An enlarger function to replace the jump to a Google Maps site page is now working on the places index page, and I think it might be possible just to integrate the same functionality right into the popup.
Enbedded GMap functionality now implemented for listing page. More work to do on popups.
...which found over a thousand errors in key references, of which I've fixed a couple of hundred. More to do here...
Prepped some materials and taught a brief intro to XPath and Schematron for the DR students; they've now created a list of a dozen constraints they'd like to apply based on their own experience and consultation with GL, and next week we'll start coding them up.
I've linked all files to a new Schematron file, and added some rules; I then made substantial fixes throughout the collection to eliminate spaces from the beginnings and ends of name elements, an annoyance that has been plaguing us for a while. All files now validate, and we're ready to start work on Schematron with the DR students.
I have the places index now working with a much-modified version of the Google Maps code used on MoEML; there's still substantial work to be done there. We also have to make similar but not the same changes to the popup place references.
I've made some substantial changes to the way
<geo> tags work in the Despatches. Geo-locations are now encoded using a single
<geo> tag with multiple coordinates in it; each coordinate pair (lat/long) is delimited by a comma, and the list of pairs is delimited by spaces. This brings us into line with the way other projects such as MoEML handle geolocations.
The documentation has been updated, as has the PDF generated from it which is in the docs folder alongside the XML. All current places in the places.xml file have been updated, and the code which generates KML for display on Google Maps is also now working with the new location info.
This means we can use Greg's excellent Vertexer tool to get the geo coordinates for a location, which is a lot simpler than what we used to do.
This was in the pipeline for a while, but became urgent because Google is about to pull the rug out from under us by abandoning support for the KML file setup we're currently using for these maps. Since this problem also afflicts the MoEML project, I needed to bring the two into line so that I can work on a single replacement for the functionality in both projects.
The last few still need to be proofed and edited by GL. Discussed plans for the next work they're going to do.
Finally decided to register the project with OAI PMH as a data source, and began the process of getting validated with their service. There are a lot of cases in which I've written my code to fail gracefully/usefully, when in actual fact the protocol requires hard errors to be returned; I now need to rethink some of the code to handle this. I'll probably reduce the granularity of the dating for simplicity's sake. A fair amount more work to do here.
Did a brief XML intro for the DR students, and then worked with them as they encoded their first bios and got familiar with SVN. They're now published, and credited on the site; they look good. Also solved a puzzle with CH -- John Henley was only wounded, not murdered as one despatch says, and the man killed was Bill Brady, not "George _____".