer
<dl><dd><p> <--- in non-transcluded header on next page
End of Para. 1 </p></dd>
<dd><p> Para. 2 </p></dd>
<dd><p> Para. 3 </p></dd>


.... and all 3 paragraphs seem to display equi-distant now when transcluded. George Orwell III (talk) 22:47, 28 October 2010 (UTC)

Thanks! —Spangineer (háblame) 12:44, 29 October 2010 (UTC)

Use of the index ribbon for a book?Edit

Gday Thomas. Is there any way to individually use one of the ribbons to automatically monitor the proofreading status of a book, similar to how it is undertaken in the Special:IndexPages? I am trying to have the ability to quickly identify the progress of selected works for Validation month, and the only way that I can see such is to open the respective Index: ns pages, which inconvenient for an active monitoring task. This could be some code to display the ribbon beside a list of names of Index: files, as per User:Billinghurst/PotM_watch, or a means to select the books to be displayed via the Special pages. Thanks. — billinghurst sDrewth 09:33, 2 November 2010 (UTC)

did you consider adding the keyword "potm" to the index pages ? ThomasV (talk) 12:16, 2 November 2010 (UTC)
Didn't know that it was possible to add keywords individually, nor that it would allow such a filtering. Phe has given me some hints, and I will await the lag to the next indexing pass and see how it progresses. Thanks. — billinghurst sDrewth 13:17, 2 November 2010 (UTC)

Common style problem - initial page number missing in layoutsEdit

Hi, I thought that it was just me, however, I now have another report. When transcluding to a page like British Medical Journal/1924/Alexander Charles O'Sullivan the page number on the initial page goes missing. I can replicate in FF (v3.6 & v.4) and Chrome, and logged in using monobook and logged out with vector, and different PCs. Actually to note that in Chrome utilising layout 2 (either vector or monobook) the page number for the second page flicks to the near the top of the article. — billinghurst sDrewth 01:08, 4 November 2010 (UTC)

fixed. did I not mention hat <pages/> should not be placed inside a div ? ThomasV (talk) 09:13, 4 November 2010 (UTC)
I did mention it, see Help:Layout ThomasV (talk) 09:16, 4 November 2010 (UTC)
Thomas, we have tens of thousands of pages that have transclusions within a <div> as it was previously a requirement so page numbers appeared inboard from the left boundary. So the means that this process has taken place and the implications on appearance are not evident, and subsequently have not been communicated clearly to the community. That it appears in an unlinked file doesn't make it universal knowledge. :-( — billinghurst sDrewth 12:14, 4 November 2010 (UTC)
Sorry, but this is so obvious to me that I thought you had gotten it : the whole point of dynamic layouts is to that users do not need to add layout instructions anymore.
As Cygnis stated it, the text you type in should be layout neutral. As far as understand, this "layout neutral" policy already existed at before, and texts using transclusions were kind of an exception to the rule. They are no longer an exception.
I have tried to make the previously existing classes (lefttext, indented-text) as harmless as possible, so that previously existing pages are still correctly rendered. the drawback of this approach is that some users might continue to add these divs without knowing that they are unnecessary. Perhaps a robot cleanup will be necessary.
ThomasV (talk) 09:15, 6 November 2010 (UTC)
If I understand the comment correctly, billinghurst is pointing out that indented-text, the standard class (i.e. 'layout neutral with page numbers'), is affected by this implementation. Removing that by a bot is what I anticipated would happen, whether we have one or many layouts for transclusions, but that depends on stability and consensus to have this scheme. I anticipate that will happen eventually, and it can be widely announced and added to the help pages.
I pointed out the bug because I imagined the information might be useful to you, the positive effects of the layout options scheme already outweigh the loss of the first page link. I don't see that this could have progressed any other way, the merits become 'obvious' when it is implemented, but users applying the 'standard' class seems prudent as they 'wait and see' what emerges.
Billinghurst and I are very active in advising others of what is going on, I'm sure I speak him too when I say that your work is greatly valued, and hope the bug reports are helpful as you optimise the software. Cheers, cygnis insignis 09:59, 6 November 2010 (UTC)
I was more trying to reflect that when we make a change that affects the community in conjunction with the way that we do things, that we should be explicit in the information provided and that we wave under the noses of the natives, not leave it as implicit. We need the broadest understanding, not for the select who are positively paying attention, especially when we all have life interrupting our WS realities. I am not saying that it is your responsibility, more that as the person making the changes it that helps when you at least nag other people so we can as community get the message out there. For late Sep through to the past week, my brain has had plenty of other distractions, and has had to have some throttling on issues, and global formatting was one that didn't get its full attention.

