keys and ids (again)

CSL Development
1

OK, another design question:

Right now, citeproc expects to associate a citation with mods:mods/@ID.
In other words, identifier of the mods document happens exclusively by
doing:

<mods ID=“doe99” …

Is it OK for me to keep this convention for now? It’s easy for
Matthias to support this, of course, since he’s generating the MODS
records from the DB. In the context of a mult-user XML DB, though, it
becomes a little more tricky I think.

Bruce

2

Right now, citeproc expects to associate a citation with
mods:mods/@ID. In other words, identifier of the mods document
happens exclusively by doing:

<mods ID=“doe99” …

[ I don’t think this matters in the context of citeproc but refbase
[ allows users to cite a record not only by its cite key but also by
[ its database serial number. We have users that don’t use any cite
[ keys at all and refer to records solely by their database serial
[ numbers. If there’s no cite key for a given record (and user)
[ refbase will ommit the mods:mods/@ID attribute completely and leave
[ the generation of generic cite keys to bibutils when converting MODS
[ records to Bibtex/Endnote/RIS. Anyhow, if a user uses citeproc for
[ document processing and refbase as a DB backend than he’ll also use
[ cite keys.

Is it OK for me to keep this convention for now?

I think so, yes.

It’s easy for Matthias to support this, of course, since he’s
generating the MODS records from the DB. In the context of a
mult-user XML DB, though, it becomes a little more tricky I think.

In addition to ‘<mods ID=“doe99”…’ refbase does also output the
cite key as:

doe99

AFAIK, this is also recognized by bibutils. I’m not sure if bibutils
preferes one over the other, though.

Matthias

3

Yeah, Chris and I decided better to be safe and duplicate the content
for now. Citeproc doesn’t yet look at the identifier element though.

Bruce