felix-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Richard S. Hall" <he...@ungoverned.org>
Subject Re: Implementation of unreleased spec and community
Date Wed, 18 Jan 2017 23:40:48 GMT
On 1/18/17 18:03 , Guillaume Nodet wrote:
> The problem is when the "what" is being prototyped in an Apache project,
> not the "how".  The "how" can be done by any committer, but if the "what"
> is still being discussed, you really have no saying over the "how" yet.
> Basically, no one will ever question an implementation decision until the
> spec work is finished.  Which means that no one will question anything
> until the implementation (if that's the RI) is finished too, at which
> point, there's nothing to say anymore, unless you want to rewrite the whole
> thing.
> That's definitely not community work.
> So you're right, it's temporary until the spec work is published, at which
> point the implementation is also finished and it's too late.  Especially if
> the RI will have to be released roughly at the same time than the spec.
> Ray has listed a number of things that have been implemented during the
> past few months.  All of them have been written by a single committer who
> also happen to be the one modifying the spec document.  So I don't see any
> real ground for invoking a community involvement, when there's definitely
> none.  If what you think about community work is limited to reading
> existing code and eventually providing a few comments, those things can
> easily be done anywhere.

I disagree that it is not possible to have say over the "how" if you 
don't have any say over the "what" while it is being developed, unless 
there is an active effort by the people involved to not answer questions 
and/or hide what they are doing.

Is it more work when you don't have a complete spec? Sure, because it 
involves actually engaging others to get answers or combing through code 
to figure it out (like you have to do in most open source projects).

I'm not sure this discussion is bearing any fruit. Is there something 
you'd like to propose? If I were to agree that this is an issue, then 
the only thing I can see to resolve it is banning all work on 
provisional specs, which definitely doesn't seem like a good course of 

