The lack of vibrant community is what points to Jakarta Commons as the more appropriate place, a place where a community could grow for JCS. I'd rather see JCS as a full sub-project, but without a community to support the software, it would be misplaced as such.

Henri Yandell wrote:
So your preference, as the development-community of JCS, is for a
top-level-jakarta project, ie) at the log4j level?

If so, we can take that up with the PMC and see what views there are. As
the development community, your (and James) views count a lot, though the
smallness of community is the worrying thing.

Hen

On Thu, 4 Dec 2003, Aaron Smuts wrote:


The core of JCS is ready for a release.

The project is basically a hub for 4 types of plugins, or what are
called auxiliaries in JCS: memory, disk, lateral distribution, and
remote sever.  It requires that you use a memory plugin, but the others
are optional.

For each type of plugin there is an efficient implementation that people
are using.  These include: LRU memory manager, indexed disk cache, TCp
lateral distribution, and RMI remote server.

There are experimental versions of each type of plugin in an
experimental source directory: a b-tree disk cache, a database disk
cache, a javagroups lateral, a MRU memory manager, and others.

The core of JCS is then the hub and these 4 non-experimental plugins.
Currently there is only one small bug in the lateral cache recovery
process, that I will fix very soon.

There are additional features that are mostly extensions of the plugins.
I wanted to clean up the group handling features, but this is not
crucial.  I wanted to add run time defragmentation to the indexed disk
cache.  I also want to implement clustering on the remote server.
Basically, this will involve hooking up remote servers via the TCP
lateral cache.  All that has to be done is to work out a way to prevent
circular calls for there to be clustering.  The client can already fail
over.

I'm not sure what all the levels are called, but if we put JCS at the
level of log4j, I guess as a jakarta subproject, and then issue a
release, we can find out what else people might want and some more
people may be interested in contributing.

JCS does not need an overhaul or any significant amount of work on the
core features.  Most conceivable future development will involve tuning,
bug fixes, improving configuration, creating sample applications, and
extension development.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to