Update: The instructions on installing Vufind have been moved to the Vufind Wiki. Please check there for the most up-to-date instructions.
It’s official…Andrew tagged Vufind 0.8 yesterday! There have been some very significant changes to how the index functions. First, the Java code to index has been included into the trunk, so indexing time is greatly reduced. Second, the full marc record is no longer stored on the file system, it’s in the index. Because of this change, you have to re-index the entire catalog to use 0.8.
So, now that 0.8 is tagged, I plan on breaking the Java importer to start refactoring with some of the changes that Robert Haschart from UVa has made to make the indexing engine a bit more flexible. If you’re using the SVN version of the code, I will be breaking it 😉
In other news, we’re very close to implementing a nightly build of the code. This will let folks who don’t want to be building their own software every time Andrew or I break the code a place to get functional code without being Subversion gurus or chasing down Ant dependencies.
So what’s on the horizon (at least for the marc indexing mechanism)? I’m going to be refactoring most of the code into POJOs instead of the spaghetti code it is now (mostly from Robert’s rewrites). The big enhancement is going to be logging (via log4j), the ability to point the indexer to a directory of marc files (instead of one big file), and more “human” error handling than just the stack trace.