[general] publish SPECjbb2005 scores

[Egor Pasko <eg...@gmail.com>]

Folks,

..just thinking.. would be really nice to have Harmony/DRLVM
SPECjbb2005 results published officially on spec.org (!!!) given the
number of performance work recently, I believe in good results right
from the start (and I believe in perf improvements:)

IMHO, it should help Harmony to gain interest from enterprise vendors
(I mean, those nice people who do not pay attention to DaCapo scores
for some (unknown) reason)

the funniest thing is that according to [1] non-profit cost is
non-zero for some (unknown) reason and looks like $125. But since
Apache is a real NON PROFIT, Apache does not pay even 2c, right?

I wonder how Apache cooks things like that?


[1] http://www.spec.org/order.html

-- 
Egor Pasko

Re: [general] publish SPECjbb2005 scores

[Stepan Mishura <st...@gmail.com>]

On 10/19/07, Mikhail Loenko <ml...@gmail.com> wrote:
> This is what I was thinking about as well.
>
> Non-profit/educational license for JBB2005 is 125$ [1]
>
> But some logistics is not clear for me:
>
> Though it's possible that one can buy it and grant to ASF, it's not clear
> who in ASF has authority to accept the grant
>
> It's also not clear who will then be able to use the SPEC and how we will
> maintain this "USER will not permit any third party to use or copy the
> original or any copy thereof. In the event of unauthorized transfer or
> copying, USER will pay SPEC a penalty in the amount of one additional
> fee for each such transfer or copy." from [2]
>
> The third thing that's not clear for me is who should approve submitting
> the results and behalf of Apache Harmony.
>

I think these questions should be addressed to legal-discuss@apache.org

-Stepan.

> Thanks,
> Mikhail
>
> [1] http://www.spec.org/order.html
> [2] http://www.spec.org/spec/docs/sample_license_agreement.pdf
>
> 18 Oct 2007 18:51:16 +0400, Egor Pasko <eg...@gmail.com>:
> > Folks,
> >
> > ..just thinking.. would be really nice to have Harmony/DRLVM
> > SPECjbb2005 results published officially on spec.org (!!!) given the
> > number of performance work recently, I believe in good results right
> > from the start (and I believe in perf improvements:)
> >
> > IMHO, it should help Harmony to gain interest from enterprise vendors
> > (I mean, those nice people who do not pay attention to DaCapo scores
> > for some (unknown) reason)
> >
> > the funniest thing is that according to [1] non-profit cost is
> > non-zero for some (unknown) reason and looks like $125. But since
> > Apache is a real NON PROFIT, Apache does not pay even 2c, right?
> >
> > I wonder how Apache cooks things like that?
> >
> >
> > [1] http://www.spec.org/order.html
> >
> > --
> > Egor Pasko

Re: [general] publish SPECjbb2005 scores

[Mikhail Loenko <ml...@gmail.com>]

This is what I was thinking about as well.

Non-profit/educational license for JBB2005 is 125$ [1]

But some logistics is not clear for me:

Though it's possible that one can buy it and grant to ASF, it's not clear
who in ASF has authority to accept the grant

It's also not clear who will then be able to use the SPEC and how we will
maintain this "USER will not permit any third party to use or copy the
original or any copy thereof. In the event of unauthorized transfer or
copying, USER will pay SPEC a penalty in the amount of one additional
fee for each such transfer or copy." from [2]

The third thing that's not clear for me is who should approve submitting
the results and behalf of Apache Harmony.

Thanks,
Mikhail

[1] http://www.spec.org/order.html
[2] http://www.spec.org/spec/docs/sample_license_agreement.pdf

18 Oct 2007 18:51:16 +0400, Egor Pasko <eg...@gmail.com>:
> Folks,
>
> ..just thinking.. would be really nice to have Harmony/DRLVM
> SPECjbb2005 results published officially on spec.org (!!!) given the
> number of performance work recently, I believe in good results right
> from the start (and I believe in perf improvements:)
>
> IMHO, it should help Harmony to gain interest from enterprise vendors
> (I mean, those nice people who do not pay attention to DaCapo scores
> for some (unknown) reason)
>
> the funniest thing is that according to [1] non-profit cost is
> non-zero for some (unknown) reason and looks like $125. But since
> Apache is a real NON PROFIT, Apache does not pay even 2c, right?
>
> I wonder how Apache cooks things like that?
>
>
> [1] http://www.spec.org/order.html
>
> --
> Egor Pasko
>
>

