Implements DeltaV specification (versioning including branching & merging)
Implements WebDAV ACL specification
Implements WebDAV BIND specification
Supports transactions and locking
DASL support for searching
Pluggable stores for backend storage of content and metadata (may make it possible to add a type system)
WebDAV command line client
Eventing and rendering framework (Projector project)
WebDAV Construction Kit (high level WebDAV based API)
I setup my multi-user test harness to run against Slide using a JDBC configuration inside Tomcat.
My previous tests with Slide 1.0.16 rasied deadlocks within the first few iterations, the results of using Slide 2.1 are detailed below.
Using MySQL as the database I ran 13 users all doing various basic WebDAV operations against the server and it ran for 30 minutes without any errors from the Slide or database server.
I then ran the same test using an Oracle database and the same Tomcat server and again it ran for 30 minutes without any errors from the Slide or database server.
In terms of performance the time taken to complete the operations is pretty much on par with version 1.0.16 and our own repository. It may be a little slower than ours but it is supporting a lot more features.
There is a link on the Slide home page to a list of committers and contributors. It lists 15 active committers and 15 inactive committers, plus a bunch of people that have contributed to the project over it's lifetime.
Of those listed Oliver Zeigermann, James Mason and Warwick Burrows seem to be the most active on the slide user newsgroup.
There was one conversation regarding JSR170 support, they don't preclude it but it also doesn't seem to be a high priority either i.e. it is not even mentioned in the Wiki above.
Reasons not to use Slide
Architecture is a bit disorganised, to sort this out is one of the main goals of Slide 3 (see Wiki).
Server API is extremely hard to implement against, even the main Slide committer says this.
No diff between versions
In Slide 1.0.16 they used exceptions to determine the runtime code path, as the architecture is the same in 2.1 I presume this is still the case, this is not a very good programming practice - and can adversely affect performance.
Documentation is appauling so trying to get anywhere with customisation tasks will be hard.
I think Slide has come a long way since version 1.0.16 (apart from the documentation). It seems all the problems we identified have been addressed and whole new bunch of features have been added, there is also a lot more activity in the project itself and on the newsgroups.
However, it's hard to forget the first impression and although I have run the tests that failed before successfully you can't help but think there is another major problem lurking!
However, there are people around to help now and there seems to be a future for the project.