headers
Steve Poole | 1 Oct 2008 11:43

JSR-326 Project proposal


Greetings - I thought it was worth posting to this list the fact that I 
have created an incubator project proposal for JSR 326.    I work for 
IBM and  am the spec lead for the JSR.  The JSR ( 
http://jcp.org/en/jsr/detail?id=326)  is concerned with creating a new 
API for postmortem JVM Diagnostics. 

My proposal needs mentors and a champion.

It seems to me that one of the areas we will need help in is concerned 
with the fact that we want to run the JSR development in an open way.  
Effectively the aim is that the spec will fall out of the RI. There are 
and will be interesting questions in attempting to develop spec and RI 
in this way.  For instance how much of the JCP process is actually 
applicable when a project is completely open and how to position an 
"expert group" when the whole project is open to the community.

The proposal is here http://wiki.apache.org/incubator/KatoProposal

Steve Poole
Permalink | Reply | 6 comments |
headers
Steve Loughran | 1 Oct 2008 11:56
Picon
Favicon

Re: JSR-326 Project proposal

Steve Poole wrote:
> 
> Greetings - I thought it was worth posting to this list the fact that I 
> have created an incubator project proposal for JSR 326.    I work for 
> IBM and  am the spec lead for the JSR.  The JSR ( 
> http://jcp.org/en/jsr/detail?id=326)  is concerned with creating a new 
> API for postmortem JVM Diagnostics.
> My proposal needs mentors and a champion.
> 
> It seems to me that one of the areas we will need help in is concerned 
> with the fact that we want to run the JSR development in an open way.  
> Effectively the aim is that the spec will fall out of the RI. There are 
> and will be interesting questions in attempting to develop spec and RI 
> in this way.  For instance how much of the JCP process is actually 
> applicable when a project is completely open and how to position an 
> "expert group" when the whole project is open to the community.
> 
> The proposal is here http://wiki.apache.org/incubator/KatoProposal
> 

I've been involved in some grid standards we did fully in the open . A 
key feature was BSD/apache licensed tests all developed in a public 
repository (For us, sourceforge), and hook up a CI build to continually 
test things, such as lots of XML documents against the proposed schemas. 
Interop testing was more fun, but tha's a separate topic.

So
  -the test suite comes out sooner rather than later (caused some push 
back from those people who haven't written working code for a long time)
  -no secret test suite
(Continue reading)

Permalink | Reply | 2 comments |
headers
Steve Loughran | 1 Oct 2008 11:59
Picon
Favicon

Re: JSR-326 Project proposal

Steve Poole wrote:
> 
> Greetings - I thought it was worth posting to this list the fact that I 
> have created an incubator project proposal for JSR 326.    I work for 
> IBM and  am the spec lead for the JSR.  The JSR ( 
> http://jcp.org/en/jsr/detail?id=326)  is concerned with creating a new 
> API for postmortem JVM Diagnostics.
> My proposal needs mentors and a champion.
> 
> It seems to me that one of the areas we will need help in is concerned 
> with the fact that we want to run the JSR development in an open way.  
> Effectively the aim is that the spec will fall out of the RI. There are 
> and will be interesting questions in attempting to develop spec and RI 
> in this way.  For instance how much of the JCP process is actually 
> applicable when a project is completely open and how to position an 
> "expert group" when the whole project is open to the community.
> 
> The proposal is here http://wiki.apache.org/incubator/KatoProposal
> 
> Steve Poole
> 
> 

I should add that Hadoop might be interested in this, though it doesnt 
work (yet) with harmony. currently if a forked JVM is to be killed, it 
does some abuse to get the process ID, then a kill -QUIT to dump the 
threads. In an ideal world, a complete postmortem would be obtained, a 
post mortem *that can be saved into the Hadoop filesystem for later 
analysis*. That means no fancy XML content please, but instead stuff 
that MapReduce likes.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Geir Magnusson Jr. | 1 Oct 2008 12:13
Picon
Favicon

Re: JSR-326 Project proposal


On Oct 1, 2008, at 5:43 AM, Steve Poole wrote:

>
> Greetings - I thought it was worth posting to this list the fact  
> that I have created an incubator project proposal for JSR 326.    I  
> work for IBM and  am the spec lead for the JSR.  The JSR ( http://jcp.org/en/jsr/detail?id=326) 
>   is concerned with creating a new API for postmortem JVM Diagnostics.
> My proposal needs mentors and a champion.

happy to be both a mentor and a champion
>

>
> It seems to me that one of the areas we will need help in is  
> concerned with the fact that we want to run the JSR development in  
> an open way.  Effectively the aim is that the spec will fall out of  
> the RI. There are and will be interesting questions in attempting to  
> develop spec and RI in this way.  For instance how much of the JCP  
> process is actually applicable when a project is completely open and  
> how to position an "expert group" when the whole project is open to  
> the community.

