OpenJDK projects promoting proprietary builds

[Mark Wielaard]

Hi,

Since we are all in frantic-release mode just before JavaOne I wanted to
double check some of the new features against other builds. I was
somewhat surprised to find that various projects only promote and/or
publish proprietary builds. The jdk7 binaries published on the list, but
also nio2 builds, sctp builds (these were the two I was interested in,
there might be others), are published under proprietary terms and not
just under the GPL.

The terms are particularly anti-social since they explicitly forbid
sharing the binaries with others, learning what the corresponding source
code is, modifying or creating any derivative works, and agreeing that
Sun will be "irreparably harmed" if you don't keep everything you learn
from those binaries confidential.

Could we please have OpenJDK projects only publish artifacts under the
terms listed at http://openjdk.java.net/legal/

Those terms even allows publishing binaries that contain some of the
proprietary blobs as long as the rest is published under the GPL, thanks
to the assembly exception, if that really is necessary (and it really
shouldn't be necessary anymore since people have been publishing full
GPLed builds for almost 2 years now).

Thanks,

Mark

[Geir Magnusson Jr.-3]

On May 29, 2009, at 12:56 PM, Mark Wielaard wrote:

> Hi,
>
> Since we are all in frantic-release mode just before JavaOne I  
> wanted to
> double check some of the new features against other builds. I was
> somewhat surprised to find that various projects only promote and/or
> publish proprietary builds. The jdk7 binaries published on the list,  
> but
> also nio2 builds, sctp builds (these were the two I was interested in,
> there might be others), are published under proprietary terms and not
> just under the GPL.
>
>
> The terms are particularly anti-social since they explicitly forbid
> sharing the binaries with others, learning what the corresponding  
> source
> code is, modifying or creating any derivative works, and agreeing that
> Sun will be "irreparably harmed" if you don't keep everything you  
> learn
> from those binaries confidential.

Imagine!  There's a downside for giving sun complete copyright (via  
joint assignment)!  Whoda thought!

geir

>
>
> Could we please have OpenJDK projects only publish artifacts under the
> terms listed at http://openjdk.java.net/legal/
>
> Those terms even allows publishing binaries that contain some of the
> proprietary blobs as long as the rest is published under the GPL,  
> thanks
> to the assembly exception, if that really is necessary (and it really
> shouldn't be necessary anymore since people have been publishing full
> GPLed builds for almost 2 years now).
>
> Thanks,
>
> Mark
>

[Mark Wielaard]

Hi Geir,

On Fri, 2009-05-29 at 13:06 -0400, Geir Magnusson Jr. wrote:
> On May 29, 2009, at 12:56 PM, Mark Wielaard wrote:
> > The terms are particularly anti-social since they explicitly forbid
> > sharing the binaries with others, learning what the corresponding  
> > source code is, modifying or creating any derivative works, and
> > agreeing that Sun will be "irreparably harmed" if you don't keep
> > everything you learn from those binaries confidential.
>
> Imagine!  There's a downside for giving sun complete copyright (via  
> joint assignment)!  Whoda thought!

Of course. That is a reason to think long and hard before signing the
SCA in its current form, if you wish to do that. But that is a different
issue. What Sun does with its own internal proprietary forks is their
business. This is about OpenJDK project themselves publishing and
promoting proprietary builds, which IMHO is not done.

> > Could we please have OpenJDK projects only publish artifacts under the
> > terms listed at http://openjdk.java.net/legal/
> >
> > Those terms even allows publishing binaries that contain some of the
> > proprietary blobs as long as the rest is published under the GPL,  
> > thanks to the assembly exception, if that really is necessary (and
> > it really shouldn't be necessary anymore since people have been
> > publishing full GPLed builds for almost 2 years now).

Cheers,

Mark

[David Herron-2]

Unless something has changed since I left the OpenJDK project isn't
publishing binary builds.

The situation when I was still there was JDK builds were being published but
not OpenJDK builds.

