You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@incubator.apache.org by Stefan Hepper <st...@hursley.ibm.com> on 2008/02/01 12:11:40 UTC
Re: [Proposal] NoNameYet - Pluto
here my response to Endre's mail
(http://mail-archives.apache.org/mod_mbox/incubator-general/200801.mbox/%3c47A1E1D5.3050404@Stolsvik.com%3e):
about Pluto V 1.x:
Due to the JCP process guidelines at that time you could not have early
public drafts and thus you are correct that the RI got to Apache very
late. However, we discussed this with people from Apache and we also
went through the incubation process. It would have always been an option
to not accept the project.
It is not true that after the JSR was final everything stopped. In fact
once we had finished 1.0 there was still work done to get to a more
stable 1.0.1 release. After that the pluto community re-structed the
code which led to the pluto 1.1 stream, so you can see that it was an
active community not only some code drop.
Also in my view pluto was a success, take a look at all the projects
using the pluto container: http://portals.apache.org/pluto/powered.html.
about JSR 286/Pluto 2.0:
Besides myself there are 4 people from Apache projects sitting in the
JSR 286 EG and now that the JCP supports early public drafts we
published early drafts more than a year ago and started the development
of 2.0 directly at Apache. It is also based on the new 1.1 codestream
that the Apache community created.
I think you should also consider the view points of companies donating
work and code to Apache. In my opinion it would have been much cheaper
for IBM to just put the RI out there on some server instead of going to
Apache. We went to Apache because we wanted the portlet standard to be
widespread and easy to adopt by everyone.
Regards,
Stefan
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Feb 2, 2008 10:24 PM, Bertrand Delacretaz <bd...@apache.org> wrote:
> On Feb 1, 2008 12:59 PM, Paul Fremantle <pz...@gmail.com> wrote:
> > ...Maybe the interesting question is whether you think - based on your
> > experience - if is is really appropriate to try to create a JCP RI as
> > an Apache incubator project?...
>
> IMHO, Jackrabbit is a successful example of such an RI at Apache.
A key point here is that Jackrabbit was never meant to be just the
reference implementation of JSR 170. A pre-1.0 snapshot of Apache
Jackrabbit was used as the JCR RI, but that was never the ultimate
goal of the project.
Basically, a project needs to be used, not just developed. We
typically focus a lot on the diversity and health of the development
community, but perhaps we should also consider the user community.
With Pluto it sounds like (I'm just reading the thread here) IBM's
goal was to develop the product but not really to use it.
BR,
Jukka Zitting
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Bertrand Delacretaz <bd...@apache.org>.
On Feb 1, 2008 12:59 PM, Paul Fremantle <pz...@gmail.com> wrote:
> ...Maybe the interesting question is whether you think - based on your
> experience - if is is really appropriate to try to create a JCP RI as
> an Apache incubator project?...
IMHO, Jackrabbit is a successful example of such an RI at Apache.
-Bertrand
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Paul Fremantle <pz...@gmail.com>.
Stefan
Thank you for you insights and response.
Maybe the interesting question is whether you think - based on your
experience - if is is really appropriate to try to create a JCP RI as
an Apache incubator project?
Paul
On Feb 1, 2008 11:11 AM, Stefan Hepper <st...@hursley.ibm.com> wrote:
> here my response to Endre's mail
> (http://mail-archives.apache.org/mod_mbox/incubator-general/200801.mbox/%3c47A1E1D5.3050404@Stolsvik.com%3e):
>
> about Pluto V 1.x:
> Due to the JCP process guidelines at that time you could not have early
> public drafts and thus you are correct that the RI got to Apache very
> late. However, we discussed this with people from Apache and we also
> went through the incubation process. It would have always been an option
> to not accept the project.
> It is not true that after the JSR was final everything stopped. In fact
> once we had finished 1.0 there was still work done to get to a more
> stable 1.0.1 release. After that the pluto community re-structed the
> code which led to the pluto 1.1 stream, so you can see that it was an
> active community not only some code drop.
> Also in my view pluto was a success, take a look at all the projects
> using the pluto container: http://portals.apache.org/pluto/powered.html.
>
> about JSR 286/Pluto 2.0:
> Besides myself there are 4 people from Apache projects sitting in the
> JSR 286 EG and now that the JCP supports early public drafts we
> published early drafts more than a year ago and started the development
> of 2.0 directly at Apache. It is also based on the new 1.1 codestream
> that the Apache community created.
