incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Graham Charters <>
Subject Re: [PROPOSAL] Apache Aries incubator for Enterprise OSGi
Date Fri, 04 Sep 2009 13:27:23 GMT
Having read all the discussions, I still have concerns about the
suggestion to put all OSGi spec implementation under Felix.  I don't
see this approach being taken for other specification organizations
(JCP, OASIS, etc.) and I think that is to the benefit of Apache.  For
example, whilst a goal of Geronimo is JEE compliance it does not host
all the implementations of the JEE specifications.  Having OpenEJB,
OpenJPA, Tomcat, etc as separate projects allows them to evolve
independent communities, with cross-pollination.  The domains,
technologies and skills required to cover the entirety of JEE are
diverse and the same is true for OSGi.  To put these under one project
would lead to sub-groups of different interests under one disjoint
umbrella community.

I also worry about the practicalities of saying that all OSGi spec
implementations belong in Felix.  This is a moving target, both within
each specification, where extensions may be created and then
standardized and also at the spec level, where new concepts are
created and then standardized.  Must these then move under the Felix
umbrella, which will undoubtedly create a lot of churn for users of,
and contributors to, Aries?

Thinking about the problem the way JEE is handled makes me wonder if
Felix could pull in spec implementations from Aries (and have
committers on Aries) and therefore still contribute to and provide the
distribution of the entire OSGi Service Platform.  Maybe this could be
done through Felix hosting a bundle repository (Felix commons?)?
Consumers of Felix would be able to get the OSGi Service Platform and
consumers of Aries would get the enterprise OSGi application portions.

I'd like to emphasize that the desire to be separate from Felix is in
no way a criticism of that project.  I have a huge amount of respect
for what Richard et al continue to achieve and having experienced
first-hand the value the combination of Felix and Equinox bring to the
OSGi standards, I sincerely believe that OSGi is far stronger as a

The desire to remain separate is based purely on the belief that this
is the approach which will best serve the goals of evolving an
enterprise OSGi application programming model, and a community around
its definition.  This rationale is also the reason why Geronimo was
not suggested (an equally valid choice given the majority of the
Enterprise OSGi specifications leverage JEE technologies).  I guess
what I don't understand is why the Incubator would not prove or
disprove this belief, without making a prior decision?

Regards, Graham.

2009/9/4 Daniel Kulp <>:
> As a point of note, not all OSGi spec implementations live in Felix even at
> Apache today.   The Remote Services/Distributed OSGi reference implementation
> is a sub project of CXF.     I think Tuscany has an implementation as well.
> So far, there hasn't been any discussion about moving those into Felix.  Your
> argument below makes it sound like they should be.
> Dan
> On Thu September 3 2009 1:33:04 pm Richard S. Hall wrote:
>> There was no attempt to contact the Felix PMC in general that I am aware
>> and I certainly didn't know about it in advance.
>> And there seems to be a continued attempt to construe my original
>> criticisms as "all of Aries should go into Felix".
>> I, personally, do not believe that all of Aries should go into Felix, I
>> too think it should have its own identity. I was always only ever
>> referring to the independent OSGi spec implementations. I was arguing
>> that Felix is a good place to work on them, since it is part of what it
>> is trying to achieve.
>> Further, I don't really understand the implication that somehow the
>> burden is now on the Felix community to go and contribute to Aries on
>> OSGi spec implementations just because of this proposal, when there was
>> no attempt to work with the Felix community on creating OSGi spec
>> implementations in the first.
>> The only conclusions I see being drawn by people who have invested very
>> little in Felix is that we should dismantle the Felix charter so that we
>> can accommodate the fact that some people don't want to play with us.
>> At that rate, I stand by my previous "vote" and otherwise people can do
>> whatever they want in Aries.
>> -> richard
>> On 9/3/09 13:23, Niclas Hedhman wrote:
>> > Kevan,
>> >
>> > Was a contact with Felix made prior to dropping the proposal here? I got
>> > the impression that wasn't the case, which I find surprising... If I am
>> > wrong, what was the meat of such?
>> >
>> > I'm also less happy with the rhetoric here repeated over and over,
>> > seemingly uninterested in discussion of reaching a solution everyone can
>> > accept. (From both camps, btw)
>> >
>> > -- Niclas
>> >
>> > On Sep 4, 2009 12:53 AM, "Kevan Miller"<>  wrote:
>> >
>> > On Sep 3, 2009, at 12:53 AM, Niclas Hedhman wrote:>  On Thu, Sep 3, 2009
>> > at 3:19 AM, William A. Ro...
>> > Totally agree. Had certainly hoped that Felix committers would be
>> > interested in joining...
>> >
>> > --kevan
>> >
>> > --------------------------------------------------------------------- To
>> > unsubscribe, e-mail: gene...
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail:
>> For additional commands, e-mail:
> --
> Daniel Kulp
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message