You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jcp-open@apache.org by Jukka Zitting <ju...@gmail.com> on 2006/01/15 12:32:22 UTC
Apache project as a JSR reference implementation
Hi,
I'm a committer of the incubating Apache Jackrabbit project that was
used as the reference implementation of the JSR 170 (Content
Repository for the Java Technology API). A second version of the API
is currently being specified by JSR 283 and there is interest to keep
using Jackrabbit as the reference implementation also for JSR 283.
Currently most of the active Jackrabbit committers (me included) are
members of the JSR 283 expert group, so I there shouldn't be many
problems doing this in practice, but I'm a bit concerned about how
this will affect the project as the community evolves. How should we
communicate design decisions based on API details not yet published by
the expert group?
I suppose this problem of interfacing the open source model with the
JCP model has come up in other JCP-related Apache projects as well, so
I hope to learn from existing ways to handle such a situation. I'm
sorry if this is a FAQ, I couldn't find anything related with a quick
search.
PS. Once Jackrabbit is promoted from the Incubator, I hope to have it
included in the implementation list on http://www.apache.org/jcp/.
BR,
Jukka Zitting
--
Yukatan - http://yukatan.fi/ - info@yukatan.fi
Software craftmanship, JCR consulting, and Java development
Re: NDA requirements
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Brett Porter wrote:
> Geir Magnusson Jr wrote:
>> Yes, but as I don't follow the discussion, I didn't know if it would be
>> possible to shadow the topics w/o giving away confidential information.
>> This is the problem - what is confidential....
>>
>> I wonder if you could get assent from the EG to talk about specific
>> concepts - like tell them what you are doing "I have an interest list
>> that isn't under NDA... can I discuss $foo to see what they think?"
>>
>
> I think the problem is I have no idea what $foo is at this point.
You are on the EG :) You should know what $foo is.
It's a
> chicken and egg problem - we have people interested, but until there is
> understanding of where its at we don't know what is interesting.
>
Yes, that's true.
> Let's start where we are at now, and see where it takes us.
>
>> I dunno. I'm trying... :)
>
> Thanks :)
>
> - Brett
>
>
Re: NDA requirements
Posted by Brett Porter <br...@apache.org>.
Geir Magnusson Jr wrote:
> Yes, but as I don't follow the discussion, I didn't know if it would be
> possible to shadow the topics w/o giving away confidential information.
> This is the problem - what is confidential....
>
> I wonder if you could get assent from the EG to talk about specific
> concepts - like tell them what you are doing "I have an interest list
> that isn't under NDA... can I discuss $foo to see what they think?"
>
I think the problem is I have no idea what $foo is at this point. It's a
chicken and egg problem - we have people interested, but until there is
understanding of where its at we don't know what is interesting.
Let's start where we are at now, and see where it takes us.
> I dunno. I'm trying... :)
Thanks :)
- Brett
Re: NDA requirements
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Brett Porter wrote:
> Geir Magnusson Jr wrote:
>> Brett et al : do you think it's possible to run this as a completely
>> open list, that the onus is on you to filter and scrub information that
>> you move across the boundary? IOW, you don't ever forward mail, you
>> don't move code across the boundary... I don't know what kinds of
>> things are being discussed in the EG, but can concepts be discussed "in
>> general" on the 277 open list?
>
> I don't think this would be possible until at least the public draft is
> available. Discussing Java modularity out of context of the JSR is a
> rather large topic. My understanding of the list was to allow discussing
> the specifics of the JSR, not general discussion which there are
> probably other forums we could use.
Yes, but as I don't follow the discussion, I didn't know if it would be
possible to shadow the topics w/o giving away confidential information.
This is the problem - what is confidential....
I wonder if you could get assent from the EG to talk about specific
concepts - like tell them what you are doing "I have an interest list
that isn't under NDA... can I discuss $foo to see what they think?"
I dunno. I'm trying... :)
geir
>
> What did others think/expect?
>
> - Brett
>
>
Re: NDA requirements
Posted by Steve Loughran <st...@apache.org>.
