Great, glad to hear it. Unfortunately, the public evidence (mailing
lists, wikis, etc) do not back up your statement. I understand that
people who are working with the ATF team find the people on the ATF
team to be friendly, technically competent, open to new ideas, and so
on - you are not the only person to reply about that. However, W, the
set of people working with the ATF team is much smaller than, U, the
set of users would would cooperate more with the project if it were
open and transparent (which is in turn smaller than, E, the larger
I have collaborated with the ATF team for several months now (to the
point that I am now part of it ), and I would qualify it as open and
transparent. This not the kind of "closed-shop" project you describe, as
I can witness.
This is all great stuff. How hard would it be to send an email to
atf-dev now and then saying "hey, I'm working with the Mozilla team on
issue X"? Then other people who are working on X or interested in X or
have experience with X could join in -- right now, because nobody knows
that the team is working on X (except you and a couple others), nobody
It is being adopted by folks at JBoss, MyEclipse, Aptana and others.
It made important contributions to getting a Mozilla browser widget
generally available in the core 3.3 platform.
It does some important collaboration with other projects such as Mozilla
XULrunner and JavaXPCOM.
It is a small team that delivers.
Actually, my comments are not about buzz (marketing), they are about
collaboration with the larger Eclipse community in development. There's
a certain minimum level of communication required to be an Eclipse
project. My critique is that ATF is not meeting that minimum. Clearly
there are projects that are at the top end of the communication scale
(Mylyn springs to mind) - I'm not asking for that level of
communication, but I am asking the project to meet a minimum level.
Some project have more resource to dedicate than others to
communication, and some projects are quieter or less outgoing, pretty
much the same way some people can be quieter and less outgoing. We are
probably not the best buzz generators are there for sure :-)
The statement was "normally we hold open meetings but we are not doing
so right now". That's not open.
Note that the statement you are referring to below is just about a
project lead stating that he is taking some vacations and trying to push
a release out of the door that very week. This is a small team, and
getting our fearless leader out is something serious.
Saying he is not available this week is both polite, open and
transparent in my book.
What is wrong with that?
A statement of "I'm going on vacation and will not attend this week's
open meeting" is a very different statement. Perhaps that is what he
meant to say, but it is not what he said.
Since the Board-driven rules for conformance labeling where in place by
mid March and your build was at the beginning of April, yes, I think
that the drop should be labeled incubation. Note that I'm not asking
for you to change the 1.0 file names. On the other hand, I'm willing to
listen to an argument about why I'm being unreasonable.
Would you want us to rename all the legacy drops file names?
It would not make sense to me. Please elaborate. Are Eclipse process
requirements retro active?
You are not required to use milestones. But it is a good idea as a way
to communicate with the rest of the Eclipse community. You are not
required to use six week milestones, but you are required to have some
process that make is clear and open.
There has been ongoing work to update the version numbering scheme.
I am not aware of requirements for milestone numbering beyond those for
I am aware of a guideline for incubation to use milestones, a common
practice to use milestones number in the released file names by several
projects, but not a requirement.
Do we should or do we must?
Great. How about saying some of that on your website? Or changing your
website to direct people to the wiki and then saying that on a wiki
page? It's easy to update a wiki page.
We are striving for the "just enough process" Eclipse is advocating for.
For now the project is going more towards frequent iterative drops, and
making sure those both conform to the requirements, and make sense from
a software engineering perspective.
My whole point is that the fact that we are having this conversation
about the ATF team's plans means that I can't find this information in
the public communications from the project. Not the website, not the
wiki, not the mailing list archives, not the newsgroup archives, ... I
shouldn't have to ask around (or, as some would say, insult the
developers) in order to find out the project plans and status.