Re: [general] publish SPECjbb2005 scores

[Alexey Petrenko <al...@gmail.com>]

2007/10/24, Tim Ellison <t....@gmail.com>:
> Alexey Petrenko wrote:
> > Yes, it would be nice to publish the results. This can be a good step
> > for Harmony.
> > Tim, can you please comment on the possibility?
>
> I believe that there were discussions between SPEC and the ASF, before
> my time, that came to nothing.  Not sure if anything has changed since.
Could you please check it at your time? ;)

SY, Alexey

Re: [general] publish SPECjbb2005 scores

[Alexey Petrenko <al...@gmail.com>]

24 Oct 2007 12:14:45 +0400, Egor Pasko <eg...@gmail.com>:
> On the 0x379 day of Apache Harmony Tim Ellison wrote:
> > Alexey Petrenko wrote:
> > > Yes, it would be nice to publish the results. This can be a good step
> > > for Harmony.
> > > Tim, can you please comment on the possibility?
> >
> > I believe that there were discussions between SPEC and the ASF, before
> > my time, that came to nothing.  Not sure if anything has changed since.
>
> Tim,
>
> is it necessary for us that SPEC results are published by means of
> ASF?
It would be much better if it will be published by ASF...

SY, Alexey

RE: [general] publish SPECjbb2005 scores

["Elford, Chris L" <ch...@intel.com>]

It would certainly be possible for an individual committer to purchase a
license.  I believe SPEC will sell to anyone.  It is a bit unclear how
an individual would be entitled to the non-profit pricing unless it is
documented somehow they will use it exclusively for this purpose
(http://www.spec.org/order.html#educational).

In addition, it would be somewhat inefficient for an individual to
license it in this way.  Because only a single license is needed for a
whole organization (http://www.spec.org/order.html#license) if the
license is held by Apache or Apache Harmony, all use for publications,
stress testing, etc. by everyone in Harmony (committers only?) would be
possible.  Of course Apache/Apache Harmony would be required to maintain
appropriate access controls.  Other projects seem to be able to manage
similar controls to licensed 3rd party components (for example Geronimo
controlling access to J2EE JCK).

Does Harmony have a way to review/accept the license to the SPEC
benchmark if a cd/package arrived at Apache?

Thanks,
Chris Elford
Intel SSG/Enterprise Solutions Software Division

-----Original Message-----
From: Tim Ellison [mailto:t.p.ellison@gmail.com] 
Sent: Wednesday, October 24, 2007 1:20 PM
To: dev@harmony.apache.org
Subject: Re: [general] publish SPECjbb2005 scores

Egor Pasko wrote:
> On the 0x379 day of Apache Harmony Tim Ellison wrote:
>> Alexey Petrenko wrote:
>>> Yes, it would be nice to publish the results. This can be a good
step
>>> for Harmony.
>>> Tim, can you please comment on the possibility?
>> I believe that there were discussions between SPEC and the ASF,
before
>> my time, that came to nothing.  Not sure if anything has changed
since.
> 
> Tim,
> 
> is it necessary for us that SPEC results are published by means of
> ASF? Maybe it is possible that some committer buys a license, measures
> all by him/herself and publishes the result? or more complicated way:
> give someone a right to publish scores?

I don't know the details of SPEC reporting.  That would be an issue
between the individual and SPEC organization.

> also, if the discussion you mentioned is googleable, I would be happy
> to take a look at the devil details, since, really, times could have
> changed. Sorry, could not find any relevant link :(

My information is just based upon an informal discussion with a Geronimo
developer.  Sorry, I don't know more about it than that.

At least this link [1] implies somebody was trying to run it.