Brett Porter wrote:
> Geir Magnusson Jr wrote:
>> Brett et al : do you think it's possible to run this as a completely
>> open list, that the onus is on you to filter and scrub information that
>> you move across the boundary? IOW, you don't ever forward mail, you
>> don't move code across the boundary... I don't know what kinds of
>> things are being discussed in the EG, but can concepts be discussed "in
>> general" on the 277 open list?
>
> I don't think this would be possible until at least the public draft is
> available. Discussing Java modularity out of context of the JSR is a
> rather large topic. My understanding of the list was to allow discussing
> the specifics of the JSR, not general discussion which there are
> probably other forums we could use.
>
> What did others think/expect?
I think its best to keep its closed. That way nobody on the mail list
needs to be circumspect.
One thing I would like you to strive for is making that test suite
public once the spec goes up, so that when apache goes to do the impl,
we dont have a secret test suite that you can't discuss in -dev lists,
post snippets of on bugreps, etc.
Re: NDA requirements
Posted by Brett Porter <br...@apache.org>.
Geir Magnusson Jr wrote:
> Brett et al : do you think it's possible to run this as a completely
> open list, that the onus is on you to filter and scrub information that
> you move across the boundary? IOW, you don't ever forward mail, you
> don't move code across the boundary... I don't know what kinds of
> things are being discussed in the EG, but can concepts be discussed "in
> general" on the 277 open list?
I don't think this would be possible until at least the public draft is
available. Discussing Java modularity out of context of the JSR is a
rather large topic. My understanding of the list was to allow discussing
the specifics of the JSR, not general discussion which there are
probably other forums we could use.
What did others think/expect?
- Brett
Re: NDA requirements
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Steve Loughran wrote:
> Geir Magnusson Jr wrote:
>> What's conflicting? We're trying to avoid messes like these....
>
> well, its just the problem of having do another NDA. Theres no conflict
> of interest yet, but if I get to find out something on the mail list
> that I shouldnt disclose to HP, then I have a problem, and I dont want
> to go there. Its a bit like how I make a point of never looking at the
> source to unix, winnt, vms or java, even though they are on filestores
> somewhere, because it makes it harder to write OSS code.
Right. The issue here seems to be the confidentiality requirement of the
material, rather than the NDA itself. If it was a verbal agreement,
you'd be bound by same conditions, and have the same problems. This is
why the JCP thing is so awful, and why we push to open it where and when
we can. The whole discussion of whether or not a member signs an NDA,
or agrees to same conditions, doesn't change anything.
Brett et al : do you think it's possible to run this as a completely
open list, that the onus is on you to filter and scrub information that
you move across the boundary? IOW, you don't ever forward mail, you
don't move code across the boundary... I don't know what kinds of
things are being discussed in the EG, but can concepts be discussed "in
general" on the 277 open list?
geir
Re: NDA requirements
Posted by Steve Loughran <st...@apache.org>.
Geir Magnusson Jr wrote:
> What's conflicting? We're trying to avoid messes like these....
well, its just the problem of having do another NDA. Theres no conflict
of interest yet, but if I get to find out something on the mail list
that I shouldnt disclose to HP, then I have a problem, and I dont want
to go there. Its a bit like how I make a point of never looking at the
source to unix, winnt, vms or java, even though they are on filestores
somewhere, because it makes it harder to write OSS code.
-steve
Re: NDA requirements
Posted by Geir Magnusson Jr <ge...@pobox.com>.
What's conflicting? We're trying to avoid messes like these....
Steve Loughran wrote:
> Geir Magnusson Jr wrote:
>>
>>
>> Brett Porter wrote:
>
>>>
>>> a) are non-members whose employer has signed the JSPA required to sign
>>> an individual NDA? (specifically, Steve has raised this with regards to
>>> HP - note that HP is not on the 277 EG itself though)
>>
>> Yes. I'm sure that someone better versed in law will put this better,
>> but any arrangements between Sun, HP and Steve (in whatever order)
>> have nothing to do with arrangements between Sun, ASF and Steve. It's
>> not like a drivers license or something, for which states grant
>> reciprocal recognition. It's not done for gun permits, for some
>> reason... :)
>>
>>>
>>> b) if members of a) are required to sign an NDA, are Sun employees any
>>> exception to this?
