james-server-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Burrell Donkin <robertburrelldon...@gmail.com>
Subject Re: Project status and perception
Date Sun, 10 May 2009 16:53:10 GMT
On Sun, May 10, 2009 at 1:58 PM, A. Rothman <amichai2@amichais.net> wrote:
> Hi all,
>
> Here's a thought/feeling I've been having for a while regarding JAMES, as a
> user of many years (both at work and at home). It's not at all meant as a
> rant, but as the personal experience of a long time user who really wants to
> see this project thrive - this is how it looks from the 'outside'. [reposted
> from elsewhere - oops :-) ]

it's good to talk :-)

> Speaking for myself, I can't exactly say that I'm interested in more
> releases and features from a 2.x codebase with no active developers,
> targeted at Java 1.4 (won't even compile on 6), with chunks of the project's
> wiki (even task-related pages) last updated 4 years ago, based on some
> framework projects that have been discontinued years ago... at least that's
> the impression the project is making in the past couple of years - that it
> is a dying one.

(here's the way i see things - others may disagree)

the 2.3.x code base is mature. it's production tested and secure. a
point release is overdue since there are a few bugs fixed since the
last one.

most active developers now use 3.0 code. so, we need more active users
of 2.3.x to contribute by testing backports and improving the
documentation if they want to see further releases.

it's not really possible to develop this code base any further.
problems with avalon, backwards compatibility, revised design
requirements and JDK changes meant that active, evolutionary
development stopped several years ago on 2.x.

sometimes (when this happens) a way forward can be agreed. at other
times, the existing developers fall out and the project explodes. i
started developing james 3.x when the developers were exploding. the
fall out from those disagreements meant that most of the server
developers are now much less active. so, the server's short of
developers (but the libraries are still reasonably active).

sometimes, time is the only healer. i think enough time has passed and
that progress would now be possible if new contributors want to get
involved.

> But, I wouldn't want to abandon it either. As a user, the alternatives are
> either the above, or a trunk 3.0 version whose status, progress and future
> is not quite clear either - there's a v3 wiki page that was last updated at
> the beginning of 2005! Perhaps it's time for a proper alternative? Perhaps
> an alpha/beta of 3, with a fresh feeling of liveliness? Something to show a
> clear sign of activity and progress? cleaning up the wiki from ancient
> roadmaps and plans which have been abandoned long ago? I think that feeling
> of freshness would make it much easier to contribute than the current
> uninviting feeling of gloom... I contribute to several projects in what
> spare time I've got, and this one always gets pushed to the bottom of the
> list because of that feeling of 'who knows if it's not already dead? Are
> there more than a dozen users still using it? haven't seen any movement
> there in months, years... why not work on that other well-supported project
> instead?' Or to paraphrase the well known quote - "contributors contribute
> to those projects who help themselves"...

i've done the whole popular project thing before and i decided on
james that i'm focussing just on the code. i'm willing to help anyone
interested in documentation but it's not something i'm willing to take
on.

but maybe james would come higher on your personal development agenda
if you saw it as a far of doing cool stuff rather than a chore

i'm interested in advanced, next generation mail architectures. in
particular, in pulling in multiple information sources (news, email,
atom, jabber, etc), to process them (especially using machine learning
for intelligent filtering), persist in document-centric storage then
pushing them out through multiple protocols (IMAP, POP3, twitter).
there's a lot still to do, and if anyone's interested in this stuff
it'd be great to work together,

this agenda is controversial and i don't think we'd be able to agree a
road map along those lines. but a lot of the work needed is well
aligned with other visions for 3.0. so, i use james 3.0 which
motivates me to improve and fix it.

IMO IMAP's a big enabler. creating a reasonable IMAP implementation
has been a PITA but i think it'll prove worthwhile. if people are
interested in exchange alternatives, it would be possible to push on
from this base. approaching new protocols as independent, pluggable
libraries is my preferred approach so there's less to learn. if
anyone's interested in improving exchange compatibility then i'd be
willing to help get you started.

the elephant in the room is avalon. the death of avalon has really
wounded james. avalon it a PITA and makes the code difficult to work
with. it need to be replaced but the topic of replacement has proved
controversial in the past. i suspect that this isn't so heated these
days.

the good news is that - by chance - this wait has had some advantages.
in the meantime, OSGi has emerged and matured. it also plays much
better with spring these days. the karaf microkernel which was
recently donated to felix looks really cool.

IMHO moving to karaf would be really great for james. the server could
be split into a number of loosely coupled libraries which could be
dynamically loaded and re-configured. if advanced OSGi sounds
interesting to anyone then jump in.

another personal interest of mine is improving JPA and JCR support. if
this sounds interesting to anyone, then please jump in.

- robert

---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org


Mime
View raw message