As per request, BBT charmap table

[retorquere]

I'm not entirely sure that

1. you want all of these characters, or
2. my assumption that you do math-mode by wrapping each individual construct with '$...$' is true

but here it is (in response to https://forums.zotero.org/discussion/59190/accented-character-u/#Item_12)

[adam3smith]

@dstillman your opinion on this would be useful -- this would definitely improve import/export, but we're talking about 9.5k additional lines in the bibtex translator, making it about 5* its current size. Is that a concern?

[dstillman]

But this is still only relevant for non-UTF-8 import/export, right? I just continue to care very little about that. People should use an implementation that supports UTF-8. Reproducing all of Unicode seems incredibly dumb.

(As for the impact of this, I don't know. It might be fine, or it might slow down loading of the translator, and it's certainly a lot of data to send down to every Zotero client whenever the translator is updated and to the connector for any non-client save that uses BibTeX (once every 24 hours, though we should do better caching). And it certainly will make the translator more annoying to edit.)

[adam3smith]

That's right. But BibTeX is really a non- (or pre-) UTF8 format, which is why the majority of straight bibtex implementations don't work with utf-8 and certainly don't assume it. Should people be using biblatex instead? sure. But I'm not seeing that as widely as one might expect.

[dstillman]

Should people be using biblatex instead? sure. But I'm not seeing that as widely as one might expect.

More people might, though, if tools stopped catering to the now-ridiculous demands of a pre-Unicode format…

[retorquere]

I wouldn't fault zotero for skipping this table because I think Dan's assumption that this is a solved problem is a valid argument, but a) I put the pull request here because I was requested to do so, and b) BBT users don't seem to think it's dumb; as @adam3smith points out, a lot of places still demand bibtex and the author gets zero say, and when people submit error reports I also get their BBT settings and even biblatex users have this translation active quite often.

[zuphilip]

How about taking only portion of this list? I think that the correct spelling of author names is more important than all the mathematical symbols, which can be then also added manually. Here is a rough statistic about this PR:

• LATIN -> 777 lines
• GREEK -> 268 lines
• CYRILLIC -> 762 lines
• Rest -> 7645 lines

(The numbers are simply search results for these terms and maybe only rough estimates.)

[dstillman]

That seems reasonable.