incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Geir Magnusson Jr." <g...@pobox.com>
Subject Re: Checkpoint on Harmony (Re: [discussion] Harmony podling to ask for vote for graduation)
Date Fri, 20 Oct 2006 03:02:04 GMT

Greg Stein wrote:
> On 10/19/06, Roy T. Fielding <fielding@gbiv.com> wrote:
>> On Oct 19, 2006, at 3:32 AM, Geir Magnusson Jr. wrote:
>> ...
>> > I'd like to ask that those who have asked for a release to assuage
>> > concerns about community health and capability to please read those
>> > 3 testaments from the mentors (ok, in Leo's case, 71 or so...) and
>> > please consider withdrawing your request for a release.
>>
>> That's silly -- why would people who think a release (or at least the
>> release process) is a useful learning process drop such a *request*?
>> I could understand your concern if it was expressed as a requirement,
>> but it most certainly wasn't.
> 
> Right. I said it would be useful to see if the community can make it
> happen. I know that *some individuals* can, but that is different. I
> didn't vote, I didn't say it was a requirement, just asked: why can't
> you pull together a *developer* release. Not TCK'd, not for Joe User,
> but something more formal than "here is a snap of Subversion".
> Something that developers in and around the Incubator can try. Or
> simply for (Incubator) people to observe how you plan to organize the
> community to get a release out the door.

People can already use the snapshots. We don't really just do a snap - 
there's discussion about if we believe that the parts are at least 
stable enough to do it.  We break things - what we're doing is pretty 
hard - so we don't do snaps when we know it's broken.  I'm not trying to 
toot our horn here, but this is pretty hard stuff, and IMO we're just 
not ready to be putting something out there with even an implied 
statement of stability or usefulness.  This isn't about TCK - we'll 
start that ASAP, and probably complete it sometime mid next year if all 
goes well.

Look, if incubator people really feel the need to observe that process, 
that's fine.  It's clear the delivered artifact is not what matters, but 
how the community does it. I can think of a few things that we could do 
  as a single-jar release that would require work to setup, and we can 
go through the release process to produce.

> 
> Without seeing that, you could graduate to a TLP that may be fully
> incapable of producing a release. Ever. I have zero indication that
> the Harmony community can actually produce releases. Code, yes.
> Releases, no.

I think it unlikely that it couldn't given the amount of decision making 
the project has already gone through (which is the key thing, IMO), but 
it's really hard to disprove your assertion - I take you at your word 
that you have zero indication.

What is clear to me is that anything we do now will have little 
resemblance to what we do to release 1.0, as there are tons of issues 
that remain to be sorted, such as which platforms to support, how the 
TCK integrates into our test framework and how we use it, etc.  The only 
thing in common would be how the community works together, votes 
together and makes decisions together, and as you know, I feel that 
there is ample evidence to support that those skills have been demonstrated.

> 
> Don't you think it would be a useful experiment for the Harmony
> community? 

While any activity like this with a young community is a useful 
experiment, I'll point out that we're now literally experimenting on our 
podlings. :)

I personally don't think that it would add any more information than 
what has been summarized by the mentors and is in 18 months of mail 
archives.  However, that's my personal view, and I've had close contact 
with the community.  It's clear that you haven't come to the same 
conclusion yet.  Roy suggests that unanimity of opinion is just an 
ideal, but it's one I think it's worth working towards.

>Or are you viewing it solely as make-work? 

To be frank, yes, I think it is make-work as it isn't in our plans, 
although I'm sure that if this is required, we'll be happy to do it.  I 
was hoping that enough info would be provided by mentors to satisfy 
people's need for information (as it did Justin).  Hey, I had to give it 
a shot, right? :)


> That you're
> confident that the Harmony community will not suffer problems that
> we've seen in other communities around their release processes?  The
> decision is certainly the community's (actually... where are they? why
> is Geir the only active Harmony person in this conversation?).    I'd
> like to hear the community say, "nah. we think that would be busy work
> with no utility." Fair answer, but I will note that nobody has made a
> clear statement like that. Talked around the issue, but never directly
> answered.

This is an interesting notion, as I think that this "measurement" is 
clearly going to affect the system being measured.

If you read the mail lists, the "steady state" of the project is that 
we're heads-down on code, working together, happy with snapshots, and 
working hard to get things integrated and completed.  From afar, it's 
reasonable to assume that we aren't that interested in a release right 
now because it's never been mentioned.

Now, the Giant IPMC God is going to ask the project if it's interested 
in doing something that it so far has shown no interest in, yet it's 
clear that Significant Members wish to see it happen in order to give 
the project something that it *does* have a demonstrated interest in, 
namely graduation from the Incubator.  What do you think the answer is 
going to be?

Removing my mentor hat for a moment, and putting my community 
member/committer hat on, I'm going to say "Sure! As you wish!", do it, 
and then go back to helping get the JVM stabilized (so we can do that 
'developer release' sometime in the near future....)

:)

geir



---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org


Mime
View raw message