Confessions

[Stefano Mazzocchi <st...@apache.org>]

It is with extreme sadness that I enter this mail list as an Apache
member. This is for a number of reasons:

 1) the ASF failed to provide assistance to this project
 2) the ASF failed to remove focus from "corporations" and move it over
to "individuals"   3) the ASF left you guys alone, without guidance and
without anybody to teach you the "Apache spirit of collaboration"...
suggesting rather than flaming, donating help rather than code,
indicating best strategies instead of having you to discover them the
hard way.

It is not a secret this project is the worst project hosted under Apache
in terms of community friendliness and direct individual collaboration.

Don't get me wrong, almost everyone is to blame, not only the ASF or the
XML PMC, but we thought that, as a successful software product, Xerces
would have had an easy life to develop a community around it.

It is clearly not the case.

I'm guilty as well because I already knew this, I already knew that good
code is used, while bad code is fixed.... and when fixing you learn...
when learning you enjoy it... when enjoy it you do it again.... and you
contribute your time to the community.

There are several signs that make the Xerces project a clear example of
what "not to do" when bootstrapping an open source project out of
external contributions:

 1) all the initial developers were paid by the same entity
 2) all the initial developers were managed by the same person
 3) the manager was also a project coordinator and PMC member
 4) the code was ready and more-or-less feature complete
 5) the code was indeed rock solid and great

even if it seems weird, all the above 5 points (which any person used to
closed development model would deeply appreciate) are responsible for
the failure (or, at least, slowness) of the evolution of the Xerces
community.

But the events are changing:

 1) diversity is building up (no single manager, no single check payer)
 2) xerces2 is a new effort and therefore not even close to be complete

In my mind, this clearly means that Xerces2 will be more successful than
Xerces1 in the long run and the community will be much stronger and will
reject corporate-based wars due to the fact that individuality is back
into the right central place.

So, no more "the IBM guys" or "my colleages at SUN" or "my team in
Cupertino" shit: we want people that "enjoy" working on this project and
if they can get payed to do it, well, what can I say, you're luckier
than me since I don't get a dime out of this (well, at least not
directly).

So, Xerces2 requires not only a new technical architecture (which I know
you guys are building and I think looks very nice) but an entirely new
spirit in the management of this community.

Since I've helped several projects to "happen" under the Apache
umbrella, I feel qualified to do the same here. Also, because I don't
have @ibm.com or @sun.com mail address (and won't in the future!), I can
tell anyone of you clearly what's wrong, why and what to do to fix it.

If you allow me, I'd like to volunteer to be your community advisor,
your main ASF link.

Another thing: the XML PMC failed to provide help to this project and
some of us ASF members already proposed to remove container-level PMCs
and promote all codebases to projects, much like the Apache Web Server
project has been doing.

This is to avoid blinding you since PMCs are almost useless: the real
thing happens here and foundation-wide discussions are dealt by the
members.

Did you ever thought about becoming an ASF member? wouldn't you feel
proud if that happened?

It's possible, believe me. I started with a small patch to JServ when I
was 22 and I became a member after two years: I was proposed for
membership because my contributions were considered important for the
foundation, important for the mission.

I'm proud to be a member, it's with honor that I serve the foundation
mission to provide a solid, open and free web infrastructure for
everybody, it's with pride that I receive comments from thirld world
countries where my software is helping people to establish their
business, earn some money and improve (even if in a small way, I know)
the economy of their countries.

There is a whole world out there... don't forget this.

You guys were left alone, we forgot to tell you how different open
development is from closed development, how different metrics are, how
strange things might look from your perspective.

It is not pointing fingers that we'll make this project successful..
it's not throwing flames around that we'll make you understand what you
did wrong.. it's not by forcing you to do something that you'll do a
good job on it.

I would like to share some of the Apache spirit with you guys, just like
wonderful people like Brian Behlendorf or Jon Stevens did with me when I
first subscribed to an apache mail list three years ago.

Apache is and will remain focused on its communities rather than on its
code. This is the main difference with other open source organizations.
We believe that an healthy community is *much* more important than a
great software. Why? simple: an healthy community will necessarely
create great software, while a great software will *NOT* create an
healthy community around it.

And how do you know a community is healthy? well, there are no
algorithmical metrics, but the emotional one is pretty simple: 

"a community is healthy when the original author/s leave/s the project
without killing it"

Now, please, try to answer this question about this community.

I honestly think that both Xerces1 and Crimson will die if the original
authors/maintainers walk away from them.

But I see a bright future this new effort, even if, to be entirely
honest, I think that a name change would make it much more "neutral" and
easier for everyone.

Let's clear the whiteboard: new name, new strategy, new architecture,
new community spirit based on individuals.

Both Crimson and Xerces have corporate strings attached... do we want to
keep them forever with us?

Corporations used us as a battlefield for their own political strategies
and we allowed them to play this game on us, hurting us badly in a
technology (XML parsing) which is not only vital, but it's *the*
foundation of all other apache XML projects.

This is why I propose a radical fork, some fresh air: think about it...
we have the chance to start again from scratch with a new design and new
architecture, wouldn't it be great to go back to feel part of it no
matter who pays your checks, no matter what political decisions are made
behind your back?

