Met with JJ and CB to do some preparation for the meeting tomorrow with PS. More info coming...
September 10, 2012
Two new people to add to the pay timesheet system. Created new timesheets for them. Still waiting for Sept 1-15 hours for 4 people.
September 5, 2012
Codebase is now clean of name/@type. In the process, found an extra file in the db which wasn't in versioning (DITC1), so added it.
September 4, 2012
In our born-digital documents, we're sometimes assuming that the page title will be taken from the teiHeader, and sometimes assuming that it will be the first div/head in the page; this is very confusing for code which is trying to render a title and then provide a TOC as well. I've put a very messy set of rules in place that attempts to render the page title based on what it can guess about the author's intent, but we're going to have to firm up our practice here. I think we need to formalize what we're doing here. For the page title, I think we should be using text/front/head in all situations, even where we have text/group following it. That would put the page title in the text without its being confused with a toc (which would be generated based either on body or on group). I'll have to figure out how best to make the changes smoothly, though.
After checking that CB has removed all instances of name/@type[.='pageant'], I've started removing the now-redundant @type='person'. I only got as far as /info/ before I found a different bug to focus on, though.
I've rewritten the TOC code so that it can cope with both the old <group> structure and a simpler setup based on <div> nesting. I've also converted all of the files in /info/ that were using multiple <text> elements so that they use only one <text> with nested <div>s. All of these are now rendering OK, and the old <group>-based pages elsewhere are also still rendering with TOCs while the team decides what to do about them. We can leave any pages that actually need multiple <text> elements alone if we want to, but any we do change will now get working TOCs.
I also have a handy XSLT tranformation that can turn a <group>-based document into a single-<text> document automatically, so conversions don't have to be done manually.
div[@type='placeInfo']/head from the rendering of the locations pages, so those pages are only getting one title, and it's taken from teiHeader/fileDesc/titleStmt/title. CB having some problems with those <head> elements at one point.August 30, 2012
With JJ, have been working on aligning our byline attribution phraseology with Marc Relator terms. We're almost there.
August 29, 2012
Added <eventList>, <event> and <rs> into the schema, then created the eventography file, with one sample event, and marked the event up in two files using <rs>, as a demonstration for CB. Once we have some real data in the eventography, I'll add the requisite XSLT to make it functional.
Fixed several hundred encoding oddities and errors related to name tags and docAuthor tags. Also updated Schematron to try to catch some of these. We're now in a position where we have almost all names marked up as either @type="person" or @type="pageant", and the question now is whether we should simply make "person" the default, and therefore optional, and save a bunch of space in files by deleting all instances of it. I see no reason not to do that. There are some name tags which lack it already, and they're working properly.
August 28, 2012
Working on standardizing the current byline markup so that I can use it as the basis for generating more formal <respStmt> elements using our taxonomy of responsibilities. I have all names now marked up as names, but the informal descriptions of their roles are going to be very difficult to parse into clear categories.