The Artima Developer Community
Sponsored Link

Weblogs Forum
Community Access to the Jini Starter Kit Codebase

8 replies on 1 page. Most recent reply: Apr 6, 2004 9:24 AM by Kim

Welcome Guest
  Sign In

Go back to the topic listing  Back to Topic List Click to reply to this topic  Reply to this Topic Click to search messages in this forum  Search Forum Click for a threaded view of the topic  Threaded View   
Previous Topic   Next Topic
Flat View: This topic has 8 replies on 1 page
Dale Asberry

Posts: 161
Nickname: bozomind
Registered: Mar, 2004

Community Access to the Jini Starter Kit Codebase (View in Weblogs)
Posted: Apr 2, 2004 1:45 PM
Reply to this message Reply
Summary
At the 7th Annual Jini Community Meeting, one of the discussions was about the fact that Sun did not have a version of the Jini Starter Kit in a publicly visible CVS tree.
Advertisement
I feel that having the entire source tree available would be a good thing for the community. For those concerned about forks or other nonsense, then the list of committers can remain solely in the power of the Sun Jini team.

Last year I submitted a patch to the Axis group for a very serious bug that was preventing me from moving forward. I wrote the fix, created a diff (according to their standards), along with a unit test that revealed the bug, and then I submitted the bug to the core team. An hour later, a committer put the code into the CVS tree. From what I've seen, this is considerably faster than the Jini team's response. Not only that, but anyone else who encountered the same problem could just grab the particular file from CVS and their problem would be fixed. In fact, that process helped me resolve another issue I was running into. Yeah, the codebase differed slightly from the release codebase, but everything worked. Since I had not intentions of "distributing" the code to additional parties, the problem of code drift was not an issue.

With the time between release dates for most projects exceeding a year (or more), source access allows each community member to decide if possible incompatibilities are worth the benefits.

The Jini team has limited resources and I believe could greatly benefit from the OSS process of many people chipping away small pieces of the problem - bugs or new features. We (the Jini Community) all know the power of worker processes grabbing tasks out of a common space: by publishing issues and enhancements and accepting patches and tests in return, the community would benefit from faster JSTK releases.


Dale Asberry

Posts: 161
Nickname: bozomind
Registered: Mar, 2004

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 2, 2004 2:53 PM
Reply to this message Reply
I apologize for not giving the appropriate attribution to Gregg Wonderly for the wholesale copying of the summary text.

Thanks for the quote Gregg.

Jim Hurley

Posts: 9
Nickname: jhurley
Registered: Oct, 2002

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 3, 2004 12:41 PM
Reply to this message Reply
Just a few clarifications...

* "Community Access to the Jini Starter Kit"
I know you're talking about a visible CVS tree... but the title of
this may imply to some that the Community doesn't have access
to the Jini Starter Kit. Of course - it does.

* "At the 7th Annual Jini Community Meeting, "
The Meeting is not an Annual meeting. It's whenever we can
pull it together! (Sometimes more than once a year, sometimes
it takes longer)

* "I feel that having the entire source tree available would be a good thing for the community."
Again, I know you're talking about a visible CVS tree... but to
clarify for people who may get the wrong impression - the entire
source is available to the entire Community. It's always been
that way. When you download the Starter Kit - you get Sun's
contributed implementations with all of the source code.

-Jim

Dale Asberry

Posts: 161
Nickname: bozomind
Registered: Mar, 2004

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 3, 2004 2:18 PM
Reply to this message Reply
Sorry for the confusion.

Yes, the source code IS available to all who join the community.

I should've quoted the "At the 7th Annual..." in the post so that I could blame this one one Gregg Wonderly :-)

Yes, you are correct about availability of the JSTK released code, I mean a publicly visible source tree of the code currently in development and not yet released.