Apache has the power of doing so, we have the code, we have the license
to protect it, we have everything... we just have to use this power.

This is why I propose a fork: new module, new mail list, new project.
Fresh air and *the* best XML parser ever written by the best java/xml
programmers around, standard, open and backed up with a solid and
individually-oriented development community.

Trust me, it can be done.

-- 
Stefano Mazzocchi      One must still have chaos in oneself to be
                          able to give birth to a dancing star.
<st...@apache.org>                             Friedrich Nietzsche
--------------------------------------------------------------------
 Missed us in Orlando? Make it up with ApacheCON Europe in London!
------------------------- http://ApacheCon.Com ---------------------

Re: Confessions

[Brett McLaughlin <br...@lutris.com>]

Stefano Mazzocchi wrote:
> 
> It is with extreme sadness that I enter this mail list as an Apache
> member. This is for a number of reasons:
> 
>  1) the ASF failed to provide assistance to this project
>  2) the ASF failed to remove focus from "corporations" and move it over
> to "individuals"   3) the ASF left you guys alone, without guidance and
> without anybody to teach you the "Apache spirit of collaboration"...
> suggesting rather than flaming, donating help rather than code,
> indicating best strategies instead of having you to discover them the
> hard way.
> 
> It is not a secret this project is the worst project hosted under Apache
> in terms of community friendliness and direct individual collaboration.

Agreed. As an outsider looking in, this is very apparant. And what is
more telling is the tone I hear people talk about projects like Cocoon
and Tomcat talked about as opposed to the tones that Xerces is spoken of
in. It's just not the same atmosphere, in any sense. 

<snip />
> 
> But the events are changing:
> 
>  1) diversity is building up (no single manager, no single check payer)
>  2) xerces2 is a new effort and therefore not even close to be complete
> 
> In my mind, this clearly means that Xerces2 will be more successful than
> Xerces1 in the long run and the community will be much stronger and will
> reject corporate-based wars due to the fact that individuality is back
> into the right central place.
> 
> So, no more "the IBM guys" or "my colleages at SUN" or "my team in
> Cupertino" shit: we want people that "enjoy" working on this project and
> if they can get payed to do it, well, what can I say, you're luckier
> than me since I don't get a dime out of this (well, at least not
> directly).

I hope this is the case - honestly, I don't see that. I still see a lot
of things happening behind the scenes, and a lot of things coming about
that are never discussed. I see a closed software methodology, with
dates being set for release before anyone talks about it. Honestly, it's
the first Apache project I've ever seen dates attached to

1) before requirements are done
2) as opposed to focusing on functionality

> 
> So, Xerces2 requires not only a new technical architecture (which I know
> you guys are building and I think looks very nice) but an entirely new
> spirit in the management of this community.

Again, I hope so. But I think that the way it looks so far is that
Xerces 2 is not as much innovation as it appears to be recreation. I
favor a clean slate; while the requirements being gathered are a good
start, it's too early to be coding in things, which is exactly what is
being done. I'm just not clear that, as you point out, people really
know how open source projects operate. I don't claim to be an authority;
but I've run, with Jason Hunter, the JDOM project, I've co-run with Jon
Stevens Turbine, and things are drastically different in those areas.

> 
> Since I've helped several projects to "happen" under the Apache
> umbrella, I feel qualified to do the same here. Also, because I don't
> have @ibm.com or @sun.com mail address (and won't in the future!), I can
> tell anyone of you clearly what's wrong, why and what to do to fix it.
> 
> If you allow me, I'd like to volunteer to be your community advisor,
> your main ASF link.

I'm for that, in the sense that your voice is more respected and adhered
to often than others (mine included). I'm more than happy to see
eveyone, including myself, get corrected when we need to be, and to try
and agree that you know more about the Apache vision that most at IBM,
Sun, or even Lutris ;-) [my company, if you missed the joke).

> 
> Did you ever thought about becoming an ASF member? wouldn't you feel
> proud if that happened?

Good point - I see lots of @apache.org addresses at Xerces; while I
assume that all are warranted, it was not clear how that happened, or if
they were doled out to those who started the project, as if they were
"grandfathered" in. Honestly, I'm OK if they were - IBM and co. did a
huge amount of work on Xerces when it was Ibm XML4J; but nobody knew
what happened, and I think it inadvertantly cheapened the @apache.org
address.

> 
> Apache is and will remain focused on its communities rather than on its
> code. This is the main difference with other open source organizations.
> We believe that an healthy community is *much* more important than a
> great software. Why? simple: an healthy community will necessarely
> create great software, while a great software will *NOT* create an
> healthy community around it.

Again, I think you point out something important. I know that Apache's
beginnings were, honestly, paltry compared to NCSA and Mozilla and
Netscape; but the community was good. I'd settle for the long-run
advantage that the community brings, aka Apache 1 and now 2, rather than
a short term feeling that Xerces 2 is immediately perfect. Honestly, we
have enough of a product in Xerces 1 to sustain us while we work out any
problems; if that means, for a time, sacraficing code to instill the
right attitudes, I'm +1, for what it's worth.

