This suggests to do the translation in a different direction or
(which I don't like) make 3 TEI files (for each "orth").
There may be a misunderstanding here -- I wasn't suggesting anything,
merely attempting to report what I saw. And I definitely wasn't
suggesting splitting these entries. Note also that a simpler
description would only have "a" for "adjective", and adding more
details should not be a reason for splitting entries. In other words,
I do not criticize what you've done.
My goal is to have the same information in the TEI file at the end
as in the CSV. I think, backwards-compatible must be possible. But
XML is not my favorit. I see a huge XML overhead in the XML TEI
format, which needs massive memory to phrase. What's interesting
about XML is that you can use css and xsl to display things in the
No pressure! :-) Providing a nice open-source dictionary is
definitely enough, and you are very welcome to do that.
Oh, a misunderstanding again. When I talked about semicolons and used
curly brackets, I was talking about rendering the entries in plain
text, after XSL conversion. Apologies for not making that clearer.
And by "gram.desc", I only meant that this is where the grammatical
description would be placed.