That was a conscious decision to not do so.

Mebbe the JDK7 build announcements shouldn't be published on OpenJDK mailing
lists for the same reason as I stopped (last fall) sending DLJ announcements
to OpenJDK mailing lists?

- David Herron
http://davidherron.com


On Fri, May 29, 2009 at 10:56 AM, Mark Wielaard <mark@...> wrote:

> Hi Geir,
>
> On Fri, 2009-05-29 at 13:06 -0400, Geir Magnusson Jr. wrote:
> > On May 29, 2009, at 12:56 PM, Mark Wielaard wrote:
> > > The terms are particularly anti-social since they explicitly forbid
> > > sharing the binaries with others, learning what the corresponding
> > > source code is, modifying or creating any derivative works, and
> > > agreeing that Sun will be "irreparably harmed" if you don't keep
> > > everything you learn from those binaries confidential.
> >
> > Imagine!  There's a downside for giving sun complete copyright (via
> > joint assignment)!  Whoda thought!
>
> Of course. That is a reason to think long and hard before signing the
> SCA in its current form, if you wish to do that. But that is a different
> issue. What Sun does with its own internal proprietary forks is their
> business. This is about OpenJDK project themselves publishing and
> promoting proprietary builds, which IMHO is not done.
>
> > > Could we please have OpenJDK projects only publish artifacts under the
> > > terms listed at http://openjdk.java.net/legal/
> > >
> > > Those terms even allows publishing binaries that contain some of the
> > > proprietary blobs as long as the rest is published under the GPL,
> > > thanks to the assembly exception, if that really is necessary (and
> > > it really shouldn't be necessary anymore since people have been
> > > publishing full GPLed builds for almost 2 years now).
>
> Cheers,
>
> Mark
>
>

[kelly.ohair]

David Herron wrote:
> Unless something has changed since I left the OpenJDK project isn't
> publishing binary builds.

My understanding too. OpenJDK projects are source only.

>
> The situation when I was still there was JDK builds were being published but
> not OpenJDK builds.

Correct as far as I know too.

>
> That was a conscious decision to not do so.

Yup.

>
> Mebbe the JDK7 build announcements shouldn't be published on OpenJDK mailing
> lists for the same reason as I stopped (last fall) sending DLJ announcements
> to OpenJDK mailing lists?

I tend to agree.

-kto

>
> - David Herron
> http://davidherron.com
>
>
> On Fri, May 29, 2009 at 10:56 AM, Mark Wielaard <mark@...> wrote:
>
>> Hi Geir,
>>
>> On Fri, 2009-05-29 at 13:06 -0400, Geir Magnusson Jr. wrote:
>>> On May 29, 2009, at 12:56 PM, Mark Wielaard wrote:
>>>> The terms are particularly anti-social since they explicitly forbid
>>>> sharing the binaries with others, learning what the corresponding
>>>> source code is, modifying or creating any derivative works, and
>>>> agreeing that Sun will be "irreparably harmed" if you don't keep
>>>> everything you learn from those binaries confidential.
>>> Imagine!  There's a downside for giving sun complete copyright (via
>>> joint assignment)!  Whoda thought!
>> Of course. That is a reason to think long and hard before signing the
>> SCA in its current form, if you wish to do that. But that is a different
>> issue. What Sun does with its own internal proprietary forks is their
>> business. This is about OpenJDK project themselves publishing and
>> promoting proprietary builds, which IMHO is not done.
>>
>>>> Could we please have OpenJDK projects only publish artifacts under the
>>>> terms listed at http://openjdk.java.net/legal/
>>>>
>>>> Those terms even allows publishing binaries that contain some of the
>>>> proprietary blobs as long as the rest is published under the GPL,
>>>> thanks to the assembly exception, if that really is necessary (and
>>>> it really shouldn't be necessary anymore since people have been
>>>> publishing full GPLed builds for almost 2 years now).
>> Cheers,
>>
>> Mark
>>
>>

[Mark Wielaard]

Hi David,

On Fri, 2009-05-29 at 11:20 -0700, David Herron wrote:
> Unless something has changed since I left the OpenJDK project isn't
> publishing binary builds.

