---- Original Message -----
From: "Noel J. Bergman" <[EMAIL PROTECTED]>
>
> > Would you consider :
>> ....
>
> For what year?  ;-)  There has been a lot of work on James v2.1, and it
> should be put out as a release.  For a version "3", your comments make
sense
> to consider.

That is why I was suggesting 2 branches. One focussed on scalability(future)
and another on stability(now) and separate releases focussed on these.

> Will you be around to help?  :-)

Not sure, but I hope I can. :-(
There are some pieces I can contribute that have worked for me (I actually
use these)
- Scheduler
- Better connection management and SSL library.

I think James should architect for scale. Maybe start with bare bones
protocol handlers and add pieces only if they don't hamper scalability.
Redisgn or drop pieces/ideas that hamper scalability. I don't think this
would be an incremental effort from the current architecture. The upside is
James scaling to enterprise/ISP kind of load.

Even if this is not done, it would be nice to talk more about the
scalability limitations imposed by current architecture. For instance,
mailets/linear processors are powerful but what can be done to safely
distribute processing across machines/processes. Can some form of
messaging/workflow be used instead of linear processor ?

>
> As I understand the release plan proposal, the idea is to put out a stable
> release on the current code base, followed by preparing for potentially
> major revisions.

Sound like a good idea.

Harmeet


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

Reply via email to