>
> I think you should also consider the view points of companies donating
> work and code to Apache. In my opinion it would have been much cheaper
> for IBM to just put the RI out there on some server instead of going to
> Apache. We went to Apache because we wanted the portlet standard to be
> widespread and easy to adopt by everyone.
>
>
> Regards,
> Stefan
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
>
--
Paul Fremantle
Co-Founder and VP of Technical Sales, WSO2
OASIS WS-RX TC Co-chair
blog: http://pzf.fremantle.org
paul@wso2.com
"Oxygenating the Web Service Platform", www.wso2.com
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Luciano Resende <lu...@gmail.com>.
On Feb 2, 2008 6:08 AM, Leo Simons <ma...@leosimons.com> wrote:
> However, much more importantly, proposals should be evaluated on
> their own merits, not based on what happened to some other unrelated
> project 4 years ago.
>
+1
--
Luciano Resende
Apache Tuscany Committer
http://people.apache.org/~lresende
http://lresende.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
RE: [Proposal] NoNameYet - Pluto
Posted by "Noel J. Bergman" <no...@devtech.com>.
> But it is not individuals that propose this particular project, as I
> understand it: it is IBM and BEA. And it was IBM that, in my view,
> dumped the JSR 168 RI and then fled - not any individuals as such.
And IBM is also a significant force behind Tuscany, and have definitely not fled. They are also involved in Derby and Geronimo.
> I don't think Apache necessarily is the right place to dump a JSR RI and
> TCK implementation
I agree. The ASF is not the right place to dump things. But it is the right place to build a community.
--- Noel
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Endre Stølsvik <En...@stolsvik.com>.
Bill Stoddard wrote:
> Endre Stølsvik wrote:
>> Leo Simons wrote:
>>
>>> Sure, activity is not that high, and there's not a *huge* developer
>>> community, but there does not really seem to be any problem, either.
>>> Apache doesn't require projects to be huge successes (by whatever
>>> metric) as long as they're healthy and self-sustaining.
>>
>> This was not healthy up until quite recently - I'd even question
>> whether the code and the community is _healthy_ now. And the main
>> point was that IBM, the dumpsters, fled the plot pretty fast after the
>> import was done, not bothering about creating The Healthy Community
>> around the code they off-loaded before leaving.
>>
>> This should raise at least some slight concerns in Apache, IMO.
>
> Endre,
> Disclosure... I work for IBM.
>
> I agree with your concern about code dumping and I completely respect
> your comments. However, I object to the logic you are applying here.
> IBM'ers participate on projects as individuals and it's the actions of
> individuals that should be judged. To tar all IBM'ers because of bad
> behavior (perceived or real) of a few is just wrong.
I don't group all IBM'ers, really - I actually believe IBM'ers do tons
of good in many open source projects. Also I think IBM itself is a
somewhat good open source citizen in several regards.
But it is not individuals that propose this particular project, as I
understand it: it is IBM and BEA. And it was IBM that, in my view,
dumped the JSR 168 RI and then fled - not any individuals as such.
I don't think Apache necessarily is the right place to dump a JSR RI and
TCK implementation (because, lets face that, it isn't *developed* here)
- it goes against the entire grain of Apache, AFAIU.
At least, if it is put here, then just don't pretend that it more than
that either: It's just the RI and TCK implementations, staying at Apache
as Apache are good guardians of code on a general basis.
Thus, if it happens, this particular project's name shouldn't be
anything fancy, probably not include the name "Apache", maybe just be
JSR-235 with two subfolders: RI and TCK.
On the other side, some server implementing JSR-235 could be called
Apache What-Ever, would run its own incubation, have its own
infrastructure, possibly use the RI as a code starting point - but
nothing more. This would keep the distinction very clear. The goals
seems too different to mix.
Endre.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Bill Stoddard <bi...@wstoddard.com>.
William A. Rowe, Jr. wrote:
> Roland Weber wrote:
>> I think that is a bit oversimplified. IBM has strict rules about
>> open source participation. It is either "on private time", such
>> as my involvement at Apache. Then the person is acting as an
>> individual. Or it is "on company time". Then the person is doing
>> what he or she is paid for. And if IBM is changing it's priorities,
>> or the line item that required OSS participation is closed, plenty
>> of other work will be dumped on that person, simply leaving no
>> (work) time for OSS participation. Yes, Apache attaches all merits
>> to the individual. But you cannot reasonably expect individuals
>> that got paid for working on an Apache project to continue their
>> involvement at a comparable level on private time, nor "judge"
>> them for retiring. The ultimate cause of reduced activity here
>> would be the employer's decision, not the individual's.
