Map Of London

Meeting this morning, with lots of suggestions coming out of it. I've reworked some of the document type taxonomy, created a path that lets you see all documents of a specific type, fixed many invalid documents, started work with KMF on expanding the site menu, and done many other fixes and tweaks.

Today:

• Ironed out some bugs in the code from yesterday.
• Added better handling for multiple authors.
• Added Chicago and APA renderings, after some discussion with TL. All the output formats will have to be reviewed by the academic wing.
• Added RefWorks export. The way this works is that you link to the refworks server, providing a URL which it can use as a callback to get information on the item. That callback URL provides a text file in RefWorks Tagged Format which the RW server then reads; the result is that you're taken to your RefWorks account and you see the data in an import box, and it's added to your db. I tested this with my own account on RW and it seems to work fine.
• Looked into providing export for RefMan and EndNote. Both of these can import text files in RIS format, so I think all I need to do is to provide a link which provides that text file, with the appropriate headers and mime type, and people running the desktop software will get an import; failing that, they can save the file locally and import it manually.

I've implemented a "Cite this page" system in the redesign, following discussions at the meeting this afternoon, and it seems to be working OK.

Right now it shows only a basic MLA citation, and only includes authors (we'll have to decide how/if other roles get included). Whenever you're looking at a page that has authors in its metadata, you should see the link; please click on it and check whether what you see makes sense, and report any oddities to me. Lots of documents don't yet have respStmts in their headers, so the authors may be lacking; for other pages such as site landing pages, the link won't appear.

Other decisions we'll need to make on this:

• Which style guides should we cover?
• Which roles should be included (e.g. editors)?
• Should we attempt to provide the last-changed date in the citation, as well as the date the reader is looking at the page? We could conceivably do this by parsing the SVN header in the XML document, but it would be a bit messy.

<change> elements in <revisionDesc> were in virtually random order, with people sometimes adding new ones at the end and sometimes at the beginning (I've been one of the main offenders with my automated processing). I've now re-ordered them all, and added a Schematron constraint that makes sure they're in descending order.

The search page incorporated both the search input box in the banner and its own search box, and the two had the same id, as did their parent forms. This is obviously invalid, and since there's no point in a second search box to confuse things on the search page, I've fixed the page-building stuff in includes.xql so that the banner box does not appear when you're on the search page.

Today I've implemented those components of the document type taxonomy that we've agreed on (there are still some questions to be dealt with around the categorization of various editorial documents). All documents now have at a <profileDesc> with a <textClass> containing at least one <catRef>, pointing to the taxonomy through the use of the mdt prefix, which is dereferenced to the boilerplate/includes.xml file.

I've also implemented support for this in the search page of the redesign, with five sample categories for testing purposes. This seems to be working well.

Same system as for EEBO. Note that both need to be documented in a proper prefixDef element, presumably in the includes file.

At JJ's request, I tightened up the spacing between the items in the Page Credits list, and added some XQuery and XSLT so that for each credit role (Author, Editor etc.), the formal definition of that role (according to MoEML) is provided as a title attribute, so it shows up when you mouseover the role name.

I've tweaked the JS for "fancy scrolling" such that it can now respond to the hashchange event, meaning that external sources such as news feed readers which navigate you within a page rather than between pages can still benefit from the elegant scroll effect. Surprised to find that the hashchange event now seems to be widely supported.

Added an rss 2.0 news feed to the redesign site.