Re: [Refdb-users] using refdb with a course website

[Markus H. <mar...@mh...>]

Hi Matt,

Matt Price writes:
 > are the existing backends well-documented?  I haven't had a chance to
 > poke around in the source yet, but maybe if you can point me to the
 > main locations I can try to learn enough about them to figure out
 > where to start.  
 > 

Sorry, no. You'll have to wade through C code but that contains quite
some comments. The places to look at are the backend_xyz.c files, each
of which defines a backend (except the bibliography backend which is
split in two due to its complexity). There's essentially three functions being
called by getref, one that adds a header (if required), one that
formats the data proper, and one that adds a footer (if
required). backend.c defines the internal API which is the most
convenient way to get at the individual fields.
 
 > and maybe I can learn perl while I'm at it...  
 > 

The more I think about it, the more I'm convinced that custom backends
are best implemented on top of the Perl client module, using risx as
the data exchange format. So there's indeed a good reason to learn
Perl.

 > I think risx is probably the answer for me, unless DcBook is so
 > enormously powerful that it could generate files that could be easily
 > read by EndNote (I don't know much about docBook, but it does seem
 > very cool).  The target audience here is students and researchers in
 > the humanities -- that is, people even dumber than I am about
 > technical questions.  
 > 

?? If you're thinking about getting RefDB data into EndNote, then
plain ol' RIS should be the answer. RIS as exported by RefDB may need
some tweaking to get it into EndNote cleanly (as is the case if you go
the other way), but I would certainly look no further for this task.

 > OK, I'll look at the module.  But suppose (in the interim, while I'm
 > learning) I set up a database with the same structure as refdb's and
 > enter references from the web, or via FileMaker or something, using
 > only the main table.  Will it be straightforward to add these
 > references to RefDB at a later date?  
 > 

Keep in mind that the main table lacks the journal, the author/editor,
and the keyword information, all of which I consider essential.

 > And on a similar note:  I would *love* to have the refdb tables in
 > the same database as that which is maintained by SPIP, the CMS I'm
 > working with.  This would allow me to make cross-referencing tables
 > for SPIP to use when it generates static html pages for web display.
 > I suppose this kind of thing could be done even if I were using
 > separate databases, but I'd rather not have that hurdle if I can help
 > it.  Is there any in principle problem with this?  I noticed that
 > refdb wouldn't create the necessary tables in a pre-existing
 > database, and wondered if I could work around that.  
 > 

Please look at the empty.mysql.dump and empty.pgsql.dump.in scripts
shipped with RefDB. The main use of these scripts is to allow database
administrators that don't want to run an app they don't know to create
a RefDB database. However, if you run these scripts on existing
databases, they will create the RefDB-specific tables unless there are
name clashes. empty.mysql.dump should work out of the box, whereas
empty.pgsql.dump.in needs some modification as it attempts to create
the database itself. Just remove the offending code and off you go.

 > well, I'm NOOOOOTTTT a hacker, honestly, but I will try!

Great!

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de