Sebastian writes: "After ages and ages of "real-soon-now" comments, we finally have our ZAP! Apache module ready for public release. ZAP! allows you to build WWW-based Z39.50 clients by filling in "templates" for each page in your interface. It is simple to do basic things, and yet it is possible to build quite advanced gateways. In the Apache mode, it can be extremely efficient, but it also runs as a conventional CGI script. It is freeware, but we have certain commercial options available." Nice... tried it on Yale's ORBIS using the demo page and it seems to work pretty clean.
Sebastian writes: "Index Data has just made its first release of a Z39.50 Server module for Perl. It provides a pretty simple API which hides most of the complexities of Z39.50 and network programming in general, so all you have to do is provide a bit of code to interface to your resource. You can draw on all the usual Perl tools to talk to back-end databases, create response records in XML, MARC, etc." Built using YAZ by the people who brought us YAZ to begin with... definitely worth a close look.
from freshmeat: "Significant API changes, involving provision of a unified settings system for passing parameters, and several small tweaks. Applications will need simple modifications and recompilation. Also features improved weighting, a forking network server, and a few bugfixes. Note that the Java, Perl and Python bindings do not function in this version - do not upgrade to this version if you require these." For more see open.muscat.com.
from freshmeat: "The purpose of mifluz is to provide a C++ library to build and query a full text inverted index. It is dynamically updatable, scalable (up to 1Tb indexes), uses a controlled amount of memory, shares index files and memory cache among processes or threads and compresses index files to 50% of the raw data. The structure of the index is configurable at runtime and allows inclusion of relevance ranking information. The query functions do not require to load all the occurences of a searched term. They consume very few resources and many searches can be run in parallel.
[changes include] Integration into the GNU project, complete re-architecture of the inverted index structure, major performance enhancements, and more." All this apparently from the principal author of Catalog...
Rob S. writes: "Cheshire is an OSS (Berkeley style licence) z39.50 search engine/server in active development. Also being developed is an extension
to Mozilla for the z39.50 protocol."
from freshmeat: "Distributed searching across several machines, improved writable databases, the ability to automatically select elite terms for performing queries, small API changes, and many bugfixes." see open.muscat.com for more.
Both the Ecila (French) search engine codebase, catalog-1.0, and the U.K. eLib endproduct ROADS v2+ are open source and increasingly used tools for building web-based catalogs a la Yahoo. Some eLib folks have explicitly turned to open source as a way to keep formerly well-funded projects going (see the press release describing this decision). Open source as "exit strategy" isn't terrifically sustainable, but a step in the right direction nonetheless.