Craig McClanahan wrote:
On 4/24/06, Frank W. Zammetti <[EMAIL PROTECTED]> wrote:
I look forward to feedback. Thanks for listening!
Without commenting on the merit of the proposal itself, or the reasoning
presented as its justification, it is important to note that we (the Struts
community) do not have free reign to do whatever we want in this regard.
Craig, while you aren't directly addressing me in this post, I take some
offense to the above statement. It is very insulting. You are insulting
people's intelligence.
First of all, this "royal we" of speaking in the name of the community
rubs me the wrong way. Also, it is quite obvious to me that, so-called
"Apache Way" or not, a community can add more committers if it wants to.
What discussions or proposals precede that is really pretty much beside
the point.
You collectively can certainly add Frank and other people who are
willing and able to do work as committers if you want to. Now, I'm
inferring that you don't want to do that. But then you should come up
with some other reason than this stuff about community and the "Apache Way".
As
part of Apache, the Struts community is accountable to the Apache Software
Foundation's Board of Directors, and one of the things that gets watched is
how well the various projects implement "the Apache Way". And one of the
key tenets of that "way" is how new committers get selected.
My casual observation suggests that the "way" in which new committers
get selected is entirely arbitrary.
In fact, it strikes me that what you are saying is 180ยบ from the truth.
If I want to know how somebody who has a ph.D. in a field has that
degree, I can go dig up the guy's doctoral dissertation and see how he
got the degree. If somebody is a member of the New York Bar, they must
have passed the New York Bar examination, and I could dig up examples of
the test that person would have had to pass. Guilds of medieval
craftsmen had a system whereby, to become a full member of the guild, a
"master" of the craft, you had to present some kind of major work. This
is where the term "masterpiece" comes from IIRC.
As regards the question of why one person is a "committer" and somebody
else is not, there is simply no equivalent way of pointing to why
somebody is a committer and another person is not.
Anyway, I think the text you wrote above is a further insult to people's
intelligence. I don't find what you're saying credible. I am quite sure
that the "Struts community" can add whoever they want as a committer and
the ASF Board of Directors is not going to say boo.
Are there any precedents you can point to?
Like, project FooBar added Joe Blow as a commmitter and the ASF Board of
Directors actually took issue with this?
Does this ever happen?
So I get the very very real sense here that you guys just make up this
stuff as you go along. Though I don't know which aspect is more
distturbing -- that you do this, or that other people just meekly go
along with it...
Regardless of the details, anything along the general lines of what you
describe would be quite different from the way Apache has grown from the
very beginning to what it is today -- and it would, therefore, be looked at
with a *lot* of skepticism by Apache as a whole.
Well, you seem to be implying that other ASF members who are not at all
involved in Struts are going to second-guess your decisions about who to
add as a committer and why you added the person(s) in question.
This just does not ring true to me.
Trying to argue that the
Apache Way is flawed, and needs to be fundamentally changed like this, does
not seem likely (to me) to get much agreement Apache wide, given the ASF's
history, and the success of many of its projects at building communities of
committers that buy in to that culture -- and, by the way, produce some
pretty popular software to boot.
Well, much (maybe most) of the software under the ASF brand was
developed originally outside of ASF. In the case of what is going on
with Struts right now, you are currently in the process of relabelling a
competing product, Webwork, as "Struts Action 2". Since the existing
Struts is going to be "Struts Action 1", there is a tacit acceptance of
the fact that the codebase developed within ASF, and presumably
according to the "Apache Way", was technically superseded by the
codebase developed outside of ASF.
It is quite clear that Struts 1.x development stagnated terribly and, to
keep saying that you can't do anything any differently from what was
done heretofore because it runs against this so-called "Apache Way" --
this strikes me as lunacy.
So what the heck is this "Apache way" thing? It is a culture; a way of
gathering a community that interoperates ...
In another part of this overall discussion, Niall Pemberton took me to
task (or tried to) because he characterized my attitude as wanting the
"demise of apache" and I did not deign to elaborate on that. (An example
presumably of me "walking away from discussions")
Well, one basic benefit of a hypothetical "demise of ASF" would be that
one would no longer have to hear any of this claptrap about the "Apache
Way". Frankly, that would be one major improvement in the situation. <sigh>
and it can be tough at times to
write down a descrption that makes sense.
Well, the other possibility is that, if you can't write a description of
the "Apache Way" that really makes sense, maybe it's because it doesn't
really make any sense...
Also, if it's left deliberately vague what the "Apache Way" is, kind of
mystical, you can set yourself up as the interpreter of what the "Apache
Way" is and this gives you leeway to stifle needed discussion by
jumping in and saying: "this runs counter to the Apache Way".
But [1] is a pretty good starting
point. I'd also recommend looking through the documentation for the
Incubator Project (where a specific exit criteria is that the folks
participating in the new project "get it" about the Apache way).
You know, recently, in private correspondence, somebody who is on the
PMC of a project that recently graduated from the "incubator" told me
that he didn't really understand what the "Apache Way" was.
The real fact of the matter is that projects in the incubator want to
get in on ASF because, like, it's a big ball of wax, you know. They know
that their work will get a lot more visibility by being on apache.org
than where it was previously. So, you know, you guys spout this stuff
about the "Apache Way" and they basically say: "Whatever you say, man..."
That's the way it really is...
It reminds me of the scene in the Shawshank Redemption where Red (The
Morgan Freeman character) is asked by the parole board: "Do you consider
yourself rehabilitated?"
"Oh, yessah, I am most definitely rehabilitated..."
Similarly:
"Do you consider that you now "get" the "Apache Way".
"Oh, yessah, we most definitely do..."
I know that in private, at least some of these guys will tell you that
they don't even understand WTF the "Apache Way is!!! And they've been
through this incubator and so on...
Frank
Craig
[1] http://www.apache.org/foundation/how-it-works.html
PS: If you like the code but don't like the community, one of the best
things about Apache is that the license gives you the right to fork and
build your own community, operated according to whatever rules please you.
So, even if a proposal like this is not accepted, you still have the
opportunity to go build "a better Struts" -- that doesn't have to be done
here.
Yeah, in other words: "Why don't you go fork off?"
Well, this begs the question really. First of all, it is not a special
characteristic of Apache projects that you can go fork them off
somewhere. That's true of open source projects generally.
The problem is that, once you're working on a non-canonical version of
the code, your work only gets a small fraction of the usage and
attention that it would if it was part of the canonical Struts project
on apache.org. You are basically in the position of being out in the
cold, shouting into the void.
I would bet that if you reviewed the list of current Struts committers,
you would find that there are people there who have contributed nothing
or hardly anything substantial in years. Why should they be committers
when somebody who wants to do something, like Frank, is out in the cold?
If that is what the Apache Way says, I think then, most pragmatic people
would say, to h*ll with the Apache Way.
OTOH, I don't think that either. I think the so-called "Apache Way" and
all the vacuous verbiage around it is basically a kind of MacGuffin you
guys trot out.
So stop it. Please answer the question without any nonsense about the
Apache Way:
What is the reason not to let Frank (or others who are able and willing
to do so) commit code?
Jonathan Revusky
--
lead developer, FreeMarker project, http://freemarker.org/
FreeMarker blog, http://freemarker.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