>>
>> That's a great question, and I still think the answer is yes, because
>> we have no visibility or reliance on arrangements between Sun and its
>> employees.
>>
>> TO me, the cleanest approach is to ask everyone to sign the NDA. It
>> was written deliberately to be as simple and loose as possible because
>> we're just trying to demonstrate responsible oversight, rather than
>> try and hose people with stupid process. If Steve has a problem, and
>> has a suggestion to make it even simpler or looser, I'm certainly all
>> ears...
>
> It's probably best I steer clear of conflicting NDAs, because that only
> complicates my life, as does talking to lawyers.
>
> If that means staying off the 277 list until I do sit down and talk to
> the legal dept, then so be it.
>
> -steve
>
>
Re: NDA requirements
Posted by Steve Loughran <st...@apache.org>.
Geir Magnusson Jr wrote:
>
>
> Brett Porter wrote:
>>
>> a) are non-members whose employer has signed the JSPA required to sign
>> an individual NDA? (specifically, Steve has raised this with regards to
>> HP - note that HP is not on the 277 EG itself though)
>
> Yes. I'm sure that someone better versed in law will put this better,
> but any arrangements between Sun, HP and Steve (in whatever order) have
> nothing to do with arrangements between Sun, ASF and Steve. It's not
> like a drivers license or something, for which states grant reciprocal
> recognition. It's not done for gun permits, for some reason... :)
>
>>
>> b) if members of a) are required to sign an NDA, are Sun employees any
>> exception to this?
>
> That's a great question, and I still think the answer is yes, because we
> have no visibility or reliance on arrangements between Sun and its
> employees.
>
> TO me, the cleanest approach is to ask everyone to sign the NDA. It was
> written deliberately to be as simple and loose as possible because we're
> just trying to demonstrate responsible oversight, rather than try and
> hose people with stupid process. If Steve has a problem, and has a
> suggestion to make it even simpler or looser, I'm certainly all ears...
It's probably best I steer clear of conflicting NDAs, because that only
complicates my life, as does talking to lawyers.
If that means staying off the 277 list until I do sit down and talk to
the legal dept, then so be it.
-steve
Re: NDA requirements
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Brett Porter wrote:
> Geir Magnusson Jr wrote:
>>
>> Brett Porter wrote:
>>> Sorry for the late reply. Not been at the computer most of the last week.
>> I know the feeling. Been sick myself and then a touch of travel.
>>
>> List is done. You have been added as a moderator. If you don't mind,
>> please continue to ask that all get NDAs signed, but don't let that stop
>> you from adding a Member - out-of-order is fine for members on the
>> theory that there's some [weird] implicit NDA in place...
>>
>
> Thanks Geir. Can I get clarification on one point about the NDA, since
> you are the one with the legal responsibility:
Ultimately it's the ASF, but given we have no D&O insurance, I suppose
that in some twisted and alternative universe that I hope never to live
in, my house is on the line... I keep thinking that it would be prudent
to put in my wife's name, so when I go off to JCP Process Prison, her
and the kids aren't tossed out into the mean streets :) I actually
don't think there is real risk here, but of our keister is well-covered
by a simple piece of paper...
>
> a) are non-members whose employer has signed the JSPA required to sign
> an individual NDA? (specifically, Steve has raised this with regards to
> HP - note that HP is not on the 277 EG itself though)
Yes. I'm sure that someone better versed in law will put this better,
but any arrangements between Sun, HP and Steve (in whatever order) have
nothing to do with arrangements between Sun, ASF and Steve. It's not
like a drivers license or something, for which states grant reciprocal
recognition. It's not done for gun permits, for some reason... :)
>
> b) if members of a) are required to sign an NDA, are Sun employees any
> exception to this?
That's a great question, and I still think the answer is yes, because we
have no visibility or reliance on arrangements between Sun and its
employees.
