This has been an eventful day. I learned that I have mis-implemented
a large and difficult section of the citeproc-js disambiguation
routines:
http://forums.zotero.org/discussion/7457/should-there-be-a-no-givenname-disambiguation-default-style/
That thread contrasts with the detailed discussion here:
http://sourceforge.net/mailarchive/forum.php?forum_name=xbiblio-devel&max_rows=25&style=threaded&viewmonth=200903&viewday=12
It’s good to catch this at this point, rather than discovering it
after release, with everyone and his dog complaining about broken
styles. But this has probably cost a month or more of design and
coding effort, and it highlights a weakness in the current design
process that might be fixable.
The test suite helps to remove the ambiguity from design decisions,
but we’re not yet getting much benefit from them, because they receive
only sparse review. We can’t look to the user community for that.
Authors have only intermittent, even one-off exposure to a given
style, and are often working under deadline pressure to boot. It’s
not reasonable to look to them for review of subtle effects produced
by particular style options.
So I’m wondering, is it possible to involve publishers themselves in
CSL spec review? They are repeat players with respect to their own
style, they have staff who know it cold, and they have an strong
incentive to reduce their own costs through more efficient manuscript
preparation. It seems like a link-up just waiting to happen.
…
So I’m wondering, is it possible to involve publishers themselves in
CSL spec review? They are repeat players with respect to their own
style, they have staff who know it cold, and they have an strong
incentive to reduce their own costs through more efficient manuscript
preparation. It seems like a link-up just waiting to happen.
Ideally, publishers would develop and host their own styles, and so
would have an investment in the success of CSL. We’ve started to see
that interest with small journals, but I don’t really have a clue how
to involve larger ones. Being able to comment on the test suite
requires a unique mix of skills, unfortunately; knowledge of XML in
general and CSL in particular, as well as of the domain itself.
Perhaps one possibility is to use the 0.9 release as an opportunity to
raise the public profile of this work and to solicit comments from
organizations like Chicago, APA, etc.?
Realistically, what would that take? Probably a new, sexy (new CSS,
logo, etc.), streamlined (ditch some of the no-longer relevant stuff
and focus only on CSL), and rather more extensive web site. Among the
content might be a page dedicated to the test suite, complete with
previews and such. We also need the spec*, and tutorial.
But I don’t think you and I alone can or should do all this. We need some help.
Bruce
- I’ve been hoping the spec can be automatically generated from the
schema comments.
Convincing journal editorial boards to review and test piece of code
doesn’t seem very likely nor very efficient proposition. My guess will
be that it is best to contact the people who write the online article
submission software. From what I see a lot of biomedical journals use
online submission systems designed by eJournalPress:
http://www.ejpress.com/index.shtml
Good point.
I should say that I’m unlikely personally to go through the trouble of
contacting people on this.
Thinking grander, it’d probably make sense to consider journal needs
in designing the online CSL Wizard I’ve been promoting. Imagine a new
journal simply being able to choose a style from the wizard, and get
the style they want, plus (probably a link to) human documentation.
Bruce
A first step toward appealing to institutions that maintain the
citations end of publishing standards would probably be to prepare a
receptacle for receiving money (either a university-based center or a
non-profit vehicle). Not that anyone is going to get rich off of
this, but it would help to signal that CSL will persist beyond the
enthusiasm of a few individuals.
You’re not the first to mention this idea, and it makes some sense in
the abstract. I just really don’t know what such a thing should look
like, how to bring it about, and how to avoid a time sink.
Bruce