Amen Cyg, I do really appreciate ThomasV's work and his provided application. Sometimes we do the doing and are remiss on managing the implications and the associated messages. Useful for each of us to take a step back, to be reflective of what it means and have we communicated what it means successfully. — billinghurst sDrewth 12:26, 6 November 2010 (UTC)

No. of pages on Index pageEdit

Is there an easy means to identify how many pages are in a boo without having to go to the Index: page? I am trying to get a list of small works that are not validated, and find that I have to either open the work or have scroll through lists and lists of works. I either would like be able to sort the works or to be able to run a bot through. Ideas? Thanks. — billinghurst sDrewth 00:17, 26 November 2010 (UTC)

Really this ought to be accessible through the API by querying imageinfo on the DjVu file. I've submitted a bug. Hesperian 01:33, 26 November 2010 (UTC)
Fixed already! Just waiting for it to be deployed to the servers. If you notice this API result returning the number of pages in the file, let me know and I'll code us a list of works sortable by distance to completion. Hesperian 00:16, 28 November 2010 (UTC)


Gday, some of our pages refer to Template:PageQuality, though I understood that this was now overtaken by a tag that undertakes a similar process. That being the case, should we be looking to take the information and update to your new scheme. Have you got newer documentation available? — billinghurst sDrewth 07:15, 6 December 2010 (UTC)

pages with the old template are being converted automatically everytime you edit them. There are 6376 validated pages that have the old template, though. I might convert them with a robot when I find time; if you want to do it you’re welcome of course :-). no doc available but all you need to do is program your bot to do what the javascript does ThomasV (talk) 11:19, 7 December 2010 (UTC)
I obviously wasn't clear and I understand that process and the conversion. Here I was talking about the informational pages oldwikisource:Wikisource:ProofreadPage#Proofreading path and its direct reference to the Template. I am not sure of how to amend that page to reflect the new reality. I will add the task to Bot requests. — billinghurst sDrewth 13:19, 7 December 2010 (UTC)

you’re right; I updated it in a very deletionist way [7] ThomasV (talk) 13:49, 7 December 2010 (UTC)
Yep, saw that. That is even within my skill-set! Thanks. — billinghurst sDrewth 13:57, 7 December 2010 (UTC)

User:ThomasBot/Help:Bots and ThomasbotEdit

Another question. Is this page still relevant with respect to your bot? — billinghurst sDrewth 13:27, 7 December 2010 (UTC)

hmm this robot is broken ; I really have no time to investigate what is wrong with it atm ThomasV (talk) 13:50, 7 December 2010 (UTC)

A quirk in some of the stats generatedEdit

Hi ThomasV, To me we either have an amazing coincidence or a quirk in the stats being generated. At Wikisource:Administrators'_noticeboard/snapshot those stats are fed into WS:AN, and {{PROOFREAD PAGES}} outputs 23,508 and {{VALIDATED PAGES}} outputs 23,508. The same. Thanks. — billinghurst sDrewth 07:59, 3 January 2011 (UTC)

Pages remain in pagequality categories after being movedEdit

Hi Thomas,

Page:Crainquebille, Putois, Riquet and other profitable tales.djvu/236 was moved and is now a redirect. Looking at the page itself, it is not in a pagequality category. But if you look at Category:Validated, you will find it in there: web, api. It seems that moving a page does not trigger an update to the category tables, if the category is applied by your extension tag.