-> richard

> Guillaume
> 2017-01-18 21:52 GMT+01:00 Richard S. Hall <heavy@ungoverned.org>:
>> On 1/18/17 15:28 , Guillaume Nodet wrote:
>>> 2017-01-18 20:23 GMT+01:00 Richard S. Hall <heavy@ungoverned.org>:
>>> On 1/18/17 14:06 , Guillaume Nodet wrote:
>>>> Let's take a clearer example, as I have a feeling I'm still not
>>>>> understood
>>>>> correctly.  My problem is definitely not the fact that there is an
>>>>> implementation based on an unreleased spec or RFC (as my email title
>>>>> seemed
>>>>> to indicate).
>>>>> If a committer comes and say : I'd like to implement rfc-xxx based on
>>>>> the
>>>>> public document (spec or rfc), that's very fine with me, because all
>>>>> committers have the same level of information and can get involved.
>>>>>    Fwiw,
>>>>> that's what usually happens and I haven't seen anything different in
>>>>> Felix
>>>>> so far.
>>>>> If someone comes and say: I'd like to work on some code which will
>>>>> eventually become the RI while writing the rfc within the OSGi
>>>>> alliance, I
>>>>> don't think that's fine, because what we have in this case is not an
>>>>> implementation of something, it's a prototype for designing the spec
>>>>> only OSGi Alliance members who are actually working on the rfc can
>>>>> really
>>>>> change the api.
>>>>> Is that more clear now ?  Am I the only one thinking that if not all
>>>>> committers can work on the code, there's a real problem ?
>>>>> This is what I assumed that you meant and I don't really see an issue
>>>> with
>>>> it. Yeah, it might be more painful than if there was a public document
>>>> somewhere, but this is a little bit of a chicken-and-egg situation that
>>>> will eventually clear itself up.
>>>> http://community.apache.org/committers/decisionMaking.html
>>> The first phrase is the following:
>>>      The most important thing about engaging with any Apache project is
>>> that
>>> everyone is equal.
>> They are equal in the eyes of Apache, but that doesn't mean that they are
>> equal with respects to their external memberships. You appear to want to go
>> down a rat hole that has no escape. No matter what, committers who are
>> members of standards bodies are "not equal" to committers who are not in
>> the eyes of the standards bodies. This is true with regardless of whether
>> the spec is released or not, so I'm still not sure of your point.
>> People in the standards body have more say in how things will go in what
>> will ultimately be implemented, but they do not have any more say in how it
>> will be implemented in any given Apache project. They just get a leg up on
>> implementing it before the spec is completely public, but that doesn't make
>> their kung-fu more equal than anyone else and they will still be
>> accountable for any technical discussions/decisions/debates that ensue as a
>> result.
>>> Of course, if you had someone purposely trying to keep people in the dark,
>>>> then this might be an issue, but I don't think this is a real concern and
>>>> certainly not something we've ever experienced. I have to assume that
>>>> someone doing the implementation work here would want to discuss with
>>>> other
>>>> people and get input, otherwise why would they be doing the work here in
>>>> the first place?
>>> Well, I have to assume that someone doing the implementation work here
>>> would want to obey the Apache rules, otherwise why would they be doing the
>>> work here in the first place?
>> But you seem to be implying that they aren't obeying the rules because
>> they have access to unreleased specs which doesn't really hold water with
>> me.
>> The only thing that you are saying that is true is that people will have
>> some difficulty contributing if they don't have the spec to read. Sure,
>> this is probably the case, but this certainly doesn't prevent people from
>> commenting and/or contributing to it. Most open source work doesn't have a
>> spec for people to work against and they still generally jump in and
>> contribute. And, as I said before, it is only temporary.
>> -> richard
>>>> -> richard
>>>>> 2017-01-18 15:29 GMT+01:00 Neil Bartlett <njbartlett@gmail.com>:
>>>>> On 18 Jan 2017, at 12:36, Guillaume Nodet <gnodet@apache.org> wrote:
>>>>>> Fwiw, I think Christian was referring to the JAX-RS WHITEBOARD, not
>>>>>>> JAX-RS spec itself.
>>>>>>> That one is an RFC from the OSGi Alliance...  RFC-127 afaik.
>>>>>>> This is pretty much my point. Why raise an issue with the “Whiteboard”
>>>>>> half of “JAX-RS Whiteboard” but not with the “JAX-RS” half?
Why don’t
>>>>>> your
>>>>>> arguments also apply to JCR specs, or IEEE or W3C specs?
>>>>>> Regards,
>>>>>> Neil
>>>>>> 2017-01-18 13:34 GMT+01:00 Neil Bartlett <njbartlett@gmail.com>:
>>>>>>> Christian, your example of JAX-RS Whiteboard is fascinating,
>>>>>>>> JAX-RS was designed by the Expert Groups of the JCP, not
by the
>>>>>>>> Apache
>>>>>>>> community. The same is true of many of the JavaEE specifications
>>>>>>>> implemented within Apache.
