Mariage

Fixed the "enormous title bug" (on TOC pages, the HTML title that shows up in the browser caption bar was the concatenation of all the titles of all the documents), the bug which inserted duplicate hidden text blocks into TOCs, and a bug which prevented bibliographical references in TOC introductions from turning into working popups.

After discussion with CC, we've decided to move to one centralized bibliography, containing all items. Articles etc. will link to items in that bibliography, and any items referenced in an article will be retrieved with the article and listed at the bottom. This makes for more consistency.

Checked the whole site for broken links, using the W3C link checker, and fixed a bunch of expired or obsolete links, mostly caused by changes to site structure.

In the course of site-wide link checking, I found that the hidden/revealed TOC introductions are for some reason included twice, resulting in duplicate content and ids. The Gallery page is an example. This needs to be fixed.

This has actually come out a bit better than I expected, because I've added a couple of features at the last minute. You can now retrieve a list of all links to a particular reference from two locations, the main "Allusions-notes" page, and the popup you see when you click on a reference link in a document. In each case, at the bottom of the reference explanation itself, you see the link "Liens à cette référence dans les documents". When you click on this link, the browser retrieves a list of links in the background and adds them into the page below the link. The links are sorted by document (so if there are two documents, there are two lists of links). When you click on an actual link, you will be taken to the location in the document where that link occurs; and the link itself will flash yellow a few times, so you can see it easily. Even if you're moving to a new page, the same reference item should open up on the page you're navigating to, so it should appear as though the reference popup stays open.

In the process, I've also done some more revisions to the layout and style, making the text a bit bigger in the popup, and moving it a little to the right; I've fixed some colour issues in the alternate stylesheets, and refactored some JavaScript code.

The link which retrieves references has been added, with a call to a JS function not yet implemented, along with a container for receiving the results from the AJAX call.

To do our planned round-trip reference linking, we need to be able to jump to a specific instance of a reference link in a document. This means that all links need to have unique ids, which is achieved by using their key, supplemented by a counter which makes them unique, counting 1-based from the beginning of the document. This has now been achieved, so you can navigate to a unique instance of a reference link in a document.

In the process, I found some bugs caused by the fact that the main reference file structure was severely broken; failure to close div elements had led to many references being nested several levels deep inside preceding ones. This is now fixed. We have to watch for this!

I'm planning a Mariage version of a feature which already exists in the Despatches project, which will enable you to see all the links to a particular reference in the documents. Title from CC: Liens à cette référence dans les documents.

Integrated the article display into the site based on instructions from CC, and also made an adjustment to the left position of the main content div (div#content), changing it from 18em to 26em. This makes for more blank space on the page, which makes reading easier. This is the first of several layout adjustments that I think we need to make, to make the site more elegant, and I'd also like to create a better default stylesheet (with black text on a light background).

Worked most of the day on the rendering of editorial articles (modern content). The tricky part has been figuring out how best to re-use a lot of the rendering functionality we already have (such as reference and footnote popups) while still separating out the article rendering so that it can be differentiated where necessary from the anthology content, without duplicating code. I have a new XSLT file called article_p5_to_xhtml.xsl, which is not long, but takes control of the processing at a crucial point, and a new rendering @mode attribute in the XSLT which can be used to fork the processing at various points. All documents still commence their processing with index.xsl, though, and inherit all the main site components at that point.

The basics:

• Articles are retrieved from the URL /article.htm?id=whatever.
• Bibliography items and lists which are fully defined in the original document are rendered visibly in the output, before the footnotes; they're also available as popups from links in the main body.
• Bibliography items which are referred to, but which reside in the main collection bibliography, are retrieved into the document but hidden; they are shown when the reference link is clicked on.
• Footnotes are rendered at the bottom of the document, but also shown as popups from buttons in the text.
• References should work like main-collection biblio items, although there are none in this article anyway.
• Links to anthology documents can go to the document, or to a page in the document (if there are pages with numbers captured by the appropriate tags), or to annotations in image markup documents.