(I found this bug when scripting up a list of Page: namespace pages that are not in a pagequality category. I got lots of valid results, but also many invalid results, many of which are somehow associated with page moves. In case it help you with debugging, the list is at User:Hesperian/Script.)

Hesperian 12:53, 7 January 2011 (UTC)

Possibly related: Page:United States Statutes at Large Volume 1 - Congress 3-5.djvu/227 has been moved and is now a redirect. The page's Special:WhatLinksHere/Page:United States Statutes at Large Volume 1 - Congress 3-5.djvu/227 states that it is transcluded into United States Statutes at Large/Volume 1/5th Congress/2nd Session/Chapter 54; and the latter page, when in edit mode, lists the former as a page it transcludes. Both are wrong. I think there is a problem with the ProofreadPage extension that means page moves are not triggering the necessary updates, and the database is being slowly corrupted. Hesperian 05:26, 9 January 2011 (UTC)

I've done a cleanup of redirects in page namespace, so the evidence of the problem is gone. I'm pretty sure the problem is still there though. Hesperian 01:37, 10 January 2011 (UTC)

DjVu file with 2 "page" numbers per pageEdit

Hi Thomas: I have a problem I've never seen before. I just uploaded Index:American Archives, Series 4, Volume 1.djvu, which has the interesting characteristic of numbering each column, rather than each page. There are two columns per page, so Page:American Archives, Series 4, Volume 1.djvu/521 has on it numbers 913 and 914. Citations to this work are based on the numbers 913 and 914 (naturally). Do you have any suggestions for how to handle this, in terms of the index <pagelist> and the page links that appear in the main namespace? Is splitting all 1000 pages in half and re-creating the djvu file the best option? —Spangineer (háblame) 17:47, 19 January 2011 (UTC)

As far as the transclusion part of your dilema, I had a similar concern awhile back and started to tinker with {{Pagebreak2}} to eventually work-around the same 2-column 2-item designation problem but never finished it since I lost access to the work in question in the interim. See the sandbox for starters and hopefully you can gleam where I was going -- George Orwell III (talk) 19:33, 19 January 2011 (UTC)
Looks promising, but I can't figure out how to eliminate the extra page number... {{pagebreak2}} generates two numbers, one in the margin and one in the text. The one in the text needs to be removed somehow. —Spangineer (háblame) 20:05, 19 January 2011 (UTC)
Sorry I was just messing around in both Sandboxes - Do you still see that? (remember: I am IE impaired :-( George Orwell III (talk) 20:08, 19 January 2011 (UTC)
I see the double page numbers in the link you gave to the Sandbox in IE, Firefox, and Chrome (in the latter two, it's about 8 lines up from the number in the margin). I also saw the same thing in both IE and Chrome when I added {{pagebreak2|916|left}} to Page:American Archives, Series 4, Volume 1.djvu/522. —Spangineer (háblame) 20:19, 19 January 2011 (UTC)
Duh! I realized eliminating the output would let the pagenum class stand alone. Check it again. Of course, now the trick is to get it point back to the right page & anchor in the Page: namespace. -- George Orwell III (talk) 20:50, 19 January 2011 (UTC)
Well it's not pretty but it seems that adding a 3rd parameter (title) to Pagebreak2 gets the link working somewhat right. Check the last Sandbox now. --George Orwell III (talk) 21:17, 19 January 2011 (UTC)
Got it working at Continental Association. I agree it's not pretty... it especially won't be pretty when a work starts in the second column (the first column's page number will appear instead). Better than nothing, at least, until we find something else. I found Help:DjVu_files/other_pages#double_pages_in_djvu, where there's discussion about image splitting. Not sure if that makes sense... —Spangineer (háblame) 21:23, 19 January 2011 (UTC)
Well the template needs to be tweaked and optimized (probably could use a proper anchor in the Page: namespace that won't transclude to the mainspace). I'm the only editor that's used Pagebreak2 until now. It's really just Pagebreak with addition of a 4th "hidden" parameter which I added at somepoint just for EOs.

