[Koha] Goods and bads...

[Ernesto Silva]

.. hi folks.

I´ve been away from this list for some time (years?). I installed Koha
at Universidad ORT Uruguay in 2004-2005 with Andrés Tarallo
(http://biblio.ort.edu.uy  = 61000 biblios) and adapted the system to
feet our needs (version 2.2 or 2.4). Before Koha we had MicroISIS.

I still work at ORT University but I also work at CES, which is the
secondary schools headmaster organization at country level.... well,
perhaps it's no so big, Uruguay is not a big country, only 3 million
people.

We have 270 schools with a library on each one, a big library (I don't
know the size yet) and I believe there are a few other small libraries
around.
Much of the catalogs are in MicroISIS and OpenBiblio and we want to
migrate to a fully centered web based architecture, our schools are in
the process of being digitally interconnected.

So of course my first alternative is Koha as this is the platform I
know best, but... (there's alway a 'but') the 2.2 version of Koha had
a lot of good code and a few very, very bad pieces, not that I'm a
good programmer.

I feel this new project will also need some level of code alteration
and I'm worried about the actual code status.
I haven't got the time to setup a new Koha installation, and I have
some questions perhaps some of you can answer, sorry for my laziness
but I really don't have the time to read the code:

-How is the code today? Does it still have library specific hard-coded pieces?
-Are Authorities and Serials working well?
-How about acquisitions?
-How does Zebra performs with complex queries, say, 5 to 10 terms ( A
and B and not C or D... etc.)


Some history of my work with Koha.
I've modified a lot the Koha version a got from Internet because ORT
University wanted some processes to work in a certain way not fully
supported by Koha. I contributed to Koha source with a few patches and
I also made a fully new search system using MySQL full text search
capabilities. I "invented" a query language (reinventing the wheel?)
and a translator to SQL. With this search improvements I've managed to
reduce the complex searches delays from minutes to seconds, say 15
minutes to 10 or 20 seconds.
We hired Jean Pierre Ducassou, he was the one who finished the search
system and developed a new record display system based on XML+XSL+XSLT
which we still don't have online but will have in a few months. We
need to alter some MARC data to unify criteria here at ORT.
Of course this code is available for free to everyone who wants it,
just ask for it.


My regards to the list and Paul Paulain.

Ernesto Silva.