I've seen this mentioned on IRC yesterday and thought some of you might not be aware of it. fr.wikisource uses an interesting hack to show the source image that is being transcribed on a page, like here: http://fr.wikisource.org/wiki/Page:Fermat_-_Livre_1-000067.jpg
Now the really cool thing is that this even works on the edit page: http://fr.wikisource.org/w/index.php?title=Page:Fermat_-_Livre_1-000067.jpg&...
It's a hack based on Monobook.js; the function is called nsPage: http://fr.wikisource.org/wiki/MediaWiki:Monobook.js
Essentially, it checks whether the page title begins with "Page:", and if so, tries to load the image of the name that follows on both the view and the edit screen. This may not work in all browsers and certainly required JavaScript; I've only tried it in Firefox. But I think it's very cool in any case, as it gives you similar functionality to what Distributed Proofreaders is using, but with a wiki.
It may be worth for Wikisources to request a proper Page: namespace if this method proves to be viable.
well, this is still experimental, but it starts to look good, so I guess I can make an official announcement :-)
The whole point of this is to render texts through transclusion. Each book has a logical organisation (chapters, etc), and a physical one (scanned pages).
We should use scanned pages in order to correct typos, because that way we can compare the text to the page.
OTOH the text should be rendered as chapters for readers.
here is the main page of the book : http://fr.wikisource.org/wiki/%C5%92uvres_de_Fermat_-_I
if you try to edit it, all you'll see is a list of tanscluded pages. if you click on "liens vers les pages" (assuming you have javascript), links to the Page: pseudo namespace show up.
Once you have the page displayed, clicking on the image activates a zoom.
I started to write up a help page on this : http://fr.wikisource.org/wiki/Aide:Affichage_par_pages
I plan to translate it to English once it is more mature, but feel free to go ahead.
Thomas
wikisource-l@lists.wikimedia.org