As for works starting in the second column - you'd be doing a labeled section transclusion at the top either way wouldn't you? -- George Orwell III (talk) 21:38, 19 January 2011 (UTC)

I'd use <pages>. For example, if the images were all split, I could do <pages index="blah.djvu" from="Page containing column 2 of page x" to="Page y" fromsection=anchor /> instead of <pages index="blah.djvu" from="Page x" to="Page y" fromsection=anchor />. Result of the first is that the leading page number will be correct and point directly to the column of text I want. Result of the second is that the leading page number is incorrect and doesn't point anywhere. —Spangineer (háblame) 21:49, 19 January 2011 (UTC)
Well I wrapped the entire template in another span that duplicates the id of the second variable span if it makes any difference. Also, I created another option labled "tran" that mimics the one we were testing. Please use that instead of "left" as the second value from now on. I figure best not to stray too far from the original template's layout in case templates get merged, sliced or forked at some point down the road or something. -- George Orwell III (talk) 23:08, 19 January 2011 (UTC)
← I see you added #(((1|))) at the end of the title= parameter. I don't believe that is a valid string for that - once I removed the [2nd] wrapping span tag and it stopped going directly back to the second column from the main namespace (still goes to the right Page: though).

I tried #((anchorencode:(((1|))))) too; no difference. I'm assuming you see the same? -- George Orwell III (talk) 16:28, 20 January 2011 (UTC)

It appears to be working fine in Chrome and Firefox, but not IE. Not sure why though; if you have any ideas I'm all ears. —Spangineer (háblame) 16:46, 20 January 2011 (UTC)

Generation of /‾source‾\ tabEdit

Gday ThomasV, I had been of the understanding, that the /‾source‾\ tab was generated by <pages, however, now I see that as a misunderstanding. At Greaves, Edward (DNB00) I have transcluded text with <pages>, then also used #lst transclusion to pull in erratum from a later volume and appended to the page. This act now displays the /‾source‾\ tab of the #lst transclusion, which is not the desire result. Is there a workaround to my dilemma? Thanks. — billinghurst sDrewth 11:01, 7 February 2011 (UTC)

not currently. in the future the software could detect it and display a warning ThomasV (talk) 16:50, 21 February 2011 (UTC)
Later thought. If two different sources are being transcluded, how does it decide which source link to display in the tab? I have tried reversing the order of transclusion, and that didn't seem to have an effect. Billinghurst (talk) 03:03, 21 March 2011 (UTC)

Pre-populating index pagesEdit

ThomasV, question for you: how are new Index pages pre-populated with the text found at MediaWiki talk:Proofreadpage index template? I thought it was part of the ProofreadPage extension, but I recently installed that extension on my wiki, and all I get are blank edit windows when I go to create a new index page. Wikisource doesn't have any preloading extensions installed that I can tell, so I'm at a loss. Can you explain? As always, thanks for your help. —Spangineer (háblame) 23:04, 9 February 2011 (UTC)

Thanks for pointing me to oldwikisource:Wikisource:Shared Scripts. —Spangineer (háblame) 16:17, 10 February 2011 (UTC)

Page:Best Russian Short StoriesEdit

Hi. Is it necessary? [8]? It is really pleasure these stories to transcribe :) Tommy Jantarek (talk) 17:17, 12 February 2011 (UTC)

That wasn't Thomas, it was me who set the bot working. It's supposed to make it easier to proofread works that are already in the Mainspace. Beeswaxcandle (talk) 00:21, 13 February 2011 (UTC)
Tommy, please see User:ThomasBotbillinghurst sDrewth 00:27, 13 February 2011 (UTC)
OK, I understand. ThomasV, I'm sorry ;). Tommy Jantarek (talk) 02:47, 13 February 2011 (UTC)

