Last week and the beginning of this week I've brought the styling of the CC site to almost its final look.
April 15, 2008
Stayed half an hour late assisting with CP's presentation in the CALL on Monday. Quit an hour early Tuesday to do Half-Baked business.
About 260 files were missing <idno type="received"> tags in the header, because the format of their original text slightly varied from what was expected ("registered" instead of "received", followed by the date). After some testing and one false start, I managed to fix this using an XQuery against the db. I really could do with some code which would back up (i.e. bring down from the db) only files which have changed since a certain date, because right now when I run one of these large XQuery routines, I have to backup all the files from the DB, then move them all over to replace my copies; I'm unable to tell which files were actually changed by the XQuery, because files backed up all get the same date/time. I'll have to try and figure that one out.
Found a handful of mistyped CO number/volume data in BC files, and fixed it after checking with Jim. Also normalized the form of Carnarvon so that it always includes his title (Parliamentary Under-Secretary). It was in two different forms before.
April 14, 2008
Up to now, my appreciation of digital art has fallen into two dutiful categories: useful art, the kind of sticky website combination of visual, textual, useful applied art you find in websites that keep you coming back; The second kind, what I privately think of as "geek art," beautiful (in its own way) but with a limited, coded, coterie audience to marvel at technique (OMG who knew Flash/Java/etc. could do that?).
I'm old enough to realize I'm channeling Wiliam Morris: "Have nothing in your houses that you do not know to be useful, or believe to be beautiful. " But this weekend, I was moved by two websites where the message and the new medium come together seamlessly.
Stayed late manning the office while others were at JL's presentation in the CALL.
Wrote XQuery, XSLT, JavaScript and CSS to retrieve the bios of anyone tagged with the right <name key="blah"> tag in the documents. The first demo page is the first BC 58 document, which seems to work pretty well except that the Elliot bio (the last one linked) is not being retrieved for some reason. I'll work on that tomorrow. Then we can figure out how practical it will be to create these links by search-and-replace.
I've now finished 349 footnotes, and 149 remain. Found a couple of marginalia items which should have been removed to the documents themselves in the first handful of files (I probably hadn't finally decided how to handle them at that point), so those have now been removed, leaving us with 498 notes at the moment.
IE 7 returns different values for offsetTop and offsetLeft from other browsers, so when you're setting the position of an object you have to be a bit careful. I've fixed this on other sites, and had forgotten to add a fix here, so footnote popups were showing up above where they should have. I've now fixed that, using the findPos function described here.
Since Sean can't make it in today, I've done some remaining tweaks to the biographies file -- mainly changing some surnames to title case, and changing one name tag from surname to forename. I also updated the XQuery so that it sorts the namelist by lower-case version, so that any case problems don't affect the sorting.