TO me, the cleanest approach is to ask everyone to sign the NDA. It was
written deliberately to be as simple and loose as possible because we're
just trying to demonstrate responsible oversight, rather than try and
hose people with stupid process. If Steve has a problem, and has a
suggestion to make it even simpler or looser, I'm certainly all ears...
geir
>
> Thanks,
> Brett
>
>
Re: NDA requirements (was: list stuff)
Posted by Brett Porter <br...@apache.org>.
Geir Magnusson Jr wrote:
>
>
> Brett Porter wrote:
>> Sorry for the late reply. Not been at the computer most of the last week.
>
> I know the feeling. Been sick myself and then a touch of travel.
>
> List is done. You have been added as a moderator. If you don't mind,
> please continue to ask that all get NDAs signed, but don't let that stop
> you from adding a Member - out-of-order is fine for members on the
> theory that there's some [weird] implicit NDA in place...
>
Thanks Geir. Can I get clarification on one point about the NDA, since
you are the one with the legal responsibility:
a) are non-members whose employer has signed the JSPA required to sign
an individual NDA? (specifically, Steve has raised this with regards to
HP - note that HP is not on the 277 EG itself though)
b) if members of a) are required to sign an NDA, are Sun employees any
exception to this?
Thanks,
Brett
Re: list stuff
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Brett Porter wrote:
> Sorry for the late reply. Not been at the computer most of the last week.
I know the feeling. Been sick myself and then a touch of travel.
List is done. You have been added as a moderator. If you don't mind,
please continue to ask that all get NDAs signed, but don't let that stop
you from adding a Member - out-of-order is fine for members on the
theory that there's some [weird] implicit NDA in place...
geir
>
> Geir Magnusson Jr wrote:
>>
>> Brett Porter wrote:
>>> Obviously +1 on this. Let us know.
>>>
>>> Does mail stuff include the 277 list by any chance?
>> Yes.
>
> Have you had a chance to look at this?
>
> Are there other people doing ML admin than yourself and Roy? I'd have to
> learn the ropes, but my offer still stands to help out if needed. I
> think I learn quickly :)
>
>> I haven't received a
>>> response to my most recent email. I've had some more feedback from
>>> Stanley and need to catch up to where we are at here.
>> re the list? What did he say?
>
> I'll send a separate mail on this.
>
>> (I'll go respond to the last one now...)
>
> Did you send this? I didn't get anything.
>
> Thanks,
> Brett
>
>
Re: list stuff
Posted by Brett Porter <br...@apache.org>.
Sorry for the late reply. Not been at the computer most of the last week.
Geir Magnusson Jr wrote:
>
>
> Brett Porter wrote:
>> Obviously +1 on this. Let us know.
>>
>> Does mail stuff include the 277 list by any chance?
>
> Yes.
Have you had a chance to look at this?
Are there other people doing ML admin than yourself and Roy? I'd have to
learn the ropes, but my offer still stands to help out if needed. I
think I learn quickly :)
>
> I haven't received a
>> response to my most recent email. I've had some more feedback from
>> Stanley and need to catch up to where we are at here.
>
> re the list? What did he say?
I'll send a separate mail on this.
> (I'll go respond to the last one now...)
Did you send this? I didn't get anything.
Thanks,
Brett
Re: list stuff
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Brett Porter wrote:
> Obviously +1 on this. Let us know.
>
> Does mail stuff include the 277 list by any chance?
Yes.
I haven't received a
> response to my most recent email. I've had some more feedback from
> Stanley and need to catch up to where we are at here.
re the list? What did he say? (I'll go respond to the last one now...)
geir
>
> Let me know if I can help.
>
> - Brett
>
> Geir Magnusson Jr wrote:
>> Works for me. I have some other mail stuff to do, so I'll do this later
>> today... we'll keep this one up for a little while to drive people to
>> -discuss
>>
>> geir
>>
>> Roy T. Fielding wrote:
>>> Umm, the config for this list was, well, highly creative. I just
>>> fixed that to be the usual moderated list with open subscription.
>>> However, the archive is private (and yet full of spam).
>>>
>>> I suggest that we change this list to jcp-discuss with a public
>>> archive and delete the spam. We already have the jcp at apache
>>> list for private discussion.
