incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: Why solve a problem that doesn't exist?
Date Thu, 07 Aug 2003 12:49:02 GMT

On Thursday, Aug 7, 2003, at 08:17 Europe/Rome, Jochen Wiedmann wrote:

> Quoting Greg Wilkins <gregw@mortbay.com>:
>
>> However, open process is at least as important as open software.
>
> Agreed. But the ASF has just given a bad example on this (IMO).

Allow me to point out that "incubation" is the process by which the ASF 
*screens* communities to understand if the have a future inside apache.

The incubator was created as a solution to the 'thermal death' of some 
projects, which were killed by the fact that the mail developers left.

Apache is about healthy communities, not about code.

The reason for this is that we believe that a healhty community will 
make good code, while the opposite is not necessarely true.

Why so? well, if a community is healhty and diverse, it is respectful 
and open to collaboration, alternatives, new ideas and new proposals 
and with no political bias. It is balanced and long-lasting.

On the other hand, good code without a healthy community around it, 
it's too tight to the current developers and their orientation. lack of 
diversity of affiliation, also, has strong impacts on the way the 
community might evolve because the company paying them could go 
bankrupt and leave the code dead.

> Following the discussions on Geronimo in the last days, my
> impression is that a lot of decisions (in particular architecture)
> has already made behind the scenes.

Wait, you are getting the wrong impression.

Apache has a rule for project incubation: while we focus on 
communities, you need something that works. We found out in the past 
that incubating a healthy community out of "ideas" is impossible: 
because it's much easier to talk than to write code, people talk a lot.

We call the initial codebase "the seed".

In this case, "the seed" is composed by code that was written by the 
people that donated the code to start this project.

It's the beginning of the design phase, not the end.

> I do not even know who took
> those decisions, or how they look like. I just read in some mails,
> that they are "soon to be published".

"the seed" is not yet on CVS because the ASF is screening all possible 
legal issues because we are *very* sensible to copyright issues: Apache 
does *NOT* have code on our CVS modules that the foundation does not 
own. It's a normal process during incubation to "legally validate" the 
seed, but in this case, additional care should be applied because of 
the legal sensitivity of this project.

> Not that *I* am the one who could influence that, but there have been
> some prominent names expressing interest in Geronimo on this list,
> who could.

When a project is being incubated, the group of apache members 
sheparding the incubation will look exactly at how the community 
handles disagreement and proposed changes.

After the seed enter the CVS, normal community open development will 
take place.

Let me repeat: a seed is the point to start, not the point to end.

There has been already on this list the question on whether or not to 
use Avalon for the containment. The choice of JMX vs. Avalon will be 
discussed several times. The development community will find consensus 
on how to move forward.

>> The high attrition rate of significant contributors to the JBoss 
>> project
>> over the years indicates that at least for some there is a problem, 
>> that
>> hopefully the open process of apache will address.
>
> That's definitely a point. On the other hand, I still have mixed 
> feelings.
> My impression is that the Apache side behaves very, well, formal.

Yes and no. What you percieve as "formal" is a collection of best 
practices on open source software development. "design patterns", if 
you want. They have been written by tens of different individuals after 
years of practice in creating open source communities, mostly out of 
jakarta.apache.org and xml.apache.org, I, myself, helped in the 
incubation of several communities that now count up to 20 active 
developers with a high degree of diversity.

Apache cares about the people more than they care about software.

And the people who wants to participate in this, they have to show they 
share the same view of software development.

> Right,
> there might be reasons for doing so, but the typical behaviour between
> various open source projects should be different, say friendly 
> competing.

This is exactly one of the parameters the shepards will use to judge 
the incubation evolution. If the community is not friendly, it will 
never become diverse, then it will never exit incubation.

You must understand that "incubation" is a process that have 
potentially two ends: the incubated community becomes a official apache 
project or it gets kicked out.

> While I reject words like "controlling every popular, open-source,
> significant project", I still would prefer a public statement like
> "we have attempted to do this and that, but that failed because ...".

I agree with you that it was not made clear that this project is *UNDER 
INCUBATION* inside Apache. We apache members know what this means, but 
several people outside don't.

I hope that my words clarify the position of the process.

> And, very important, followed by a "We are still interested in 
> discussions
> and open for exchange of ideas and possibly even sources, if license 
> and
> copyright allows." What good does it, to close the doors?

Believe me, every ASF member watching this process agrees 100% with 
you: if the development community behind Geronimo "closes doors", it 
will be a very bad sign and it will be picked up.

At the same time, it's *way too early* to jump to conclusions.

First, the seed didn't even enter CVS.

Second, the community didn't even start working or releasing stuff.

Third, there is a lot of FUD floating around and, IMO, all this will 
have to calm down before the community starts operating nicely and 
focusing on code, instead of fighting personal attacks and friction.

so, my point is: incubation is a 'trial period' that is going to end 
when the community becomes "healthy and diverse" enough for the ASF to 
be "confident" on the long-lasting behavior of it.

I hope this helped clarifying the ASF position on Geronimo.

If not, please don't hesitate to ask more information on this list: 
last thing we want is to give a wrong impression on what we are doing 
and why.

--
Stefano.


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


Mime
View raw message