Robert Burrell Donkin ha scritto:
On Mon, Apr 28, 2008 at 7:21 PM, Stefano Bagnara <[EMAIL PROTECTED]> wrote:
Robert Burrell Donkin ha scritto:
who's interested in releasing IMAP? not me, for one. i can live very
happily enough without any releases for the forseeable future.
releases attract users, not developers.
True. Just take into consideration that many JAMES developers knew JAMES as
users and then decided to start hacking the code.
1) Releases attract users that could be future developers
2) Releases give some developers motivations/satisfaction to keep working.
i don't have the energy to cope with users. even developers are
difficult since the codebase is in flux...
Even if I lost my motivation in coping with the PMC I think I'm still
doing user support as much as possible. Yes, I reduced my activity there
too, but I try to keep running with users as I think they are the only
key for JAMES issues: we need new users, new developers, a new community
driven by new smart people. Since I joined james (2005) I replied almost
an average of 1 message per day on james-user. This does not catch up
with user questions, but this is A LOT. Norman also helped in the last
years. Danny, Serge and expecially Noel worked very hard on the
james-user list in the previous years. Unfortunately I guess both the
users community and the developers community has been reduced since
2002-2003 by a 5 factor.
I, for one, started as an user of the 2.1 and then 2.2 and then I've had to
fix/change a lot of code and decided to try to be a developer, I'm very
happy for 2.3.0, and then I stopped because I had no enough
energy/motivation to make another release like that one and I see no way to
release again. But this is me. And this is past.
i don't have the energy to do everything. i can see a route towards
new JAMES releases involving considerable code from trunk if not trunk
itself. i just can't see any chance of releasing trunk at all. but if
no one else who wants releases has energy to contribute then i'm not
going waste my time.
I would +1 any attempt to branch or release milestones from trunk even
tomorrow. I would test the resulting milestone, I would report bugs, I
would *probably* report patches, too. But I won't push this, as I've
been accused to push things and I'll wait for others to push now :-)
I hope you get me as an "opinionist" and not as a contender: I really
admire your work and your motivation (I've been there) and I hope you
understand that in practice I will rarely use my -1, but I now like to
scream as loud as I can because most of our PMC tends to sleep, and I
want them to hear and take reasoned decisions ;-)
IMAP is orthogonal to the community issues surrounding trunk. IMAP is
not tightly coupled to the rest of JAMES. if the community issues
remain unresolved when IMAP is ready for release, the easiest approach
would be to backport to a uncontroversial version.
the only way that trunk is going to get released is if someone steps
forward
I agree. What I didn't agree is that removing IMAP is a step forward and
that this is a community opinion.
i never said removing IMAP is the majority opinion: it's bound to be unpopular
You are right, I made a conclusion based on a couple of sentences and
maybe I was wrong.
I told:
"IMHO there is no shame in releasing code marked as unstable together
with stable code. We introduced modules also for this, didn't we?" and
you replied:
"yes but IMHO that's not an opinion shared by the majority of those
with binding VOTEs."
If I understood this was what made you decide to start removing IMAP,
but maybe I misunderstood.
I just want to tell you "please check that someone else in the PMC, in
addition to Noel, thinks that releasing unstable modules from trunk is a
shame", because I only count him from my mailing list archive lookups,
and I really hope we don't still count Noel as the majority ;-)
I could accept to give *yours* idea (as active developers against
speak-only people, including me) the power of majority, but not Noel's ;-)
But I already said that I trust you on
this. I will not -1 this.
If this is needed then let's make another "top level" project,
svn/james/imapserver (with no weird trees, please ;-) ), or let's push back
to sandbox (I would hate this, but if THE COMMUNITY think it is better, I
will not vote down this too).
JAMES suffers from everyone doing anything interesting in the sandbox.
this is bad. the reasoning behind the modular build is that anyone
should be able to try new stuff without having to fork JAMES. that's
now easy in trunk.
ATM JAMES suffers from everyone doing nothing :-(
The sandbox ERA is already past.. it lasted few months...
As I said I would hate moving IMAP to sandbox. It was simply a statement
to tell you what I would "accept" without a -1. If I understood it
correctly you was fine with keeping IMAP in trunk or moving it to
svn/james/imapserver, am I wrong? (I just didn't like the
svn/james/server/imapserver solution, but I hope this is a minor
technical detail)
i think that the lack of releases is unhealthy for the community. i
also understand that many developers feel frustrated. i see no reasons
why JAMES couldn't release a couple of components a month for the next
year or so provided that the people who want more releases step
forward.
I think you understand that when this community try for real and concretely
to release something I always try to help someway (site stuff, release
tests, maven issues). Just don't ask me to push things. I gave up with
pushing :-) . I will join when I see something I consider concrete and
realistic.
if no one's willing to push forward releases then they aren't going to
happen. previously, you were pushing against too many in the community
to have a realistic chance of success. the art is to approach from a
different perspective.
Have you understood WHO was against and WHY? I'm still waiting to
understand why and to get counterproposal or to get some "sorry, I was
wrong, please try again now". Sometimes it is good to read again that
old messages. Read the motivations.. people was scared because they need
some more months to complete their work... I guess they had their months
now, maybe now everything is ok to release ;-)
I still hope you will try to push, because if you don't do this, no one
else will do now. I don't expect anymore the community that complained
me and Norman at the end of 2006 to get a better solution to our
proposal (I checked my spam folder and is not there ;-) ).
i think releasing mailet-api, mailet-base, std-mailet and crypto-mail
in the next month is not unrealistic if someone else were to step
forward to act as release manager so we can spread the load. we will
then have started to release the 3.0 code base in a way that allow
it's encorporation within the 2.x codebase.
I will not do the release manager, but I can try helping with mailets
and coping with dependencies and code analysis as I did in the original
message identifying what mailets/matchers can be moved or not and why.
there is a lot of interest in lightweight, embeddable protocol
handling libraries. one way to reassemble JAMES would be to extract
our popular protocol into separate products. we then build the
headline JAMES server from loosely coupled components with separate
versioning. the more i look at the codebase, the more i think that if
only avalon were not so intrusive this would be a realistic
possibility. this would allow SMTP to be released when it were ready
and arguments about that protocol to be restricted just to that
codebase. it would also allow JAMES components to be easily reused in
other projects (there are several who would be interested if only
JAMES were not so monolithic).
- robert
I don't consider JAMES a so big monolithic application. IMHO JAMES has a
small-medium codebase. I'm used to work on much larger source trees and
I don't feel this need to break things apart (but I guess you already
got this ;-) ).
That said I'm working on avalon free, seda based, protocols libraries
outside ASF/JAMES so I really agree on the rest of your sentence.
The more we talk, the more I think you're simply late (wrt JAMES
involvement) and that if you was here 2 years ago we would have james
*4.0* out now!
Stefano
PS: OK, I realize I should not consume so much of your time. My "bar
discussions" are not concrete steps to release.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