>>>>>>>> So, Apache has always worked pragmatically to implement
>>>>>>>> specifications
>>>>>>>> emerging from external standards bodies. It seems odd therefore
>>>>>>>> single
>>>>>>> out OSGi.
>>>>>>>> Neil
>>>>>>>> On 18 Jan 2017, at 11:25, Christian Schneider <
>>>>>>>> chris@die-schneider.net
>>>>>>>> wrote:
>>>>>>>> I agree with Guillaume that the way the specs are defined
is not
>>>>>>>>> fully
>>>>>>>>> compatible to the way apache projects are managed.
>>>>>>>> In apache the idea is that the design of a component is defined
>>>>>>>>> the
>>>>>>>>> community.
>>>>>>>> Like in jax-rs-whiteboard .. if it was a pure apache thing
>>>>>>>>> changes
>>>>>>>>> in the interfaces would be proposed on the dev list and
agreed on
>>>>>>>> there.
>>>>>>>> As the interfaces are part of the spec this is out of direct
>>>>>>>>> for
>>>>>>>>> the aries community.
>>>>>>>> On the other hand I understand that the final decision about
the spec
>>>>>>>>> has to be at the OSGi alliance and even that only members
>>>>>>>> decide.
>>>>>>>> So I think this gap can not be fully solved but maybe we
can improve
>>>>>>>>> it.
>>>>>>> So what I could imagine is this:
>>>>>>>> - Changes on the spec should be immediately visible to the
>>>>>>>>> community. This could be done using a github repo where
the source
>>>>>>>> of
>>>>>>>> the
>>>>>>> spec resides and an automated snapshot build. So all changes
could be
>>>>>>>> followed directly and the newest spec jars  would always
>>>>>>>> available.
>>>>>>>> - Protocols of the expert group meetings could be posted
to the dev
>>>>>>>>> list
>>>>>>> Both improvements would shorten the feedback loop and give the
>>>>>>>> community at least more visibility of the spec progress.
>>>>>>>> community
>>>>>>>> could then also directly give feedback to the protocols as
well as
>>>>>>>> api
>>>>>>>> changes on the dev list. So this would of course still not
allow the
>>>>>>>> apache
>>>>>>> community to drive the spec but I think it would be a good compromise.
>>>>>>>> Christian
>>>>>>>>> On 18.01.2017 11:59, David Bosschaert wrote:
>>>>>>>>> Hi Guillaume,
>>>>>>>>>> First of all, the OSGi Alliance is a very open standards
>>>>>>>>>> development
>>>>>>>>>> organization. Any organisation can join. RFPs and
RFCs are
>>>>>>>>>> developed
>>>>>>>>>> in
>>>>>>>> the
>>>>>>>> open, specs are available for free and are free to be implemented
>>>>>>>>> anyone.
>>>>>>>>> There is also an open feedback channel available where
everyone can
>>>>>>>>> post
>>>>>>>> feedback, described at https://github.com/osgi/design
>>>>>>>> OSGi works very hard in defining specs that are portable
and can be
>>>>>>>>>> implemented without the need to pay for any licenses
or anything of
>>>>>>>>>> that
>>>>>>>> sort.
>>>>>>>> History has shown that spec implementations are really quite
>>>>>>>>>> portable.
>>>>>>>>>> Implementation bundles can be mixed from different
sources and
>>>>>>>>>> everything
>>>>>>>>> just works as long as you use the specced APIs.
>>>>>>>>>> Every new spec that is being worked on in OSGi needs,
besides the
>>>>>>>>>> RFP/RFC
>>>>>>>>> and spec chapter, a Reference Implementation and a Conformance
>>>>>>>>> Testsuite.
>>>>>>>>> Over the past 10 years or so, Reference Implementations
>>>>>>>>> primarily
>>>>>>>>> been
>>>>>>>>> implemented in open source. This has the benefit that
everyone can
>>>>>>>>> see
>>>>>>>>> what
>>>>>>>>> the implementation is going to be and also it allows
everyone to
>>>>>>>>> provide
>>>>>>>> feedback and participate in the implementation. Apache committers
>>>>>>> have
>>>>>>>> free
>>>>>>>>> access to the relevant CTs as well.
>>>>>>>>>> I think this is all goodness. Or would you rather
see that
>>>>>>>>>> Reference
>>>>>>>>>> Implementations are implemented in private?
>>>>>>>>>> Best regards,
>>>>>>>>>> David
>>>>>>>>>> On 18 January 2017 at 10:41, Guillaume Nodet <gnodet@apache.org>
>>>>>>>>>> wrote:
>>>>>>>> I'm a bit concerned by some subprojects in our communities.
>>>>>>>> The ASF is supposed to be "community over code", so the very
>>>>>>>>>>> thing
>>>>>>>>> for a project is that people can get involved.
>>>>>>>>>> However, I see more and more code developped as a
>>>>>>>>>>> implementation
>>>>>>>>> of a spec which is not publicly available, because it's
still being
>>>>>>>>>> developed at the OSGi Alliance.  I find that very
>>>>>>>>>>> because
>>>>>>>>>>> there's no way the community can get involved
unless they are OSGi
>>>>>>>>>>> Alliance
>>>>>>>>> members, and that's clearly not acceptable imho.
>>>>>>>>>> Thoughts ?
>>>>>>>>>>> Guillaume Nodet
>>>>>>>>>>> --
>>>>>>>>> Christian Schneider
>>>>>>>>> http://www.liquid-reality.de
>>>>>>>>> Open Source Architect
>>>>>>>>> http://www.talend.com
>>>>>>>>> --
>>>>>>> ------------------------
>>>>>>> Guillaume Nodet
>>>>>>> ------------------------
>>>>>>> Red Hat, Open Source Integration
>>>>>>> Email: gnodet@redhat.com
>>>>>>> Web: http://fusesource.com
>>>>>>> Blog: http://gnodet.blogspot.com/

View raw message