encoding of lacunas and omissions across block-level boundaries

[arlogriffiths]

I am copying from the closed issue #2 in erc-dharma/aditia-pdh.

@arlogriffiths wrote — this concerns EGC 5.5.11.2 and ​​5.5.12.2

• our encoding rules, if followed correctly, ought in principle to ensure tags for lacunae and omission always come in pairs
• ought we perhaps to change the rules of encoding in the EGC to require explicit matches between start and end tags with @xml:id and @match or @corresp?
• is it possible to have the schema control whether every such tag has a corresponding tag?

@ajaniak responded — I have pushed very quick schematron. Two for lacunae and two for omission. Indeed without real identifier, it is incertain to check that the matching pattern is respected since I can only discriminate based on the witness. So, I have also added a rule comparing the number of lacunaStart and lacunaEnd; as Michaël as done to underline the issue in the file. I will try to figure a more restrictive rule.

Michael: you may resume the discussion with Axelle and propose steps to be taken after finishing with urgent work for Aditia's dissertation.

[arlogriffiths]

at the moment, with Axelle's schematron still in place, several editions are massively invalid. see screenshot.

I would prefer the relevant lines of code to be hidden and perhaps to be unhidden later if and when we change the encoding rules.

[ajaniak]

The message just says that something in your encoding is wrong: one of your omissionStart isn't closed. I have deleted those rules.

[arlogriffiths]

thanks. well let me first check those cases then and try to remove the errors. if other files don't massively invalidate, we may keep the rule in place.