Frank Sommers

Posts: 2642
Nickname: fsommers
Registered: Jan, 2002

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 4, 2004 5:11 PM
Reply to this message Reply
Well, I always thought that the JSK from Sun is just one implementation of the Jini specs. As it is a Sun project, I am not sure how others could contribute to the implementation itself. Since the Sun code is very high quality, production-ready code, I don't see how the community would benefit from having access to the development source tree. The analogy would be the Java VM itself: While the source code for Sun's implementation of the Java VM is available (under the SCSL license, AFAIK), it's a Sun product, and its development source is not available outside of Sun.

Now, what we don't have in Jini are fully open implementations of the entire Jini specs. With the JVM, there are fully open-source VMs. However, Rob Gingell once mentioned to me in an interview (available here on Artima), that VM's that never passed Sun's compatibility tests actually pose an interesting problem: While they claim to be Java VMs, they are really not "officially" sanctioned VMs. Other than their users, no one really knows if they function exactly to the specs. With JVMs, you apparently had to license the test harnesses from Sun, which did cost money, and which many open source projects did not want to pay for. Note that in 2002 Sun had remedied that situation by eliminating the fee associated with those test kits.

I wonder whether the current Jini licensing mechanism would allow a fully open implementation of the Jini specs. I used to think that the answer was an unequivocal no, but after doing a bit of research on licensing and the JCP process, I'm not sure any more. I think the answer really depends on the copyright and license on the Jini specs themselves, and then the copyright on the Jini name itself. While I'm not sure about an open-source implementation, I know that Jini implementations have already been created inside the SCSL licensing scheme.

I do think, though, that being able to create a fully open-source implementation of the Jini specs would benefit the broader Jini and Java community, since implementations would soon find their way into numerous open source projects and distributions. That, IMO, can only further the Jini cause, and contribute to Jini's ubiquity.

Additionaly, many academic research projects prefer open-source licenses, since the students working on those projects often desire the broadest possible dissemination of their work. That's in keeping with the spirit of academic cooperation (which actually contributed to the collaborative spirit of the open source movement itself, according to Eric Raymond). Open source licensing also simplifies distributing code, since many university departments are not set up to police various software license distribution mechanisms (especially if the college can't create revenue from those projects). If you use Jini in a research project, you can't just post a tar ball of your code on a Web site. That's a problem. I personally encountered that problem with a project I worked on in the past, and that project's source code is irretrievably buried on some university computers.

At this point, I'm inclined to support an open source Jini initiative. In the past, I believed that the SCSL license was important to protect Jini's integrity. I think an objective way to test that hypothesis is to review the experience with the SCSL and Jini from the past 5 years, and see actual cases where the SCSL prevented abuses. In other words, I'd like to see cases where Sun actually had to go after Jini impostors. If that never happened, then the SCSL might be a sheild against a strawman, and it might truly restrict Jini's wide and ubiquitious dissemination. Of course, we wouldn't really know if that's the case, until an open source Jini implementation becomes an option.

Dale Asberry

Posts: 161
Nickname: bozomind
Registered: Mar, 2004

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 5, 2004 10:22 AM
Reply to this message Reply
Before I start, I want to point out that your post brings out two issues: access to the CVS source tree AND SCSL licensing problems. I'm wanting to address the SCSL in a future blog entry so I'll bring up those points then.
 Well, I always thought that the JSK from Sun is just one implementation of the Jini specs. As it is a Sun 
project, I am not sure how others could contribute to the implementation itself. Since the Sun code is very high
quality, production-ready code, I don't see how the community would benefit from having access to the
development source tree. The analogy would be the Java VM itself: While the source code for Sun's implementation of
the Java VM is available (under the SCSL license, AFAIK), it's a Sun product, and its development source is not
available outside of Sun.
The difficulty with this point is that, unlike the JVM and JDK, the Jini team is seriously resource constrained. In addition, the JSTK has been contributed to the Community while the JDK has not been.
I do think, though, that being able to create a fully open-source implementation of the Jini specs would benefit 
the broader Jini and Java community, since implementations would soon find their way into numerous open source projects
and distributions. That, IMO, can only further the Jini cause, and contribute to Jini's ubiquity.
I completely agree, however, the opportunity cost of doing so is very high. Organizing such an effort with people that have the time resources to commit is so high that I doubt it could ever come to pass. By putting the CVS tree in a public forum, the opportunity cost for contributing is lowered significantly. Those of us that are excited about the technology can easily contribute and Sun can still maintain control. Also, would anyone in the community truly be able to submit a real open source implementation? We'd likely have to exclude the people who have implemented partial solutions like Geoff Arnold, Dan Creswell, and Mark Brouwer (and cohorts) since they've seen/used the Sun contributed code.