>>>
>>> ....Roy
>>>
>>>
>
>
Re: list stuff
Posted by Brett Porter <br...@apache.org>.
Obviously +1 on this. Let us know.
Does mail stuff include the 277 list by any chance? I haven't received a
response to my most recent email. I've had some more feedback from
Stanley and need to catch up to where we are at here.
Let me know if I can help.
- Brett
Geir Magnusson Jr wrote:
> Works for me. I have some other mail stuff to do, so I'll do this later
> today... we'll keep this one up for a little while to drive people to
> -discuss
>
> geir
>
> Roy T. Fielding wrote:
>> Umm, the config for this list was, well, highly creative. I just
>> fixed that to be the usual moderated list with open subscription.
>> However, the archive is private (and yet full of spam).
>>
>> I suggest that we change this list to jcp-discuss with a public
>> archive and delete the spam. We already have the jcp at apache
>> list for private discussion.
>>
>> ....Roy
>>
>>
>
Re: list stuff
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Works for me. I have some other mail stuff to do, so I'll do this later
today... we'll keep this one up for a little while to drive people to
-discuss
geir
Roy T. Fielding wrote:
> Umm, the config for this list was, well, highly creative. I just
> fixed that to be the usual moderated list with open subscription.
> However, the archive is private (and yet full of spam).
>
> I suggest that we change this list to jcp-discuss with a public
> archive and delete the spam. We already have the jcp at apache
> list for private discussion.
>
> ....Roy
>
>
Re: list stuff
Posted by Geir Magnusson Jr <ge...@pobox.com>.
seems right.
Brett Porter wrote:
> Trying to put all this in JIRA (INFRA-705). Is this the correct list of
> items?
>
> * create a publicly archived, moderated subscription jcp-discuss list
> * subscribe those from jcp-open, redirect posts to jcp-open
> * close down jcp-open as a list
> * cleanse jcp-open archives of spam and leave in private-arch for
> historical purposes
>
> I wasn't sure about "moderated subscription" (though that is consistent
> with similar lists like community, repository and legal-discuss), and
> whether the archives remain where they are or get copied to the new
> location.
>
> I've volunteered to make time to get this done.
>
> - Brett
>
>
> Roy T. Fielding wrote:
>> Umm, the config for this list was, well, highly creative. I just
>> fixed that to be the usual moderated list with open subscription.
>> However, the archive is private (and yet full of spam).
>>
>> I suggest that we change this list to jcp-discuss with a public
>> archive and delete the spam. We already have the jcp at apache
>> list for private discussion.
>>
>> ....Roy
>>
>
>
Re: list stuff
Posted by Brett Porter <br...@apache.org>.
Trying to put all this in JIRA (INFRA-705). Is this the correct list of
items?
* create a publicly archived, moderated subscription jcp-discuss list
* subscribe those from jcp-open, redirect posts to jcp-open
* close down jcp-open as a list
* cleanse jcp-open archives of spam and leave in private-arch for
historical purposes
I wasn't sure about "moderated subscription" (though that is consistent
with similar lists like community, repository and legal-discuss), and
whether the archives remain where they are or get copied to the new
location.
I've volunteered to make time to get this done.
- Brett
Roy T. Fielding wrote:
> Umm, the config for this list was, well, highly creative. I just
> fixed that to be the usual moderated list with open subscription.
> However, the archive is private (and yet full of spam).
>
> I suggest that we change this list to jcp-discuss with a public
> archive and delete the spam. We already have the jcp at apache
> list for private discussion.
>
> ....Roy
>
list stuff
Posted by "Roy T. Fielding" <fi...@gbiv.com>.
Umm, the config for this list was, well, highly creative. I just
fixed that to be the usual moderated list with open subscription.
However, the archive is private (and yet full of spam).
I suggest that we change this list to jcp-discuss with a public
archive and delete the spam. We already have the jcp at apache
list for private discussion.
....Roy
Re: Apache project as a JSR reference implementation
Posted by "Roy T. Fielding" <fi...@gbiv.com>.
