Note: We are a consortia of 249 public libraries on a centralized union
catalog, with 7.8
million items, 1.1 million patrons, and a FY2003 circulation of 15
million.  This has a
dramatic impact on any technical solution we make use of.

> What programming-type projects have you undertaken that really had
> some benefit in that context?
> (Alternatively, what projects would you like to undertake if the time
> and/or skills were available?)

My first library project was to a create a statistical reporting system
to alleviate some of the
inadequacies in our existing system.  We ended up dumping our database
into MySQL
every night (our existing reporting system was already segregated and
working on day-old
data in a separate environment because of performance issues), and used
Perl's DBI and
CGI modules for creating a web-based frontend where users can create,
schedule, and
download reports (which are made pretty with Spreadsheet::WriteExcel).

Since then we have done a lot of data work with MARC::Record, for the
purpose of
migrating other libraries into our system, and for cleaning up our
database (looking for
duplicate records), and have made much use of Perl for system
administration.

We've filled in a lot of gaps that our vendor has refused to address.
We created a webpage
for uploading "standalone"  transactions (for bookmobiles and when the
network is down)
so that libraries could get instant feedback on errors, and eventually
created a friendlier
and prettier standalone client (using Tcl/Tk of all things) that was
actually stable and could
manage the uploads itself if brought into a network.

We created a web-based management interface so that libraries could
manage their own
closed dates, login PIN's, and override codes, and not have to go
through a few individuals.

We provided a MySQL instance for 3rd party software requiring access to
our patron DB,
since our vendor didn't support the SIP2 protocol at the time.

We replaced a component of our ILS to allow for randomization when
fulfilling holds, since
our ILS was preferring the last copy added to the catalog for any given
title, and this
became very apparent when we migrated a single library system which had
a really good
collection--they were getting an abnormally high number of hold
requests.  The vendor
didn't think it was "feasible" to fix this.

We're considering in-house development of our own ILS (which would be
open sourced),
since there isn't a single vendor out there that meets all of our
needs, and it is very irksome
to be treated as a "minority customer" with needs so unique that they
are not worth
catering to.

--
Jason Etheridge, PINES System Support Specialist
Office: 404-982-3488