[bug] BibTeX key generation

When BibTeX.js removes the small words from the title to form that part of a BibTeX key, it apparently removes both the beginning space, so that a title like

Love on a Platter by Smith in 1980

ends up with

smith_loveplatter_1980

rather than the expected

smith_love_1980

It will even merge more than two words to make a reallylongbibtexkey if those words each have the small words in between them.

I'd be glad if this were fixed, and can't manage to do it myself.

--

and (if you have the inclination):

I'd also be glad to have the list of ignored small words (typically including prepositions, articles and conjunctions) expanded to include the articles of --at a start--- German, Spanish and French.

I suggest that the following might be added to the regex:

la/el/los/las/le/les/der/die/das/dem/den/des/

(plus perhaps the l' of the elided French article. But I expect that would take it's own pass, since it's not separated from its noun by a space.)

This is a change I can make for myself, but I think others might appreciate it as well. Of course a more ambitious expansion could include prepositions and conjunctions in those languages, a larger selection of English prepositions, and perhaps venture into other languages, but the above change is pretty simple and would take takes us a good deal further than we are now. Articles are any way more common as initial title words than the other small words.

Thanks
  • I forgot to add these articles to the small words list. I was reminded by the documentation for bibtool

    un/une/ein/eine/einen/eines/einer/unos/una/unas

    That should cover, as I said, French/German/Spanish articles.
  • I also see the bug with the "Love on a Platter by Smith in 1980" title exported to "smith_loveplatter_1980" key.

    And I would vote for adding the pruposed keywords to the list.
  • I may have a patch that fixes both of these things. I'll test some more and post it in the next few days.
  • edited June 5, 2010
    Patch submitted to Zotero-Dev for both of these issues. I noticed no bad side effects at all when exporting my fairly large library. It made about 15% of my BibTeX keys more friendly (300 of 1800). And by my stopwatch-benchmarking, it also speeds up BibTeX export by about 15%, even though it triples the number of articles and common more-grammatical-than-lexical prepositions it excludes from the BibTeX key.

    I'd be very glad if the patch could be considered for inclusion.

    http://groups.google.com/group/zotero-dev/t/10039e31bad0d5ce
Sign In or Register to comment.