I now have a basic XSLT transform working, which does the hard stuff -- it creates tagging which can be used to render in a human-readable way multiple coloured underlines for nested codes. I still have to do the control panel and work on the CSS, but the hard bit is done.
I now have Oxygen able to insert linked anchors and prompt for analysis codes. This is coming along nicely. I think we should have a pilot version ready by the end of next week. We'll need to write a fairly sophisticated two-step XSLT transformation to create an HTML version that's useful for reading/proofing, but there's nothing there that we haven't done before.
The requirement to include components such as spaces and square brackets in the GET URL for the server is severely problematic, because the eXist httpclient:get function rejects them; if they're escaped using percent escapes, the Solr server doesn't like them, but if they're escaped using XML entities, eXist rejects them.
This may be gotten around by using httpclient:post, with httpclient:field elements; see the testSolr.xql file currently in the Graves app, where I've been doing some early testing.
I may actually need to go back to running my own local Solr and local eXist for this dev work, because it's too awkward having to edit stuff in the live site (which is the only way to call the real Solr instance).
I've made a start on creating the taxonomy, with one sample major category with subheadings, and I've now written some comprehensive ODD processing that not only builds these values into the schema, but also extracts and preprocesses the Schematron, for later speedy validation. I've also updated the ODD to add new things we need, and added some Schematron rules to constrain how anchors are going to be used. Talked at length with EC about this work. We will need some processing into HTML to enable easy comparison of two coded texts side-by-side, for reconciliation/checking purposes.
Only one interview left.
Got started on a plain statement of our principles, arising out of what we've learned, which may form the basis of our white-or-not-paper. Lots of input from various people, and it's beginning to take shape.
Did the keywords for the second of my assigned interviews; also noticed a common tendency to include leading or trailing spaces in name elements, so I added a Schematron rule for that, and fixed all existing instances.
Did DM's timesheet, which required Windows, which required updating, which took four hours in total. FFS.
Interviewed DJ, processed resulting audio.
Went through the first interview assigned to me and added topics in the keywords section of the header. Also added the required components for this to the schema, and handling for them in the Author Mode CSS. Then helped EC merge conflicting versions after we both edited the same file at the same time.