>
> You are absolutely right...
Yep, Roland and Bill, I agree. Thanks to the efforts of Ken Coar and
others, IBM, "the corporate entity", has organizational knowledge (read
"process") for formally reviewing proposals to contribute code to open
source communities. I've personally shutdown more than a few IBM
proposals that would have been 'code dumps'. You might be surprised at
how I most often do it... I show them how difficult it is to get through
the internal review process . When they discover how hard it is to
'dump' code, they quickly lose interest. No one here would waste their
time going to the effort to 'dump' code. The word 'dump' is a really
bad metaphor. So is the process perfect? Of course not and that's a
silly question as there is no such thing as a perfect process in the
real world. I'm certain some IBM sponsored projects have not lived up
to expectations.
Open invitation to anyone reading this... if you detect a pattern of IBM
dumping code, let me know. It's certainly possible that our internal
process for vetting proposals is not strict as it should be. Convince
me there is a problem and I will work to solve it on the IBM side.
Bill
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Santiago Gala <sa...@gmail.com>.
On Feb 4, 2008 4:54 PM, William A. Rowe, Jr. <wr...@rowe-clan.net> wrote:
> Roland Weber wrote:
> > I think that is a bit oversimplified. IBM has strict rules about
> > open source participation. It is either "on private time", such
> > as my involvement at Apache. Then the person is acting as an
> > individual. Or it is "on company time". Then the person is doing
> > what he or she is paid for. And if IBM is changing it's priorities,
> > or the line item that required OSS participation is closed, plenty
> > of other work will be dumped on that person, simply leaving no
> > (work) time for OSS participation. Yes, Apache attaches all merits
> > to the individual. But you cannot reasonably expect individuals
> > that got paid for working on an Apache project to continue their
> > involvement at a comparable level on private time, nor "judge"
> > them for retiring. The ultimate cause of reduced activity here
> > would be the employer's decision, not the individual's.
>
> You are absolutely right...
>
> Perhaps we should force all initial committers to divulge if they
> are strictly involved in the effort as a work assignment, or if they
> have a broader interest in the new podling?
>
> And certainly, we should judge contributing corporations on their
> prior projects successes and failures, and this should be one of
> the many factors that go into the +/-1 decision of accepting a project.
> Not the only factor, but one of many. The failure of corporations
> to 'play nice with each other' is also one of those factors, if they
> are capable of participating in an open, transparent and collaborative
> development methodology required at and by the ASF.
>
> That said, we never "judge" people per-say for choosing to move on
> to some other projects or interest in their lives. The code is here
> for the public, and if the public can't be bothered to contribute,
> then it's simply shelved. No different than any commercial technology
> when a company looses interest in it.
>
+1 I think the relevant issue for incubation efforts is wether there
is a reasonable expectation that the current ("dropped") code base
will attract enough people from outside the donating company to
graduate. I say "reasonable expectation" and not "certainty", as
incubation is a bet, and it can succeed or not.
Once incubation is going on everything can happen, from success to
withdrawal before graduation, passing through stagnation before or
after graduation.
Regards
Santiago
> Bill
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Roland Weber <os...@dubioso.net>.
William A. Rowe, Jr. wrote:
>
> Perhaps we should force all initial committers to divulge if they
> are strictly involved in the effort as a work assignment, or if they
> have a broader interest in the new podling?
+1
> That said, we never "judge" people per-say [...]
I know. Just couldn't resist the pun ;-)
cheers,
Roland
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by "William A. Rowe, Jr." <wr...@rowe-clan.net>.
Roland Weber wrote:
> I think that is a bit oversimplified. IBM has strict rules about
> open source participation. It is either "on private time", such
> as my involvement at Apache. Then the person is acting as an
> individual. Or it is "on company time". Then the person is doing
> what he or she is paid for. And if IBM is changing it's priorities,
> or the line item that required OSS participation is closed, plenty
> of other work will be dumped on that person, simply leaving no
> (work) time for OSS participation. Yes, Apache attaches all merits
> to the individual. But you cannot reasonably expect individuals
> that got paid for working on an Apache project to continue their
> involvement at a comparable level on private time, nor "judge"
> them for retiring. The ultimate cause of reduced activity here
> would be the employer's decision, not the individual's.
