Monument 2025-06-23 to 2025-06-27
to : Martin Holmes
Minutes: 700
New urgent requests for changes to the configuration and rendering of the data
for the Monument from ST and JG led to a meeting on Monday morning, and the first
hour or two of work on the refactoring. By the end of Monday, I had investigated
and figured out the reason for discrepancies between the number of communities
listed on the site locations page and the total number appearing as components
of area pages: 12 communities no longer have anyone eligible for the Monument
being uprooted from them. Added more detailed stats and diagnostics for this to
the diagnostics page, and then refactored a lot of the XSLT to switch to the
term communities
for this lowest level of the hierarchy, instead of the
ambiguous locations
.
On Tuesday, continued this work, which surfaced a few more questions and edge cases which I’ve sent to the JCLS folks. Published a new version of the public site, and got the listings of communities inserted into the headings of areas on the output document; what remains is to do the alphabetic splitting of the larger areas (and there is a question outstanding on that), and then the rewrite of the diagnostics that test their PDF for any name discrepancies, which will have to be updated to take account of the community listings and the alphabetical sub-groups.
On Wednesday, I started on the alphabetic subsection code, starting with the plain-text output, since that’s only informational and not provided to the JCLS, but serves as a good test-bed for the alterations. By the end of the day I had this splitting working; the process is quite different from that used for creating the separate web pages, so I had to re-implement the core functions for the text output, and I’ll need to do the same thing for the HTML that is the actual product, but the approach has now been tested.
On Thursday morning, got a full implementation of the HTML page output working, and sent a link to ST and JG to get their feedback. Then tweaked it a bit to add more support for a rewrite of the PDF checking algorithm, which is next on the agenda, I think.
On Friday, got the OK from JG to send a new Word document and did so. Now we wait for the first instance of a PDF so we can rewrite the checking code. I have purposely constructed the HTML so that it can serve as the basis for checking, rather than the original TEI; we should check against the original TEI to confirm that all names appear, but only by checking against the HTML can we confirm that they appear in the right locations and subsections.