On Jan 15, 2006, at 10:38 AM, Geir Magnusson Jr wrote:
>>> You would be breaking fresh ground at the ASF if you could
>>> convince the
>>> EG to work this way. I will happily help you in any way that you
>>> could
>>> use me.
>> Cool, thanks! I'll check back with the EG and we'll see what comes
>> out
>> of this. :-)
>
> Right. Day Software is the spec lead. David Nuescheler clearly
> "gets it" and further, our own Roy Fielding (yes, Roy, we appear to
> own you :) is their Chief Scientist, so you have quite a bit of
> help on your side here. let us know.
I would, of course, be overjoyed to make it completely public since that
would remove all of the liability (both here and at Day) of playing with
various implementations before the spec is declared finished. I already
tried to convince David to make it public when 283 started. My guess
(and its only a guess) is that the IBM experts don't want to work in
public,
since they have said that multiple times, and David is trying to
accommodate
their wishes rather than cause a conflict.
BTW, we should get David on this list.
....Roy
Re: Apache project as a JSR reference implementation
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Jukka Zitting wrote:
> Hi,
>
> On 1/15/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
>> Well, if you guys can control the EG, I'd propose that you work entirely
>> in public, just like an Apache (or other community-oriented open source
>> project does).
>
> That would be very nice. IANAL so I'm not exactly sure of what are the
> legal constraints for doing that. The JSPA (2.0) section 11
> (Publicity) puts pretty heavy constraints on discussing any JSR work
> publicly, but the JCP 2: Process Document (2.6) section 2.1.1 (Freedom
> of working style) allows and even encourages transparency. Are there
> other openness constraints for JSR work? If not, then I suppose the
> expert group could very well decide to be as open as needed.
That's right. The JSPA has a constrictive structure when working
privately, all allegedly to ensure that no one gets an early advantage
in the market (except for the EG members, I suppose) and to enable
contributors not used to public discussion to be able to float ideas and
IP in private. Or something.
However, if you all work on a public list that is considered to be under
the terms of the Apache License and the appropriate contributors
agreements (like all Apache projects do implicitly, and some - like
Harmony - do explicitly (see the top of
http://incubator.apache.org/harmony/mailing.html), then the IP rules
don't apply. Now, there may be EG members who are reluctant to work
this way, so there is where lobbying and constructive engagement from us
will help.
>
>> You would be breaking fresh ground at the ASF if you could convince the
>> EG to work this way. I will happily help you in any way that you could
>> use me.
>
> Cool, thanks! I'll check back with the EG and we'll see what comes out
> of this. :-)
Right. Day Software is the spec lead. David Nuescheler clearly "gets
it" and further, our own Roy Fielding (yes, Roy, we appear to own you :)
is their Chief Scientist, so you have quite a bit of help on your side
here. let us know.
geir
Re: Apache project as a JSR reference implementation
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On 1/15/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
> Well, if you guys can control the EG, I'd propose that you work entirely
> in public, just like an Apache (or other community-oriented open source
> project does).
That would be very nice. IANAL so I'm not exactly sure of what are the
legal constraints for doing that. The JSPA (2.0) section 11
(Publicity) puts pretty heavy constraints on discussing any JSR work
publicly, but the JCP 2: Process Document (2.6) section 2.1.1 (Freedom
of working style) allows and even encourages transparency. Are there
other openness constraints for JSR work? If not, then I suppose the
expert group could very well decide to be as open as needed.
> You would be breaking fresh ground at the ASF if you could convince the
> EG to work this way. I will happily help you in any way that you could
> use me.
Cool, thanks! I'll check back with the EG and we'll see what comes out
of this. :-)
BR,
Jukka Zitting
--
Yukatan - http://yukatan.fi/ - info@yukatan.fi
Software craftmanship, JCR consulting, and Java development
Re: Apache project as a JSR reference implementation
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
Reporting back on the JSR 283 / Apache Jackrabbit issue. The problem I
raised in January was how to best combine the open source development
process with the rather strict default JCP publicity rules.
On 1/16/06, Jukka Zitting <ju...@gmail.com> wrote:
> On 1/16/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
> > Let us know how things go. It sounds like you're going to have some
> > trouble with some of the members of the EG.