[1] http://cwiki.apache.org/GMOxDOC10/specjappserver2004.html

Regards,
Tim

Re: [general] publish SPECjbb2005 scores

[Alexey Petrenko <al...@gmail.com>]

2007/10/24, Tim Ellison <t....@gmail.com>:
> Egor Pasko wrote:
> > On the 0x379 day of Apache Harmony Tim Ellison wrote:
> >> Alexey Petrenko wrote:
> >>> Yes, it would be nice to publish the results. This can be a good step
> >>> for Harmony.
> >>> Tim, can you please comment on the possibility?
> >> I believe that there were discussions between SPEC and the ASF, before
> >> my time, that came to nothing.  Not sure if anything has changed since.
> >
> > Tim,
> >
> > is it necessary for us that SPEC results are published by means of
> > ASF? Maybe it is possible that some committer buys a license, measures
> > all by him/herself and publishes the result? or more complicated way:
> > give someone a right to publish scores?
>
> I don't know the details of SPEC reporting.  That would be an issue
> between the individual and SPEC organization.
>
> > also, if the discussion you mentioned is googleable, I would be happy
> > to take a look at the devil details, since, really, times could have
> > changed. Sorry, could not find any relevant link :(
>
> My information is just based upon an informal discussion with a Geronimo
> developer.  Sorry, I don't know more about it than that.
>
> At least this link [1] implies somebody was trying to run it.
>
> [1] http://cwiki.apache.org/GMOxDOC10/specjappserver2004.html
Actually the situation is quite different for Sjbb, and also changed
for SjAS with the introduction of EAStress in 1.08 version. SPEC made
a large step forward to Open Source community and Apache – why should
not we use the momentum and publish the scores?

Tim,
Is there any process in Apache to make such decisions? It should be :)
Can we have a try to start this process again?

Thanks in advance.

SY, Alexey

Re: [general] publish SPECjbb2005 scores

[Tim Ellison <t....@gmail.com>]

Egor Pasko wrote:
> On the 0x379 day of Apache Harmony Tim Ellison wrote:
>> Alexey Petrenko wrote:
>>> Yes, it would be nice to publish the results. This can be a good step
>>> for Harmony.
>>> Tim, can you please comment on the possibility?
>> I believe that there were discussions between SPEC and the ASF, before
>> my time, that came to nothing.  Not sure if anything has changed since.
> 
> Tim,
> 
> is it necessary for us that SPEC results are published by means of
> ASF? Maybe it is possible that some committer buys a license, measures
> all by him/herself and publishes the result? or more complicated way:
> give someone a right to publish scores?

I don't know the details of SPEC reporting.  That would be an issue
between the individual and SPEC organization.

> also, if the discussion you mentioned is googleable, I would be happy
> to take a look at the devil details, since, really, times could have
> changed. Sorry, could not find any relevant link :(

My information is just based upon an informal discussion with a Geronimo
developer.  Sorry, I don't know more about it than that.

At least this link [1] implies somebody was trying to run it.

[1] http://cwiki.apache.org/GMOxDOC10/specjappserver2004.html

Regards,
Tim

Re: [general] publish SPECjbb2005 scores

[Egor Pasko <eg...@gmail.com>]

On the 0x379 day of Apache Harmony Tim Ellison wrote:
> Alexey Petrenko wrote:
> > Yes, it would be nice to publish the results. This can be a good step
> > for Harmony.
> > Tim, can you please comment on the possibility?
> 
> I believe that there were discussions between SPEC and the ASF, before
> my time, that came to nothing.  Not sure if anything has changed since.

Tim,

is it necessary for us that SPEC results are published by means of
ASF? Maybe it is possible that some committer buys a license, measures
all by him/herself and publishes the result? or more complicated way:
give someone a right to publish scores?

also, if the discussion you mentioned is googleable, I would be happy
to take a look at the devil details, since, really, times could have
changed. Sorry, could not find any relevant link :(

> Regards,
> Tim
> 
> > Thanks in advance.
> > 
> > SY, Alexey
> > 
> > 18 Oct 2007 18:51:16 +0400, Egor Pasko <eg...@gmail.com>:
> >> Folks,
> >>
> >> ..just thinking.. would be really nice to have Harmony/DRLVM
> >> SPECjbb2005 results published officially on spec.org (!!!) given the
> >> number of performance work recently, I believe in good results right
> >> from the start (and I believe in perf improvements:)
> >>
> >> IMHO, it should help Harmony to gain interest from enterprise vendors
> >> (I mean, those nice people who do not pay attention to DaCapo scores
> >> for some (unknown) reason)
> >>
> >> the funniest thing is that according to [1] non-profit cost is
> >> non-zero for some (unknown) reason and looks like $125. But since
> >> Apache is a real NON PROFIT, Apache does not pay even 2c, right?
> >>
> >> I wonder how Apache cooks things like that?
> >>
> >>
> >> [1] http://www.spec.org/order.html
> >>
> >> --
> >> Egor Pasko
> >>
> >>
> > 
> 

-- 
Egor Pasko

Re: [general] publish SPECjbb2005 scores

[Tim Ellison <t....@gmail.com>]

Alexey Petrenko wrote:
> Yes, it would be nice to publish the results. This can be a good step
> for Harmony.
> Tim, can you please comment on the possibility?

