Monument 2026-01-12 to 2026-01-16
to : Martin Holmes
Minutes: 670
On Tuesday, I met with JJ about the use of LOI prose paragraphs in the Monument site. Then I started working on the integration of the map HTML into the site, which entails a lot of additional work on the schema, since there are some elements and attributes that weren’t yet in the schema, and some structures that weren’t supported by my very basic content models. I am trying to update the XHTML5-as-ODD repository at the same time, so that we keep pushing that forward based on actual needs from real projects.
On Wednesday, I got the location map source code to be valid, and then started on the initial build out of pages. I am slowly porting over templates from the original site, revising as I go; the original site used all-TEI content, whereas our site pages are in HTML, so there are lots of tweaks to make. By the end of the day, got output from person pages and content pages which was valid, but so far not much actual content in the pages.
On Thursday, made a lot of progress with building pages. I created a procedure in the Landscapes repo which generates a map of all LOI ids of people who have casefiles to their individual casefiles; this is required for the Monument build, but the Monument repo does not contain that information, so it needs to be generated in the LOI repo and the resulting XML file copied over into Monument every time the LOI data changes. Then I added code in the Monument repo to consume that file, and that was the last piece required to generate person pages as they used to be generated.
Then I parameterized the code to generate relationship information on people pages.
The parameter is familyInfo=full|list|none, which can be supplied
to the Ant build with -DfamilyInfo=none (for example). The full
setting produces the output that we use on our private site; the list
setting produces a simple list of possible relatives in alphabetical order, with
relationships not identified, and the none setting excludes this information
altogether.
Also discussed with PS the decision about whether to have a single site with switchable languages on each page, or to have two completely separate sites. For the final output, the latter seems much preferable because it’s more intuitive, there’s no need to store a language choice and retrieve it on every page load, and it will be easier for search. However, I will initially develop the former, because it’s quicker and easier for testing and debugging, and then build the two separate sites from it.
On Friday, made much more progress with rewriting the site build code, and now it’s generating all the alphabetical listing pages, and the place hierarchy pages, although there are lots of issues still to work out with those. A major issue is the fact that captions in the old site were plain strings, so they could be concatted and so on, and were passed as string variables, but now they’re combinations of spans in both languages with strings, so have to be treated as sequences of items.