Kill the (reverse lexixon) indexes ?

[Omikhleia]

Early on in the project, we introduced "indexes" that could be used to generate reversed lexicons (English-Sindarin, French-Sindarin, etc.)

• The result was never very convincing and was not used a lot (IMHO, but it seems to be also the opinion of others that the reversed lexicons were a bit messy).
• The scope of the project was moreover smaller (initially, mostly LotR, Ety and a few HoME) and the indexing was quite direct then. What to index and how to do it properly was never really the result of a general approach.
• When new words were added, the indexes were not maintained consistently...
• They are not even very consistent between languages (so English and French, as of today - the German version was even more broken, but it's behind us anyway now).
• The way these indexes are encore in TEI XML P3 is also not very satisfying (repeating glosses, being marked in a strange way, etc.)
• Most of the scripts for generating a reversed lexicon are pretty broken now (for what of them could be recovered), as the dictionary had several structural changes since then.

Still, we have plenty of theses indexes in entries, which make the XML hard to read and maintain.

I am suggesting getting rid of those indexes.

Pros:

• Having a simpler and smaller XML structure
• Getting rid of things of the past that didn't pass the test of time...

Cons:

• Loosing the ability to generate reversed lexicons, which despite being kind of broken, as mentioned above, could also be repaired... But that's a big effort checking all entries, encoding indexes in a more satisfying way, re-working scripts, etc. I am sincerely not sure it's worth doing...

Mitigating the cons: If it ever had to be re-done one day, I'd personally would prefer to start from scratch rather than suffering from the impediments of earlier attempts.

We can leave them for a while (they don't do any harm - they are ignored in the HTML and PDF outputs), still they bother me each time I have to change/update an entry.

[L-ND-L]

Mitigating the cons: If it ever had to be re-done one day, I'd personally would prefer to start from scratch rather than suffering from the impediments of earlier attempts.

Doesn't that answer the question? I'd be in favour of reverse English and French lexicons, but if the present state is so unsatisfactory that restarting from scratch seems easier to the person who initially coded it all, then I guess these indexes won't help much anybody else either.