All of that formal involvement takes a hell of a lot of time.
As I see it they only had 2 options:
1. The path they took
2. Grab an open-source OSGi implementation whose implementation and
license were amenable and bundle and fork it as needed
Working with standards bodies in this context is a recipe for wasting
countless hours of engineer's time and countless schedule delays.
Informal conversations and discourse around approach #2, followed by an
attempt to merge the fork back into the main implementation and standard
afterwards is about all you possibly could have hoped for. The fact
that they took path #1 likely indicates a judgement call that hacking an
OSGi implementation down to the subset they actually needed, quality
checking every bit thereof, and then adding what it was missing would
take longer than just writing what they needed.
I can't really disagree with the decision.
--
Jess Holle
Eric Newcomer wrote:
> Yes, that's what they keep saying. I'm sure they mean that they speak
> when they see each other in the hallways at Java One, or after a
> presentation. Or something like that.
>
> But there is no participation in any OSGi expert group by Jigsaw
> folks. And there is no Jigsaw JSR for anyone else to participate in.
>
> The Jigsaw folks can say anything they want to anyone informally - and
> they do say a lot of things - but there is no formal exchange of ideas
> and proposals. That is the point. Without that, there's no way to
> bring things together.
>
> Eric
>
> On Mon, Jun 22, 2009 at 5:32 PM, Reinier Zwitserloot
> <reini...@gmail.com <mailto:reini...@gmail.com>> wrote:
>
>
> But they do communicate with the OSGi team.
>
> On Jun 22, 7:43 pm, Eric Newcomer <enewco...@gmail.com
> <mailto:enewco...@gmail.com>> wrote:
> > No, the point is that Jigsaw proposes to be a standard. OSGi
> already is
> > one, and I don't mean in the academic sense. I mean that it has
> been widely
> > adopted and used. My point is that the Jigsaw folks are talking
> about their
> > effort as if it were already the equal of OSGi, when it is far
> from it.
> >
> > I wish I knew what you guys were all talking about with respect
> to the OSGi
> > supporters vs Jigsaw supporters. To me this sounds like the
> principal
> > arriving in the middle of a fight at the schoolyard, and not
> knowing who
> > started it, can only blame each fighter equally.
> >
> > If we in the OSGi community invite the Jigsaw folks to
> participate (and Sun
> > is definitely a member of OSGi as well as the relevant expert
> groups), and
> > they say they will, but then they don't, and we complain about
> that, is that
> > a valid compliant or not?
> >
> > Eric
> >
> > On Mon, Jun 22, 2009 at 12:43 PM, Lloyd Meinholz
> > <meinh...@javabilities.com <mailto:meinh...@javabilities.com>>wrote:
> >
> > > Being a standard doesn't necessarily make it a good thing.
> Corba, EJB 1 and
> > > 2 are standards. Spring and Hibernate even eclipse are defacto
> standards,
> > > but not official standards. Emphasizing OSGI being a standard
> doesn't help
> > > you argument IMO. I also don't agree that retrofitting OSGI to
> meet the
> > > needs of Jigsaw is quicker. It may be, but not necessarily.
> >
> > > As a total outsider to the modularization stuff, I have to say
> that my
> > > experience is that the OSGI supporters that I see appear to be
> much more
> > > irrational than the Jigsaw supporters. That turns me off and I
> think hurts
> > > adoption of this technology. The same kind of community
> behaviour is one
> > > reason I really don't care for RoR. Yes, I could just grow up
> and ignore the
> > > jerks, but why not work with a group that doesn't include
> them? There are
> > > plenty of interesting things to work on.
> >
> > > Lloyd
> >
> > > On Mon, Jun 22, 2009 at 10:34 AM, Eric Newcomer
> <enewco...@gmail.com <mailto:enewco...@gmail.com>>wrote:
> >
> > >> But this is a completely nonsensical argument.
> Standardization is slow
> > >> because developing a standard means that you need to get
> people to buy into
> > >> it and adopt it. That takes time. There are no shortcuts.
> The same will
> > >> have to be true for Jigsaw. If it really gets adopted it
> will take a long
> > >> time.
> >
> > >> Are you suggesting it doesn't matter whether Eclipse, IBM,
> Oracle,
> > >> Progress, etc. adopt it? If it does matter it will take time.