You are absolutely right...
Perhaps we should force all initial committers to divulge if they
are strictly involved in the effort as a work assignment, or if they
have a broader interest in the new podling?
And certainly, we should judge contributing corporations on their
prior projects successes and failures, and this should be one of
the many factors that go into the +/-1 decision of accepting a project.
Not the only factor, but one of many. The failure of corporations
to 'play nice with each other' is also one of those factors, if they
are capable of participating in an open, transparent and collaborative
development methodology required at and by the ASF.
That said, we never "judge" people per-say for choosing to move on
to some other projects or interest in their lives. The code is here
for the public, and if the public can't be bothered to contribute,
then it's simply shelved. No different than any commercial technology
when a company looses interest in it.
Bill
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Roland Weber <os...@dubioso.net>.
Bill Stoddard wrote:
> Disclosure... I work for IBM.
So do I.
> IBM'ers participate on projects as individuals and it's the actions of
> individuals that should be judged.
I think that is a bit oversimplified. IBM has strict rules about
open source participation. It is either "on private time", such
as my involvement at Apache. Then the person is acting as an
individual. Or it is "on company time". Then the person is doing
what he or she is paid for. And if IBM is changing it's priorities,
or the line item that required OSS participation is closed, plenty
of other work will be dumped on that person, simply leaving no
(work) time for OSS participation. Yes, Apache attaches all merits
to the individual. But you cannot reasonably expect individuals
that got paid for working on an Apache project to continue their
involvement at a comparable level on private time, nor "judge"
them for retiring. The ultimate cause of reduced activity here
would be the employer's decision, not the individual's.
While I think that Endre's concerns are very real and justified,
there are two points I'd like to mention in this context:
1. This is not specific to IBM. It is a danger to all projects
with corporate backing, whether that comes from IBM or BEA
or SUN or whomever.
2. Community diversity is an exit criteria, not an entry criteria.
cheers,
Roland
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Bill Stoddard <bi...@wstoddard.com>.
Endre Stølsvik wrote:
> Leo Simons wrote:
>
>> Sure, activity is not that high, and there's not a *huge* developer
>> community, but there does not really seem to be any problem, either.
>> Apache doesn't require projects to be huge successes (by whatever
>> metric) as long as they're healthy and self-sustaining.
>
> This was not healthy up until quite recently - I'd even question
> whether the code and the community is _healthy_ now. And the main
> point was that IBM, the dumpsters, fled the plot pretty fast after the
> import was done, not bothering about creating The Healthy Community
> around the code they off-loaded before leaving.
>
> This should raise at least some slight concerns in Apache, IMO.
Endre,
Disclosure... I work for IBM.
I agree with your concern about code dumping and I completely respect
your comments. However, I object to the logic you are applying here.
IBM'ers participate on projects as individuals and it's the actions of
individuals that should be judged. To tar all IBM'ers because of bad
behavior (perceived or real) of a few is just wrong.
Bill
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Endre Stølsvik <En...@stolsvik.com>.
Leo Simons wrote:
> Sure, activity is not
> that high, and there's not a *huge* developer community, but there does
> not really seem to be any problem, either. Apache doesn't require
> projects to be huge successes (by whatever metric) as long as they're
> healthy and self-sustaining.
This was not healthy up until quite recently - I'd even question whether
the code and the community is _healthy_ now. And the main point was that
IBM, the dumpsters, fled the plot pretty fast after the import was done,
not bothering about creating The Healthy Community around the code they
off-loaded before leaving.
This should raise at least some slight concerns in Apache, IMO.
>
> However, much more importantly, proposals should be evaluated on their
> own merits, not based on what happened to some other unrelated project 4
> years ago.
How is that project unrelated?? It is a JSR-RI/TCK implementation that
supposedly shall be "done at Apache", one of the companies that come
with the proposal and code is the very same company that did the Pluto
stuff, and the code is delivered in the form of a finished product: a
complete RI, and a complete TCK.
Out of actual curiosity, what is their actual interest in having this at
Apache, rather than using Sourceforge or Google Code if the single aim
is getting this published in an open source fashion?
