[Koha] Some more ideas on cataloguing and Z39.50 searches

Manos PETRIDIS egpetridis at gmail.com
Tue Dec 31 08:59:51 NZDT 2013


Allow me please to share a couple of ideas I had on zebra cataloguing:

 

a) batch Z39.50 searches.


As a non-librarian, I rely on the bibliographical entries of others, and
I've found that most of my cataloguing time is spent waiting for Z39.50
search results. It may have to do with the number of targets I normally use,
or with the fact that I sometimes look for titles that are carried by some
libraries only, or none at all, it seems though that this interactive
process is counter-productive. I wonder therefore if you too find batch
Z39.50 searches an interesting idea: The user would be able to enter ISBNs
(or other similar search terms) in a table, and koha would perform searches
against Z39.50 targets in batch mode, either in the background, or
semi-interactively, i.e. by taking-up a browser window. Search elements
already processed, would be removed from the table, so that the user would
be aware when the search is complete for each item. The user would also be
able to add more search terms at the bottom of the list, even when the batch
search is active.  Search results would be stored in an appropriate
container (the "reservoir" perhaps) and koha cataloguing
(cataloguing/addbooks.pl ?) would be amended to allow for searches against
this container too. 

 

b )  use reservoir to auto-tune Z39.50 targets' proprieties. 

 

I've noticed that the koha entity named "reservoir" holds results of past
Z39.50 searches against remote targets. I wonder if these entries could be
used to auto-tune the priority of Z39.50 servers, so that one may be able to
first search targets that consistently provide results. 

As it regards priorities, it would be useful to have two sets of priority
attributes instead of one, or a scheme could be devised so that the user
would be able to split the value range available from the current single
column "rank", to denote different groups/search phases: targets to always
search first, then known quality targets to search when the first have
failed, then the rest, for example.

Those entries could be used to check/uncheck en-masse the relevant targets
in z3950_search.pl. The user would perform a search against the standard
targets (just as in  the current implementation where only the targets
defined as checked are pre-selected). Then, should this first search bear no
results, it would be easy to select the additional servers/targets with a
single click. Should even that fail, the user could  select all, and try
again. 

kind regards,

Manos PETRIDIS
Athens, Greece

 

 

 

 



More information about the Koha mailing list