> >
> > >> There is no magic in the world that anyone can use to anoint
> anything a
> > >> standard. This is a really false argument on its face. You
> cannot just
> > >> wish for something to be adopted and have it happen.
> >
> > >> That's why I say we are dealing with a false debate here. We
> are drawn
> > >> into a comparison of Jigsaw *as it might become* versus OSGi
> *as it already
> > >> is*. OSGi is a standard - Jigsaw is not. We can suppose
> that Jigsaw might
> > >> become a standard, but can we really say the time it will
> take for Jigsaw to
> > >> be as widely adopted as OSGi would be time well spent?
> >
> > >> Eric
> >
> > >> On Mon, Jun 22, 2009 at 10:27 AM, Jess Holle <je...@ptc.com
> <mailto:je...@ptc.com>> wrote:
> >
> > >>> I think one of the most compelling arguments cited was
> actually *not*
> > >>> getting necessary changes worked out through the OSGi
> standards body.
> >
> > >>> Standards body == slow.
> >
> > >>> Given that there are clear things OSGi is missing that
> Jigsaw needs and
> > >>> the non-goal of Jigsaw replacing OSGi, bypassing it makes sense.
> >
> > >>> There are other things that really gave me pause (e.g. the
> emphasis on
> > >>> native packaging to the exclusion of cross-platform portable
> packaging
> > >>> approaches!!!), but by-passing OSGi seems like a non-issue.
> >
> > >>> --
> > >>> Jess Holle
> >
> > >>> P.S. I'd contrast this with by-passing log4j, for instance,
> where log4j
> > >>> could have easily been extended to do everything Sun had in
> mind and had no
> > >>> weighty standards body to contend with.
> >
> > >>> Eric wrote:
> >
> > >>> This time issue is another false argument that attempts to
> justify
> > >>> incorrect behavior - and I mean incorrect with regard to the
> Java
> > >>> community.
> >
> > >>> It takes a lot more time to start something from scratch
> than to work
> > >>> with something that already exists. I don't just mean the
> development
> > >>> of something, but its adoption as a standard. Unless, of
> course,
> > >>> that's not what Jigsaw folks intend.
> >
> > >>> Eric
> >
> > >>> On Jun 20, 10:24 pm, Josh Suereth <joshua.suer...@gmail.com
> <mailto:joshua.suer...@gmail.com>> <joshua.suer...@gmail.com
> <mailto:joshua.suer...@gmail.com>> wrote:
> >
> > >>> Can we stop the name-calling? On the scala list serves, I
> usually reserve
> > >>> pictures of fluffy,furry, cute and cuddly kittens to help
> quell heated
> > >>> pointless arguments.
> >
> > >>> If you listen to the podcast again, you'll notice they
> specifically say that
> > >>> they didn't feel Sun could afford the time waiting for the
> OSGi-alliance to
> > >>> update themselves based on Sun's requirements. To this
> extent I do agree
> > >>> that Java really needs modularity. However it is also
> understandable how
> > >>> this would make the OSGi camp feel. Sun is basically
> saying, "Although you
> > >>> may have a good product, we can't affrod to wait around for
> your standards
> > >>> committee. We're going to do what we need and let OSGi
> follow along." This
> > >>> is far different from Sun's previous approaches (think EJB).
> >
> > >>> The good news is that it sounds like the Jigsaw + OSGi folks
> are at least
> > >>> talking. I don't think a merged approach to modularity will
> take long to
> > >>> follow the release of JDK 7. That is of course, assuming
> Jigsaw succeeds.
> >
> > >>> I think the biggest complaint coming from OSGi users (myself
> included) is
> > >>> that I don't want to have to deal with the complexity of 2
> modularity
> > >>> tools. It's painful enough dealing with 2-3 different
> logging mechanisms
> > >>> accross 4-5 libraries. Competition is good, but integration
> is hard.
> >
> > >>> I'm just waiting for the modularity communities to start
> making "abstract
> > >>> modules" that define services that can be implemented by
> venders. Anyone
> > >>> remember CORBA?
> >
> > >>> - Josh
>
>
>
> >
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "The
Java Posse" group.
To post to this group, send email to javaposse@googlegroups.com
To unsubscribe from this group, send email to
javaposse+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/javaposse?hl=en
-~----------~----~----~----~------~----~------~--~---
