I don’t really understand what do you mean by old release of citeproc,
mods, csl …
I mean that both CSL and the citeproc-xsl code have been a moving
target. We made some changes to CSL since the last citeproc release,
and I have not entirely finished changes to the XSL code that are
primarily:
- fixing bugs
- changing from MODS to an RDF internal model
- updating to new CSL version
So it’s just a difficult time to be debugging the old (0.7.1) release.
E.g. if I don’t now have time to fix the SVN version, I don’t have time
to fix an outdated version.
I feel bad about this, but I just have too much going on now. I hope
some time will open up soon so that we can finally decide on CSL 1.0
and document it better, and that I can then update my XSL to properly
reflect it.
I’m currently using the citeproc 0.7.1. In addition, I tested citeproc
with various combinations of mods, csl files extensively. But to be a
bit more clear, here’s a summary of what I tested :
- I processed citeproc 0.7.1 with valid MODS document (taken from the
LOC site examples) and citeproc provided csl . => do not get proper
results unless is one of example documents provided with citeproc
download
But I’m not clear on what you mean by “improper results” and how you
know that? In other words, you say that the title wasn’t italicized,
but are you sure citeproc is using the template (book, article, etc.)
you think it should use?
- I reprocessed the citeproc provided examples (docbook-test.xml) with
the csl from the xbiblio svn.
If you use the SVN version, you will see rendering problems, but things
should come close to working. Note, though, that you need to use the
RDF/XML, and I have not written a MODS input driver.
=> do not get proper results unless I use the csl version of the one
of csl documents provided with citeproc download
That’s correct, because CSL has changed. You need to use the CSL files
in the SVN.
=> was surprised to observe the same results whatever the csl from
xbiblio used (apa-en, chicago, harvard). I assume a default csl is
used. But which one and most of all why?
The apa style is default. Why not?
So, finally the only case which I’m able to operate citeproc is with
the example (docbook-test.xml) with the initial citeproc 0.7.1
package.
As soon I try to change any parameter it doesn’t run accurately (or
even not at all).
Are all these problems raising from different schemas in use?
i.e.
- citeproc seems not to use the MODS schema from Loc gov site
- citeproc seems to have different csl.rnc schema than the one in
xbiblio SVN
Yes, that’s the problem.
Then I would have only two questions.
a) What I have to do exactly to operate the citeproc with a validat
MODS-3 xml document and csl definitions taken from xbiblio svn?
I would appriciate using the regular MODS-3 schema and the csl styles
defined in xbiblio csl - seems they are newer and better documented.
For the short-term, the options are not attractive. You need to convert
the MODS to the RDF/XML citeproc now expects. You can either do that in
two steps, and write an input driver that does it in one. I have a
MODS2RDF stylesheet in SVN, though I’ve not tested it with citeproc.
Finally, even once you get it working, there will be some bugs until I
have time to fix them. I’m sorry about that; I’ve just been too busy
the past year.
b) In case not possible with citeproc 0.7.1 please let me know if
there is other way other then Zotero to get the output?
Right now, there is no other option that does not involve some coding.
For example, you could take the Zotero JS code and modify it to use
your own data source.
Let me give you a rough roadmap I would like to see:
- mid-June: decide on how we call CSL 1.0 (Simon will have some time
then)
- mid-July: finish CSL 1.0 and freeze it
- early-September: update citeproc-xsl to 1.0 status
As you can see, you’re just coming here at a bad time. But things will
get better again. In fact, I may end up updating the XSL code much
sooner, depending on how my own publications go 
Bruce