(On a tangent, I really would expect a Reference Implementation and a
Test Compatibility Kit to be quite bare bones, both pretty much merely
defining the standard, and to NOT evolve much. Changes would only be
fixes of actual deviations from the specification. That would be a
rather different aim than what one would want of a fully fledged product)
It is quite interesting to see the amount of action, questions and
hassle many other projects are instantly getting when being proposed,
for example the current project Thrift, compared to the meager attention
this one has gotten. The spectacular amount of answers e.g. Paul's
questions have elicited from the supposed proposers is also striking.
But in the end, I believe this project will magically just happen anyway.
Endre.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Leo Simons <ma...@leosimons.com>.
On Feb 1, 2008, at 2:19 PM, Endre Stølsvik wrote:
> this isn't exactly some court
Exactly.
Now, if I look through
http://apache.markmail.org/search/?q=pluto#query:pluto+page:1
+state:facets
pluto does not seem a problematic "code dump" project, and it also
definitely isn't a "single developer" project. Sure, activity is not
that high, and there's not a *huge* developer community, but there
does not really seem to be any problem, either. Apache doesn't
require projects to be huge successes (by whatever metric) as long as
they're healthy and self-sustaining.
However, much more importantly, proposals should be evaluated on
their own merits, not based on what happened to some other unrelated
project 4 years ago.
cheers!
- Leo
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: [Proposal] NoNameYet - Pluto
Posted by Endre Stølsvik <En...@stolsvik.com>.
Stefan Hepper wrote:
> It is not true that after the JSR was final everything stopped. In fact
> once we had finished 1.0 there was still work done to get to a more
> stable 1.0.1 release. After that the pluto community re-structed the
> code which led to the pluto 1.1 stream, so you can see that it was an
> active community not only some code drop.
I mentioned this.
JSR 168 itself was "Final Release" on 27 Oct, 2003.
Wading through the repository:
The first entry of Pluto in Subversion is Tue Sep 30 14:03:01 2003
(cvs2svn import).
Pluto 1.0.1-RC1 was tagged Thu Oct 7 09:25:40 2004.
Pluto 1.0.1-RC4 was tagged Mon Jul 25 01:06:54 2005.
Pluto 1.0.1 was "ReTagged" Mon Oct 10 18:29:33 2005.
1.0.1, the first minor-revision above the inital code drop, being a fix
of the most crazy bugs, was tagged more than two years after the inital
code drop.
Pluto 1.1.0 was tagged "[maven-scm] copy for tag pluto-1.1.0" Sun Feb 11
17:29:23 2007.
1.1.0 was, IIRC, a larger refactor to make the code somewhat more
embeddable.
If the tags don't lie, the 1.1.0 came out less than a year ago, 3 years
and 4 months after the initial drop. This time period isn't really
extreme in itself, but compared to the developer activity on a project
that literally screamed for help, it becomes sad.
The following fact I'm not _quite_ certain of, and it takes a bit too
much time to search the archives to prove it (this isn't exactly some
court), but I do believe I have my words intact if I state that there
wasn't much involvement in any of those versions after the initial dump
to come from any of you IBM folks. Repos and mailing list archives are
available. And for my comments of code quality - I might be wrong, I
might be a nit-picker, or I might just be a bad coder - but the initial
drop and all the versions are still there in the repository: have a
look-see.
There might be some tools better than ViewVC to run through the
repository and mailing lists to get a better picture of these facts. But
AFAIK, and I've been lurking lots (due to the fact that I had interests
in a embeddable standards compliant container at the time), Pluto has
*never* had much of an active community. I believe most of the 1.1.0 was
done by one man.
Had the initial code been of quite a bit better quality, or had it
had quite a bit more backing from the droppers to bootstrap any
community, it had at least had ONE more active developer, that I can
guarantee. I tried, but I could just not start anywhere with that code,
and I could not start doing a complete rewrite of that dump on my own
(I'd actually rather start from scratch, to be honest).
> Also in my view pluto was a success, take a look at all the projects
> using the pluto container: http://portals.apache.org/pluto/powered.html.
Compared to the amount of portals, this isn't exactly amazing. That IBM
doesn't use it itself (I belive?), an Apache product which they
themselves created, is quite telling, compared to the fact that IBM
isn't exactly shy of using lots of other Apache products in their
products and services.
Endre.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org