I believe that there were discussions between SPEC and the ASF, before
my time, that came to nothing.  Not sure if anything has changed since.

Regards,
Tim

> Thanks in advance.
> 
> SY, Alexey
> 
> 18 Oct 2007 18:51:16 +0400, Egor Pasko <eg...@gmail.com>:
>> Folks,
>>
>> ..just thinking.. would be really nice to have Harmony/DRLVM
>> SPECjbb2005 results published officially on spec.org (!!!) given the
>> number of performance work recently, I believe in good results right
>> from the start (and I believe in perf improvements:)
>>
>> IMHO, it should help Harmony to gain interest from enterprise vendors
>> (I mean, those nice people who do not pay attention to DaCapo scores
>> for some (unknown) reason)
>>
>> the funniest thing is that according to [1] non-profit cost is
>> non-zero for some (unknown) reason and looks like $125. But since
>> Apache is a real NON PROFIT, Apache does not pay even 2c, right?
>>
>> I wonder how Apache cooks things like that?
>>
>>
>> [1] http://www.spec.org/order.html
>>
>> --
>> Egor Pasko
>>
>>
> 

Re: [general] publish SPECjbb2005 scores

[Alexey Petrenko <al...@gmail.com>]

Yes, it would be nice to publish the results. This can be a good step
for Harmony.
Tim, can you please comment on the possibility?

Thanks in advance.

SY, Alexey

18 Oct 2007 18:51:16 +0400, Egor Pasko <eg...@gmail.com>:
> Folks,
>
> ..just thinking.. would be really nice to have Harmony/DRLVM
> SPECjbb2005 results published officially on spec.org (!!!) given the
> number of performance work recently, I believe in good results right
> from the start (and I believe in perf improvements:)
>
> IMHO, it should help Harmony to gain interest from enterprise vendors
> (I mean, those nice people who do not pay attention to DaCapo scores
> for some (unknown) reason)
>
> the funniest thing is that according to [1] non-profit cost is
> non-zero for some (unknown) reason and looks like $125. But since
> Apache is a real NON PROFIT, Apache does not pay even 2c, right?
>
> I wonder how Apache cooks things like that?
>
>
> [1] http://www.spec.org/order.html
>
> --
> Egor Pasko
>
>

Re: [general] publish SPECjbb2005 scores

[Spark Shen <sm...@gmail.com>]

Sounds a good idea. +1.

Actually, I have the similar thinking[1] several days ago.
I just thought about publish more test results. For example coverage report,

unit test report etc. on our web site. The idea is to prove harmony is in a
great shape.

IMO, coverage and unit test is especially import for a project such as
harmony -
with huge code base and fundamental functions.

Maven project has a similar link on its web site[2],
with different kinds of test reports published. How about we extend out
performance
link[3] to a similar form?

I departured from your original idea a little, do I? :-)

[1]http://article.gmane.org/gmane.comp.java.harmony.devel/29632
[2] http://maven.apache.org/ref/current/maven-model/project-reports.html
[3] http://harmony.apache.org/performance.html

18 Oct 2007 18:51:16 +0400, Egor Pasko <eg...@gmail.com>:
>
> Folks,
>
> ..just thinking.. would be really nice to have Harmony/DRLVM
> SPECjbb2005 results published officially on spec.org (!!!) given the
> number of performance work recently, I believe in good results right
> from the start (and I believe in perf improvements:)
>
> IMHO, it should help Harmony to gain interest from enterprise vendors
> (I mean, those nice people who do not pay attention to DaCapo scores
> for some (unknown) reason)
>
> the funniest thing is that according to [1] non-profit cost is
> non-zero for some (unknown) reason and looks like $125. But since
> Apache is a real NON PROFIT, Apache does not pay even 2c, right?
>
> I wonder how Apache cooks things like that?
>
>
> [1] http://www.spec.org/order.html
>
> --
> Egor Pasko
>
>


-- 
Spark Shen
China Software Development Lab, IBM