>
> OK, I'll keep you posted.
We reached a compromise in a face-to-face meeting last month. Instead
of directly opening the EG work to the public, we will periodically
send a public newsletter with details of the API changes that we've
reached a rough consensus on. This will allow us to work in public on
the reference implementation code in Apache Jackrabbit even before and
between the official specification draft releases.
I've volunteered to compile these newsletters and plan to send the
first one during this month. Time will tell how well this idea works
in practice and how much extra labour it requires.
BR,
Jukka Zitting
--
Yukatan - http://yukatan.fi/ - info@yukatan.fi
Software craftsmanship, JCR consulting, and Java development
Re: Apache project as a JSR reference implementation
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On 1/16/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
> Let us know how things go. It sounds like you're going to have some
> trouble with some of the members of the EG.
OK, I'll keep you posted.
BR,
Jukka Zitting
--
Yukatan - http://yukatan.fi/ - info@yukatan.fi
Software craftmanship, JCR consulting, and Java development
Re: Apache project as a JSR reference implementation
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Steve Loughran wrote:
> Geir Magnusson Jr wrote:
>>
>>
>> Jukka Zitting wrote:
>>> Hi,
>>>
>>> I'm a committer of the incubating Apache Jackrabbit project that was
>>> used as the reference implementation of the JSR 170 (Content
>>> Repository for the Java Technology API). A second version of the API
>>> is currently being specified by JSR 283 and there is interest to keep
>>> using Jackrabbit as the reference implementation also for JSR 283.
>>>
>>
>> Sounds good.
>>
>>> Currently most of the active Jackrabbit committers (me included) are
>>> members of the JSR 283 expert group, so I there shouldn't be many
>>> problems doing this in practice, but I'm a bit concerned about how
>>> this will affect the project as the community evolves. How should we
>>> communicate design decisions based on API details not yet published by
>>> the expert group?
>>
>> Well, if you guys can control the EG, I'd propose that you work
>> entirely in public, just like an Apache (or other community-oriented
>> open source project does).
>>
>> For example, you might offer to have the Jackrabbit project 'host' the
>> EG's work, add a separate, public mailing list for the expert group
>> traffic, and a part of your SVN for their work. I suggest a separate
>> list only to keep clear what is normal jackrabbit developer traffic,
>> and what is the EG traffic.
>>
>
> Public is good, under gump is good too.
>
> I would go for
>
> 1-a separate SCM repository for spec: api+test cases that is decoupled
> from any implementation, so that all of the spec team can have commit
> access to the spec repository.
>
> 2-a junit based testrunner that can run test suites against any
> implementation that implements the right factory interface.
>
> 3-test-centric API/spec
>
> I've got the first two up and running on sourceforge for our deployment
> project, http://sourceforge.net/projects/deployment
>
> if you grab everything from CVS you can find the abuse of Junit that
> lets me take a manifest of XML files and turn each document into a test
> case. We have two open source API impls, and one closed source one all
> sharing the same test runner, so it will be easy to compare the results.
>
> The big barriers to (3) are ignorance of the value of testing amongst
> some of the spec team (its hard to work with a team leader who believes
> we should focus on implementation ahead of testing), and not enough
> experience among team members about how to write good tests.
> Interestingly, it is the OSS projects that are better at this than
> anyone else.
Indeed.
Let us know how things go. It sounds like you're going to have some
trouble with some of the members of the EG. I sent a note to David
inviting him to join us - I suppose we wait to see what he says.
geir
>
> -steve
>
>
Re: Apache project as a JSR reference implementation
Posted by Steve Loughran <st...@apache.org>.
Geir Magnusson Jr wrote:
>
>
> Jukka Zitting wrote:
>> Hi,
>>
>> I'm a committer of the incubating Apache Jackrabbit project that was
>> used as the reference implementation of the JSR 170 (Content
>> Repository for the Java Technology API). A second version of the API
>> is currently being specified by JSR 283 and there is interest to keep
>> using Jackrabbit as the reference implementation also for JSR 283.
