Fredrik and I have been working on the detailed resegmentation PRs as per TC resolution for the comment 038. This was made possible by the TCs decision to move module and cusom data including notes out of segment and lower structural levels.
The resegmentation is allowed or prohibited explicitly using the canResegment flag that is recursively inherited through file, group, unit, to segment, and can be explictly set on all those levels. The default is "yes"
Details how this flag drives the resegmentation behavior are in the latest editor's draft
The detailed PRs are still WIP but will follow the following principles.
When joining segments
with different states the resulting state is the "earlier" state
with different approved flag the resulting is "no"
with different xml:space the resulting is "preserve"
with different translate flag the resulting is "yes", unless the modifier uses
with different directionality control characters must be used to control the flow
splitting segments with targets will be allowed with warning
who wants to prevent resegmentation with targets must use the canResegment flag explicitly to prevent it
we won't provide too detailed PRs for splitting spanning codes and markers
there will be a MUST for replacing spanning codes/markers with sc/ec or sm/em pair
and a may on using the proper spans after joining if possible
But this is all already described very well in the inline section, so we will use pointers to the pc mrk <-> sc/ec sm/em transformations
Fredrik will follow up on this e-mail once the detailed PRs are in the spec.
Any input on the above principles and caverats re the detailed elaboration is welcome
The final solution should be available in the editor's draft by the end of this weekend and will be presented in the next TC on 20th August
LRC | CNGL | LT-Web | CSIS
University of Limerick, Ireland