[Koha] Re: marc and stuff!

[Tonnesen Steve]

On Wed, 23 May 2001, Chris Cormack wrote:

> > we              may be at that point someday ... I guess the biggest question 
> > is  how to load the marc records into the database can you point me  
> > somewhere for this?
> > thanks for all the advise! And I love the way the software works....
> > 
> Ok Steve Tonnessen who is on this list has been working on this and I think
> he has a working marc importer running.
> Ill cc him in this reply.


I have a couple of schools who have been downloading marc records for
their books and importing then one at a time.  They check my automatic
conversion of the marc records and make whatever modifications they need.
Since the site they have been getting marc records from is Canadian, they
have needed to modify the ISBN to the canadian publication's ISBN in every
case anyway.

I've also started working on Z39.50 support.  Just signed the contract for
access to the National Library of Canada's Z39.50 services today.  I've
got a perl module called Zeta working for Z39.50 searches, but it only
seems to return SUTRS format (unstructured text) and not MARC format
results.  It seems that Net::Z3950 is a better module, but I haven't got
it working yet.  Also the YAZ toolkit also has a perl-based Z39.50 server
that would really be nice to tie into Koha.

LCCN should not be just overwritten on the dewey field, it should have a
new field created for it.  I guess the hitch with this is reworking the
scripts for entering and editing books...


Here comes some long rambling on my part...

I think Koha really needs to be moved to a system that is at least partly
based on Marc.  I think that the records for each bibliography should be
in MARC format, and that changes can be made either to the Koha Mysql
tables, or the underlying MARC records and changes will be made in both
places.  The way that Koha stores records currently (broken into biblio
and biblioitems) might make this a little difficult, but we could maybe
associate particular MARC records with biblioitems, I guess, and then just
have groups of MARC records associated with one biblio, as is done now.

I'm thinking maybe about having one module that is used for modifying
records.  Something like:

modifybibitem($env, $bibitemnumber, "author=Doe, Jane");


would change the author for biblioitem $bibitemnumber to Jane Doe.  The
modifybibitem routine would be responsible for modifying both the MARC
record, and the MySQL tables.

This all gets a little tricky when you consider that the MARC record
stores a fair amount more information than the MySQL tables, for example,
if a particular MARC record has a subject added entry like:

650  a Extinct animals
     x Encyclopedias, Juvenile.

The corresponding entry in the MySQL bibliosubjects table would be:

   Extinct animals -- Encyclopedias, Juvenile

Now if I edit the bibliosubjects entry to read:

   Extinct animals -- Encyclopedias

It would be difficult to determine how to modify the MARC record
automatically in a generic way.  If the entry was changed to:

   Extinct animals -- dinosaurs

It becomes even more difficult to incorporate the change back into the
MARC record.

I suppose the web interface could pop up a form saying something like,

---

Need more information for MARC record modification: 

650   a Extinct animals
     __ Encyclopedias

---

with an 'x' as a suggestion for the text entry box.


Steve Tonnesen
Coast Mountains School District