> 
> And how do you know a community is healthy? well, there are no
> algorithmical metrics, but the emotional one is pretty simple:
> 
> "a community is healthy when the original author/s leave/s the project
> without killing it"
> 
> Now, please, try to answer this question about this community.
> 
> I honestly think that both Xerces1 and Crimson will die if the original
> authors/maintainers walk away from them.

Agreed. And Xerces 2 is already facing a dichotomy. There is a group
working on the process, on requirements, on voting, and that's great.
But there is this effort in parallel to move things along, and it feels
(from experience in the commercial world) to productize it; to get
something to market fast. IMO, this isn't just IBM's fault, as the
community isn't forcibly halting that effort and reining it in. We, the
people (to coin a phrase), need to insist that things be done so that
all who want to can understand. No response is actually acceptance in
open source, and I think you point out, Stefano, that no response is
what we've been doing for too long.

> Let's clear the whiteboard: new name, new strategy, new architecture,
> new community spirit based on individuals.

Very much +1. I'd like to see everything but the requirements, which
were arrived at in very clear, voted fashions, to be wiped out.

> 
> Both Crimson and Xerces have corporate strings attached... do we want to
> keep them forever with us?
> 
> Corporations used us as a battlefield for their own political strategies
> and we allowed them to play this game on us, hurting us badly in a
> technology (XML parsing) which is not only vital, but it's *the*
> foundation of all other apache XML projects.
> 
> This is why I propose a radical fork, some fresh air: think about it...
> we have the chance to start again from scratch with a new design and new
> architecture, wouldn't it be great to go back to feel part of it no
> matter who pays your checks, no matter what political decisions are made
> behind your back?
> 
> Apache has the power of doing so, we have the code, we have the license
> to protect it, we have everything... we just have to use this power.
> 
> This is why I propose a fork: new module, new mail list, new project.
> Fresh air and *the* best XML parser ever written by the best java/xml
> programmers around, standard, open and backed up with a solid and
> individually-oriented development community.

Again, for what it's worth, +1. I'm not a Xerces committer; but I'm a
Cocoon committer, an ECS committer, a Turbine committer, a jBoss
committer, an Enhydra committer, a JDOM creator... for what it's worth,
I think Stefano is right on. And from the point of view of a designer of
one of the XML APIs for Java, JDOM (which is not as big as SAX or DOM,
but is soon to be standardized in a formal way), I am much more
interested in seeing a clear, untethered parser to build JDOM support
for and to include with JDOM than I am in maintaining the status quo.
And it does mean more work on my part, and on JDOM's part, and I think
that's a small price to pay for what we would gain.

My .02c

-Brett

-- 
Brett McLaughlin, Enhydra Strategist
Lutris Technologies, Inc. 
1200 Pacific Avenue, Suite 300 
Santa Cruz, CA 95060 USA 
http://www.lutris.com
http://www.enhydra.org

Re: Confessions

[Dirk-Willem van Gulik <di...@covalent.net>]

On Thu, 5 Oct 2000, Stefano Mazzocchi wrote:

> It is not a secret this project is the worst project hosted under Apache
> in terms of community friendliness and direct individual collaboration.

Looking at the list - I see coding happening  - I see technical opinion
differences - and, certainly, there is a certain level of a second
guessing games as to opinions and 'ulterior' motives. Which makes it
perhaps a bit less of a pleasant environment.

But all in all - code is being written - about 260 commits or so in the
last month by at least 8 differnt coders from at least 4 or 5 companies.

That is more than we are currently doing against apache-two-oh!

> Don't get me wrong, almost everyone is to blame, not only the ASF or the
> XML PMC, but we thought that, as a successful software product, Xerces
> would have had an easy life to develop a community around it.
...
> Apache is and will remain focused on its communities rather than on its
> code. This is the main difference with other open source organizations.

Hmm.. Code is what matters - and if code is bad - or people stop coding -
then a project might lose its appeal and its developers. If there is a
sufficient mass to do something new - let that mass go where it wants :-)

> I honestly think that both Xerces1 and Crimson will die if the original
> authors/maintainers walk away from them.

Yes - and sometimes this is the right thing. 

When Crimson was started it was discussed as to wether to allow such
competition to exist - and worries where vocalized as to wether such
competition would distract and dilute the developer pool. Or would lead to
confusion. 

At that time it was said that this was unlikely - that things would
ultimately merge or quietly supersede each other. Some of us, mostly those
who had witnessed the epic forks of the past from close up, had great
doubts - but the only fair way - fair to those with new ideas - and fair
to those who depend and want to work on 'live' operational code - to find
this out is to do, and to see what happens.

And in a way - by skipping to another track and kicking of xerces/II - we
now let history repeat itself - Cocoon was half abandoned and lead to
Cocoon2, Jserv lives in parallel with Tomcat.

Though one might argue that Apache-2.0 is being cooked up while we work on
a 1.3.13 release - the latter is suitable for operation.

And I would really think twice before I would abandon Xerces - which
really is damn good - and ready for ops !

Dw