Well...  all of it.  There should be no real problems.  The only  
concern I can think of is ensuring that you get a grant of "necessary  
IP" from each participant such that implementations of the spec aren't  
going to be encumbered by someone (like Sun) willing to use their  
patent portfolio to control who can create implementations and what  
they can do with them.
(Continue reading)

Permalink | Reply | 1 comment |
headers
Steve Poole | 2 Oct 2008 10:19

Re: JSR-326 Project proposal

Geir Magnusson Jr. wrote:
>
> On Oct 1, 2008, at 5:43 AM, Steve Poole wrote:
>
>>
>> Greetings - I thought it was worth posting to this list the fact that 
>> I have created an incubator project proposal for JSR 326. I work for 
>> IBM and am the spec lead for the JSR. The JSR ( 
>> http://jcp.org/en/jsr/detail?id=326) is concerned with creating a new 
>> API for postmortem JVM Diagnostics.
>> My proposal needs mentors and a champion.
>
> happy to be both a mentor and a champion
>>
> Thanks Geir - much appreciated. I'll update the proposal accordingly. 
> Can you give the proposal the once over and let me know what else you 
> think needs addressing before we can call a vote?
>>
>> It seems to me that one of the areas we will need help in is 
>> concerned with the fact that we want to run the JSR development in an 
>> open way. Effectively the aim is that the spec will fall out of the 
>> RI. There are and will be interesting questions in attempting to 
>> develop spec and RI in this way. For instance how much of the JCP 
>> process is actually applicable when a project is completely open and 
>> how to position an "expert group" when the whole project is open to 
>> the community.
>
> Well... all of it. There should be no real problems. The only concern 
> I can think of is ensuring that you get a grant of "necessary IP" from 
> each participant such that implementations of the spec aren't going to
(Continue reading)

Permalink | Reply | 0 comments |
headers
Steve Poole | 2 Oct 2008 10:48

Re: JSR-326 Project proposal

Steve Loughran wrote:
> Steve Poole wrote:
>>
>> Greetings - I thought it was worth posting to this list the fact that 
>> I have created an incubator project proposal for JSR 326.    I work 
>> for IBM and  am the spec lead for the JSR.  The JSR ( 
>> http://jcp.org/en/jsr/detail?id=326)  is concerned with creating a 
>> new API for postmortem JVM Diagnostics.
>> My proposal needs mentors and a champion.
>>
>> It seems to me that one of the areas we will need help in is 
>> concerned with the fact that we want to run the JSR development in an 
>> open way.  Effectively the aim is that the spec will fall out of the 
>> RI. There are and will be interesting questions in attempting to 
>> develop spec and RI in this way.  For instance how much of the JCP 
>> process is actually applicable when a project is completely open and 
>> how to position an "expert group" when the whole project is open to 
>> the community.
>>
>> The proposal is here http://wiki.apache.org/incubator/KatoProposal
>>
>
> I've been involved in some grid standards we did fully in the open . A 
> key feature was BSD/apache licensed tests all developed in a public 
> repository (For us, sourceforge), and hook up a CI build to 
> continually test things, such as lots of XML documents against the 
> proposed schemas. Interop testing was more fun, but tha's a separate 
> topic.
>
> So
(Continue reading)

Permalink | Reply | 1 comment |
headers
Steve Loughran | 2 Oct 2008 12:03
Picon
Favicon

Re: JSR-326 Project proposal

Steve Poole wrote:
> Steve Loughran wrote:
>> Steve Poole wrote:
>>>
>>> Greetings - I thought it was worth posting to this list the fact that 
>>> I have created an incubator project proposal for JSR 326.    I work 
>>> for IBM and  am the spec lead for the JSR.  The JSR ( 
>>> http://jcp.org/en/jsr/detail?id=326)  is concerned with creating a 
>>> new API for postmortem JVM Diagnostics.
>>> My proposal needs mentors and a champion.
>>>
>>> It seems to me that one of the areas we will need help in is 
>>> concerned with the fact that we want to run the JSR development in an 
>>> open way.  Effectively the aim is that the spec will fall out of the 
>>> RI. There are and will be interesting questions in attempting to 
>>> develop spec and RI in this way.  For instance how much of the JCP 
>>> process is actually applicable when a project is completely open and 
>>> how to position an "expert group" when the whole project is open to 
>>> the community.
>>>
>>> The proposal is here http://wiki.apache.org/incubator/KatoProposal
>>>
>>
>> I've been involved in some grid standards we did fully in the open . A 
>> key feature was BSD/apache licensed tests all developed in a public 
>> repository (For us, sourceforge), and hook up a CI build to 
>> continually test things, such as lots of XML documents against the 
>> proposed schemas. Interop testing was more fun, but tha's a separate 
>> topic.
>>
(Continue reading)

Permalink | Reply | 0 comments |

Gmane