>>
>
> Sounds good.
>
>> Currently most of the active Jackrabbit committers (me included) are
>> members of the JSR 283 expert group, so I there shouldn't be many
>> problems doing this in practice, but I'm a bit concerned about how
>> this will affect the project as the community evolves. How should we
>> communicate design decisions based on API details not yet published by
>> the expert group?
>
> Well, if you guys can control the EG, I'd propose that you work entirely
> in public, just like an Apache (or other community-oriented open source
> project does).
>
> For example, you might offer to have the Jackrabbit project 'host' the
> EG's work, add a separate, public mailing list for the expert group
> traffic, and a part of your SVN for their work. I suggest a separate
> list only to keep clear what is normal jackrabbit developer traffic, and
> what is the EG traffic.
>
Public is good, under gump is good too.
I would go for
1-a separate SCM repository for spec: api+test cases that is decoupled
from any implementation, so that all of the spec team can have commit
access to the spec repository.
2-a junit based testrunner that can run test suites against any
implementation that implements the right factory interface.
3-test-centric API/spec
I've got the first two up and running on sourceforge for our deployment
project, http://sourceforge.net/projects/deployment
if you grab everything from CVS you can find the abuse of Junit that
lets me take a manifest of XML files and turn each document into a test
case. We have two open source API impls, and one closed source one all
sharing the same test runner, so it will be easy to compare the results.
The big barriers to (3) are ignorance of the value of testing amongst
some of the spec team (its hard to work with a team leader who believes
we should focus on implementation ahead of testing), and not enough
experience among team members about how to write good tests.
Interestingly, it is the OSS projects that are better at this than
anyone else.
-steve
Re: Apache project as a JSR reference implementation
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Jukka Zitting wrote:
> Hi,
>
> I'm a committer of the incubating Apache Jackrabbit project that was
> used as the reference implementation of the JSR 170 (Content
> Repository for the Java Technology API). A second version of the API
> is currently being specified by JSR 283 and there is interest to keep
> using Jackrabbit as the reference implementation also for JSR 283.
>
Sounds good.
> Currently most of the active Jackrabbit committers (me included) are
> members of the JSR 283 expert group, so I there shouldn't be many
> problems doing this in practice, but I'm a bit concerned about how
> this will affect the project as the community evolves. How should we
> communicate design decisions based on API details not yet published by
> the expert group?
Well, if you guys can control the EG, I'd propose that you work entirely
in public, just like an Apache (or other community-oriented open source
project does).
For example, you might offer to have the Jackrabbit project 'host' the
EG's work, add a separate, public mailing list for the expert group
traffic, and a part of your SVN for their work. I suggest a separate
list only to keep clear what is normal jackrabbit developer traffic, and
what is the EG traffic.
You might wish to talk with Craig Russel (craig.russel@sun.com) about
how his work with JDO 2 is going. They are building the TCK for JDO2
there, for example. They are being very public as well. I don't think
quite as public as what I'm suggesting above, but it's always worth
trying to ratchet things up a notch.
>
> I suppose this problem of interfacing the open source model with the
> JCP model has come up in other JCP-related Apache projects as well, so
> I hope to learn from existing ways to handle such a situation. I'm
> sorry if this is a FAQ, I couldn't find anything related with a quick
> search.
With the exception of JDO, other JCP-related Apache projects are only
implementing finished specifications. Even Tomcat, which is used as an
implementation area for servlets (although not the RI).
You would be breaking fresh ground at the ASF if you could convince the
EG to work this way. I will happily help you in any way that you could
use me.
Also, look at the Groovy project - that's also a JSR based on OSS, and
their activities might also be something to learn from.
>
> PS. Once Jackrabbit is promoted from the Incubator, I hope to have it
> included in the implementation list on http://www.apache.org/jcp/.
Please forgive me for not including it now. It should be there (and
after I dig out my driveway, I'll put it there...) I just forgot.
geir
>
> BR,
>
> Jukka Zitting
>
> --
> Yukatan - http://yukatan.fi/ - info@yukatan.fi
> Software craftmanship, JCR consulting, and Java development
>
>