In my opinion, Xindice suffers from a number of problems. Probably the most significant is that the code, especially in the core database, is overly complex and suffers from a number of pre-mature optimizations that I really wish weren't there (in particular the compressed DOM). This complexity has prevented evolution of the core functionality as the core is simply too hard to change for developers unfamiliar with the code.
I attribute this problem to the fact that Xindice was originally intended as a commercial product and was engineered to eventually be a viable product in that space. Unfortunately, the commercial support is no longer there, the original developers (Tom and myself) haven't been able to contribute and that complexity is now a major barrier to new developers being able to modify the code.
I've struggled for a long time with what to do about this. With 1.1 I started trying to simplify things like throwing out CORBA and the custom application server that we used, but I never really figured out what to do with the database core. I explored various things from writing unit tests to experimenting with a complete rewrite starting from the bottom up, but I just couldn't commit the time required to make any real progress and eventually got pulled away completely.
Today I still struggle with it, and I still have no answers for the problem.
All I know is that Xindice needs new developers and those developers need to do whatever is necessary to make the system their own. There have been a lot of lessons learned about what an XML database needs to do. Maybe Xindice 1.0 needs to be called a prototype and an effort needs to be started to just start over. The project framework is here, there are a lot of talented developers who contribute to Apache projects and there is much to learn from the current system.
On Dec 3, 2003, at 2:27 PM, Antoine Lévy-Lambert wrote:
Hi Kevin,
on the "open source market" for xml databases there are Xindice and exist.
I think that XML databases are a nice thing and that it is good that the ASF
has its own.
Do you think that exist is better than Xindice, or are you just worried that
you do not have enough committers on the project ?
I am not saying I want to become one, but maybe if you are patient there
could be some joining your group.
I need to test xindice for a project I am doing for my company.
Cheers,
Antoine
-----Ursprüngliche Nachricht----- Von: news [mailto:[EMAIL PROTECTED] Auftrag von Kevin O'Neill Gesendet: Mittwoch, 3. Dezember 2003 22:03 An: [EMAIL PROTECTED] Betreff: Should Xindice be retired?
There has been very little development done on Xindice over the last three
months and it may be the case that we are doing users a disservice by not
letting users know that the project is effectively inactive.
Is it time to retire the project?
-k.
(Hard hat and flame suite on ;))
Kimbro Staken Software, Consulting and Writing http://www.xmldatabases.org/ Apache Xindice native XML database http://xml.apache.org/xindice XML:DB Initiative http://www.xmldb.org