Kim

Posts: 4
Nickname: kmarius
Registered: Oct, 2003

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 5, 2004 3:59 PM
Reply to this message Reply
What we really need is a more flexible "new".

Example:

Thing aThing = new ();

or with arguments:
Car c = new ("Honda","blue");

The compiler should fill in the necessary information

Gregg Wonderly

Posts: 317
Nickname: greggwon
Registered: Apr, 2003

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 5, 2004 4:18 PM
Reply to this message Reply
>At the 7th Annual Jini Community Meeting, one of the discussions was
>about the fact that Sun did not have a version of the Jini Starter Kit
>in a publicly visible CVS tree.
>
>I feel that having the entire source tree available would be a good thing
>for the community. For those concerned about forks or other nonsense,
>then the list of committers can remain solely in the power of the Sun
>Jini team.

One of the arguments I recall hearing is that the current distribution (which does include the source) only has Sun contributions in it, and that makes it appear that the community is not in control of Jini. There is a important distinction here that I want to make sure we all discuss. The Spec is changed by the JDP which involves the community. The starter kit is Sun's implementation of what the Spec says. Anytime the community changes the spec, Sun's JTSK has to comply right?

>With the time between release dates for most projects exceeding a year
>(or more), source access allows each community member to decide if
>possible incompatibilities are worth the benefits.

The important issue for me is whether the community needs the JTSK to do something different that the spec controls, or is there other functionality that is not part of the spec (like the start.jar stuff), but that the community has found useful that needs enhancements, but which currently does not have a community contribution path.

>The Jini team has limited resources and I believe could greatly benefit
>from the OSS process of many people chipping away small pieces of the
>problem - bugs or new features. We (the Jini Community) all know the
>power of worker processes grabbing tasks out of a common space: by
>publishing issues and enhancements and accepting patches and tests in
>return, the community would benefit from faster JSTK releases.

In my mind there are, at least, a couple of different issues to discuss.

One of the issue is how does the community contribute to things that are part of the starter kit so that ease of use and other issues can be dealt with quickly. That question is based on the assumption that a single JTSK like thing is a good idea. I think it is a good idea right now. I don't see the value in another distribution, asside from the obvious performance competition possible in the Javaspaces implementations.

Another issue is related to whether the community feels that Sun's development tree for the JTSK should really be visible to the community, or is it okay for it to be a 'product' that Sun is in control of.

Currently, we don't have a competing JTSK like thing. But, Seven's release will change that. It will be a spec compliant platform that includes extensions for container based service deployment. When that happens, will Seven augment Sun's JTSK, or will it include a particular version? I think I understand from Mark Brouwer that he wants to avoid creating another JTSK distribution. It seems to me that he want's Seven to piggy back on whatever it is that Sun is distributing.

So, the real question seems to be, where is the best place for the community to add value to the Jini platform without fragmenting the platform view that the JTSK creates by standing alone?

Kim

Posts: 4
Nickname: kmarius
Registered: Oct, 2003

Re: Community Access to the Jini Starter Kit Codebase Posted: Apr 6, 2004 9:24 AM
Reply to this message Reply
Sorry, don't know how my post got to this forum, this has nothing to do with Jini Start Kit :-(

Flat View: This topic has 8 replies on 1 page
Topic: Designing Graphic Interfaces Previous Topic   Next Topic Topic: Experiences On Running A Website For Profit, Part 1: The Beginning


Sponsored Links



Google
  Web Artima.com   

Copyright © 1996-2014 Artima, Inc. All Rights Reserved. - Privacy Policy - Terms of Use - Advertise with Us