Following the fix of sectionsEdit

I am not sure what has occurred with regard to the fix in the section transclusion while I slept, however, I see something has changed in that we now get sections. :-) I do see that the sections are not flowing and joining neatly over pages. If you knew this already. <shrug> — billinghurst sDrewth 01:59, 21 February 2011 (UTC)

yes this is known ThomasV (talk) 16:50, 21 February 2011 (UTC)


for all the fixes and a great summary…JamAKiska (talk) 12:01, 22 February 2011 (UTC)

Align and DoublewikiEdit

On Eskimo Life/Chapter 2 the Doublewiki links are broken. I tried to introduce {{align}} but this did not change the situation. What's wrong? For this book, the text is available in two other languages (Norwegian and Swedish). Do I need to call align for each remote language, as I tried to do now? One difference from your example Crito is that my example has a slash in the page name. The book's main page Eskimo Life seems to work. --LA2 (talk) 14:37, 24 February 2011 (UTC)

It works fine if the link actually contains "/", but the problem is that it contains %2F which is a hex-encoded slash. --LA2 (talk) 14:56, 24 February 2011 (UTC)
In IRC, phe said:
perhaps the url would be formed ala
the trouble is in Mediawiki:commons.js BilingualLink() : wgArticlePath.replace('$1', encodeURIComponent(wgPageName)) + '?match=' + bilang;
var bilanglink = wgServer + wgArticlePath.replace('$1', wgPageName) + '?match=' + bilang;
encodeURIComponent() translate the / in %2F and server doesn't like url ala /wiki/something%2F1?match=xx
--LA2 (talk) 18:57, 24 February 2011 (UTC)

As for align, it seems impossible to work with. It can't match lines with are headings, italics or small caps. It gets out of sync as soon as the page contains images or tables. It would be far easier if the proofread text contained numbered anchor points (similar to Bible verse numbers). --LA2 (talk) 20:31, 24 February 2011 (UTC)

Gadget doesn't play fair with ProofreadPage ... nulled!Edit

The gadget MediaWiki:Gadget-HotCat.js doesn't behave itself with ProofreadPage in when it is used it breaks your header components adding an extra header, alternatively it runs up against the status barrier and thinks that you are changing the proofread status, even when not. Either way it is problematic, so I have asked John to add a condition to the gadget that nullifies it in the Page: namespace. — billinghurst sDrewth 12:13, 7 March 2011 (UTC)

that’s fine with me ; it is difficult not to have incompatibilities between different pieces of software ThomasV (talk) 15:27, 7 March 2011 (UTC)

ProofreadPage is not populating categories on WikilivresEdit

Hello ThomasV. We're using the 1.16.x snapshot of ProofreadPage on Wikilivres, but the extension is not populating the proofreading status categories. For example, see MediaWiki:Proofreadpage quality0 category ("Without text"), Category:Without text, and Page:Platon - Œuvres complètes - Tome VIII, 3.djvu/4. Updating to the newest pre-ResourceLoader version doesn't help. Do you know why the categories might not be populated? —Pathoschild 01:49:06, 17 March 2011 (UTC)

I found the cause. In ProofreadPage.php, the pr_renderPages and pr_pageQuality handlers take the $parser argument by &reference. This is apparently broken in PHP 5, but isn't necessary anyway. Taking it by value resolved the issue. —Pathoschild 03:45:26, 18 March 2011 (UTC)

Setting a default layout for a bookEdit


I see that you wrote Help:Layout.

Can you please take a look at its talk page?

