I wouldn't mind starting to formulate a plan. Here are some suggestions. I'm somewhat inspired by the JCP, but probably everyone here (including myself) has problems with it. I think the overall idea of how to create a compatible API with compatibility standards can be done in an open/open-source manner.

0. A minor tweaking of adminstration would be good:
1. refector the project - Rationale this allows people to contribute to mailet and be mailet project people without being JAMES people. Amoung other things.
a. email.apache.org/james
b. email.apache.org/mailet (associated lists)
2. Discuss the overlying contract for Mailet containers.
3. Creation of a specification similar to EJBs which lets mail container folks know what the "should" "shouldn't" "shall not"s are.
4. Creation of a JUnit set which tests compatibility of mailet containers. A TCK if you will.
5. Require TCK to dictate API to dictate Spec (meaning nothing makes it into the spec unless it has first been proposed as a test and hence an API) - but we should have some "seed" code. I think this "test-first" "code-based" design will cut some of the natural rhetoric out of the process of writing a spec.
6. Develop two (initial) branches "propose" and "mailet-2.0". Nothing makes it into "mailet-2.0" unless


I would also like to completely reimplement and rethink JavaMail which is a bit of a performance bottleneck. I don't really have any intellectual attachment to it as a project. I'd rather like to collaborate if you guys have the same issues. Benchmarking of JavaMail has not been very promising and I find its API rather limiting and bizzarre. I suppose both projects (JBMS/JAMES) will have more problems with it as we get to IMAP.

Open to suggestions/clarifications.

Thanks,

-Andy


Serge Knystautas wrote:
On Fri, 29 Apr 2005 12:21:29 -0400
 [EMAIL PROTECTED] wrote:

Myself and the other committers are also discussing offline what might be necessary to collaborate on Mailets if there was interest from JAMES, but I don't really want to go down that path until after 1.0.


I would be interested, though we're also intending to change the Mailet API. It's very servlet API-esque, which is rather passe. Would be great to solidy the clamav and spamd mailets and continue to flesh that toolset. Anyway, I'm up for the collaboration.

--
Serge Knystautas
Lokitech

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



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



Reply via email to