Currently, the BibTeX key generation always writes the name of
the author exactly as it is written in the author field.
However, people living in strange countries tend to have strange
characters in their names. I do not want to use accented
characters in BibTeX keys. There are a couple of reasons for this:
- First, it is simply inconvenient to type those in your editor while
writing a Latex document. Yes, auto-completion, copy-paste,
etc. do help with this, but I would rather do without.
- Second, JabRef is wonderful when you name your PDF files by
using BibTeX keys. PDF links are created automatically. However,
using strange characters in file names tends to be plain and
simply asking for trouble.
Thus I believe that JabRef should optionally convert BibTeX keys
to plain ASCII. It could simply change 'ä' to 'a', 'é' to 'e', etc. (I
do not know how to implement this easily in Java. However,
one could start by supporting this conversion for Latin-1
characters only. One could easily write a table of sensible
replacements for those 96 non-ASCII characters.)
In my opinion, this optional ASCII conversion could be
implemented by using a modifier in the key patterns. Perhaps
something like '[auth:lower:ascii]' could be used.
---
This brings us to the second problem with the BibTeX key
generation: the configuration of the key patterns.
It feels a bit stupid to force users to manually copy-paste the
same key pattern into 14 separate fields in the configuration
dialog. I can do that but I find it a bit hard to explain this when I
try to encourage other people to use JabRef.
And, if other people work with the same database but don't
configure all those patterns exactly as I did, the result will be a
mess.
Thus, I have two suggestions:
- First, it would be a lot easier if there was one "master" key
pattern. Entry specific patterns could be empty by default,
meaning that the master pattern is used instead.
- Second, these entry patterns could be stored in the BibTeX file.
This way everyone can simply open the database and use
JabRef, without any need to configure the key patterns. Perhaps
the values specified in the configuration dialog could be used as
defaults for newly created BibTeX databases?
Best regards,
Jukka Suomela
jukka.suomela at sammal.pp.fi
Logged In: YES
user_id=704995
Regarding the non-ascii characters in keys, try this in version 1.7 beta - it
does replace these characters with relevant ascii characters.
About the key generation setup, I never thought of having a "master" key
pattern, but this is a brilliant suggestion!
I'll have to think about the option of storing key patterns in the database.
I'm not sure how simple it would be to implement.
Logged In: YES
user_id=704995
An update: I am implementing your suggestion about a "master" pattern for
the 1.7 release. It is available in the recent 1.7 beta 2 release, although with a
bug causing it to require a restart of the program to register changes in the
master pattern. This will be fixed by the final 1.7 release.
Is it possible to transliterate the BibTeX key automatically?
The BibTeX key generator currently composes the key copying authors' names in the language of the entry. As it is for names with accents, it's much more convenient to use ASCII for the keys (and LaTeX labels, in general, IMHO). Where can I find info on how to achieve that?
Alessandro, this is a very old, closed, bug report. The key generator handles accented and other non-ASCII characters by replacement with ASCII characters deemed equivalent. If you have issues where this is not working properly, we'd be happy to look into it.