Thank you. --Amir E. Aharoni (talk) 13:37, 4 May 2011 (UTC)

ref followEdit

Hi Thomas,

I've started trying to track down pages that use my old-style footnote overflow system, and convert them to the new ref follow= syntax. I recently encountered a page with an overflowing footnote with a table in it that itself contained a footnote. I had to use the #tag:ref magic word, and protect the table pipes with {{!}}, and it was all rather a mess. But even in this horrible case, the ref follow= syntax behaved exactly as I expected, and greatly simplified the whole thing. I really am very impressed by how elegant and simple this solution is, and how well it works. Thanks again,

Hesperian 04:51, 15 June 2011 (UTC)

Set layout by nameEdit

Hi ThomasV! As part of the code for the default dynamic layouts, I wrote a function called set_layout_by_name as a counterpart to set_layout. I thought maybe it could be incorporated into PageNumbers.js in order that others can use it easily.

function set_layout_by_name(name){
    if( self.ws_layouts[name]==undefined) return; //layout doesn't exist
    var i=0; for( key in self.ws_layouts ) { if(key==name) break; i++; }

There's no rush as the code is fine where it is. Just thinking of the other domains. Cheers, Inductiveloadtalk/contribs 10:04, 2 September 2011 (UTC)

Set layout by cookieEdit

Similarly, a convenience function for setting the layout to match the cookie would, I think, be useful as a way to revert from programmatically modified layouts. This isn't new code, just lines 78-81 of PageNumbers.js broken out into a separate function:

function set_layout_from_cookie(){
   layout = GetCookie("layout") ;
   if(layout) layout_num = parseInt( layout ); else layout_num = 0;
   if(!layout_num) layout_num = 0;

Cheers, Inductiveloadtalk/contribs 08:49, 4 September 2011 (UTC)

"Source" tab IDEdit

Hi again! Would it be possible to give the "source" tab on pages with transcluded pages an ID attribute. Perhaps "ca-source"? This will make code that looks for the source tab more robust. Thanks Inductiveloadtalk/contribs 02:46, 4 September 2011 (UTC)

Do "Bad Things" happen if you allow "normal" pages to have layouts?Edit

I've been looking though the code for PP, and it seems to me that the Dynamic Layouts can be applied to any page, not just those with page numbers or transclusions. This way, any page can have custom styling applied. I have managed to "subvert" the test in PageNumbers.js but it's not a clean break, and I was wondering if the test still has use when Dynamic Layouts have the power to modify any page beyond just making space for the page numbers. Basically, I'm wondering if anything bad happens if you kill line 44 of PageNumbers.js:

if( !self.proofreadpage_source_href && get_elements_by_classname("pagenum","span").length==0 ) return;

Inductiveloadtalk/contribs 08:53, 4 September 2011 (UTC)

Question on progress barEdit

Hi. Is there an easy way to generate the color-coded bar that appears on pages in Main namespace which use transcluded pages (the colored little bar in the upper left at the beginning of the page)? I would like to generate that progress bar for each article listed in a table, to show in the table the different quality levels of each of them. Ideally, it should generate the bar automatically, as happens normally, starting from the title page. Basically, the same that is done here, but for a set of pages that I define in a table. Hope I managed to explain myself ... Thanks --Mpaa (talk) 02:02, 10 November 2011 (UTC)

Hi. I posted this also at Scriptorium. You can answer there if you wish. --Mpaa (talk) 22:59, 10 November 2011 (UTC)

Unsuccessful confirmation of adminshipEdit

Hi Thomas,

Your 2012 admin confirmation has been unsuccessful due to inactivity, and your admin rights are requested for removal. Thanks for all your work here — as developer, admin and editor — and I hope to see you back here some time.

Hesperian 00:18, 1 May 2012 (UTC)

no problem. it's been nice to participate in Wikisource. I now have other occupations. ThomasV (talk) 13:56, 12 January 2013 (UTC)

Request for removal of sysop statusEdit

Your sysop flag has been removed following the conclusion of local inactivity procedures at [9]. Snowolf How can I help? 00:18, 1 May 2012 (UTC)

Js error into MediaWiki:PageNumbers.jsEdit

I see by Chrome js console:

Uncaught TypeError: Cannot call method 'scrollIntoView' of undefined 

and I see that MediaWiki:PageNumbers.js has your name into header. I presume that it's not a blocking error, nevertheless I think better to tell you. --Alex brollo (talk) 20:25, 13 October 2013 (UTC)