Hello! I am evaluating Xindice for production use in a long-term project.
The criteria that I am using to evaluate competing software packages naturally
include flexibility, Open Source licensing, standards conformance, maturity
(kind of difficult for native XML databases) and long-term viability.
I am interested in the conclusions you will draw. Sounds good.
Although Xindice shows promise, I am concerned about a few things:
1) The web site does not seem to be zealously maintained. Early on, I
made the mistake of following one of the many links to www.xmldb.org. I learn from the mailing list archives that this site moved to
SourceForge at least 6 months ago, yet the many links on the Xindice
site have never been updated. The Bugzilla link in the main menu
is also broken.
True, but not necessarily a reason to panic.
2) Both the user and developer mailing lists show that traffic
that was fairly steady has dropped off since the first half of 2004. I know that this can be explained a number of ways (maturity stages,
dilettantes moving on to project du jour, etc.) and is not necessarily
a measurement of community involvement. For example, the number of
active contributors (bless you guys) is respectable.
True, but again not necessarily a reason to panic. Questions get answered eventually.
3) Other XML databases such as eXist and Tamino offer fledgling XQuery support. Some of the others also already provide WebDAV. Both of these are features in Xindice's road map that have not yet been developed.
Now to my question: As users or developers of Xindice, what do you feel is its current position in the landscape of native XML databases? How happy are you with its long-term prospects?
We are building a large-scale system too and this is a question we faced as well. It seems to me that in 2-3 years something will emerge in the open source community as the clear winner in the XML db world. Right now there just isn't. Our biggest problem is of course performance and none of the products we saw (including Xindice) perform well in terms of throughput and latency for queries.
Also from looking at major DB conference proceedings (VLDB, SIGMOD, etc) one sees whole sessions devoted to storing XML data, which is a sign that the topic is still immature.
So my short answer is: right now the state of XML dbs is chaotic. In 2-3 years something that performs reasonably well may emerge.
Do post the findings of your evaluation when you do it, should be interesting.
Thanks Eno
That is it and thanks for your time. Please do not regard my concerns as criticism. I focused on them instead of the positive features I see about Xindice (ASF, license, docs, contributors, etc.) because, like you, I am just trying to get the job done.
David Corpstein
--
Eno Thereska ------------------------------------------------- Carnegie Mellon University Department of Electrical and Computer Engineering Parallel Data Laboratory HH D-Level D2 5000 Forbes Ave Pittsburgh, PA 15213 Tel: 412-268-4266
