Sencoten 2023-09-04 to 2023-09-08
to : Martin Holmes
Minutes: 380
On Monday, following instructions from SK arising out of a meeting with the
community, I made a number of updates to the build process, beginning with filtering
out entries marked as EXCLUDE. I also changed the rendering of Sen/Eng entries substantially,
and reworked the insertion of markers; I now realize that the decision to put markers in
<fo:block> elements before the entries they refer to was probably the
main cause of the errors in the forme works, because although the markers are empty, they
constitute blocks and therefore were possibly being wrapped to the next page, separating
them from the entries they apply to, despite the keep attributes I was applying.
On Thursday, after more discussion, revisited the merging of entries to allow the handling of cases where the same Sencoten headword might be listed in entries with different roots. I ended up creating a pre-processing step which creates unique but predictable ids for each entry based on the headword and a counter, and then modified the processing to use these entries instead of generated ids for page-referencing. This enables us to handle the many duplicate entries more elegantly without obscuring them; if we don’t obscure them, the language team will be able to identify the duplicates more easily. Then fixed another duplicate page-number bug, and had some discussion with SK about whether we should collapse entries in the root-based index.
More brief discussion on Friday, but no further changes needed for now.