Various groups now have links to binary build downloads on their project
page. These builds are however not published under the normal terms that
the OpenJDK project uses (GPL plus various extra permissions), but under
terms that prevent all the normal activities that Free Software licenses
grant.

> Mebbe the JDK7 build announcements shouldn't be published on OpenJDK

I might not have been clear. I would not object to having builds to
compare against. In this case I was happy to see binary builds since I
was seeing some failures in the nio2 jtreg results and wanted to make
sure it wasn't local issues (admittedly my local builds are slightly
hacked up). The problem is just that they aren't published under
acceptable terms. That was probably an oversight. I assume they should
have been published under the normal GPL terms.

Cheers,

Mark

[gnu_andrew]

I agree wholeheartedly, but have to say I long ago ceased to be
surprised by Sun builds beinge proprietary. Sadly the converse is
true; I'd be surprised by a Sun build released under the same terms as
our IcedTea builds.

On 29/05/2009, Mark Wielaard <mark@...> wrote:

> Hi,
>
> Since we are all in frantic-release mode just before JavaOne I wanted to
> double check some of the new features against other builds. I was
> somewhat surprised to find that various projects only promote and/or
> publish proprietary builds. The jdk7 binaries published on the list, but
> also nio2 builds, sctp builds (these were the two I was interested in,
> there might be others), are published under proprietary terms and not
> just under the GPL.
>
> The terms are particularly anti-social since they explicitly forbid
> sharing the binaries with others, learning what the corresponding source
> code is, modifying or creating any derivative works, and agreeing that
> Sun will be "irreparably harmed" if you don't keep everything you learn
> from those binaries confidential.
>
> Could we please have OpenJDK projects only publish artifacts under the
> terms listed at http://openjdk.java.net/legal/
>
> Those terms even allows publishing binaries that contain some of the
> proprietary blobs as long as the rest is published under the GPL, thanks
> to the assembly exception, if that really is necessary (and it really
> shouldn't be necessary anymore since people have been publishing full
> GPLed builds for almost 2 years now).
>
> Thanks,
>
> Mark
>
>

--
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8

[Mark Wielaard]

On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
> I agree wholeheartedly, but have to say I long ago ceased to be
> surprised by Sun builds beinge proprietary. Sadly the converse is
> true; I'd be surprised by a Sun build released under the same terms as
> our IcedTea builds.

And that is indeed what is sad about this. That it seems OpenJDK builds
are actually Sun builds, and by extension such things are proprietary.
And that is what I object to. OpenJDK builds should be just that,
OpenJDK builds distributed under the (GPL) terms everybody in our
community adheres to.

If a project wants to publish "early access" builds then they really
should if they feel people would like to play with the bits. But such
builds should follow the standard OpenJDK project rules
(http://openjdk.java.net/legal/) that everybody else also uses.

Going to Sun legal and requesting alternative proprietary terms and then
publishing the code and binaries under non-free software licenses is
just bad for creating a community. It is bad enough that the current SCA
rules around OpenJDK assign all rights to one commercial party, Sun. But
projects then abusing those rights by pushing proprietary derivatives as
early access OpenJDK project builds undermines the whole community of
equals.

You are right that we have IcedTea to fix that. If you get your packages
through IcedTea (derivatives) you are guaranteed that it truly is Free
Software. But wouldn't it be better if we could say that about OpenJDK
itself? Wouldn't that make the community stronger?

Cheers,

Mark

[Martin Buchholz-3]

On Fri, May 29, 2009 at 11:52, Mark Wielaard <mark@...> wrote:

>
> I might not have been clear. I would not object to having builds to
> compare against. In this case I was happy to see binary builds since I
> was seeing some failures in the nio2 jtreg results and wanted to make
> sure it wasn't local issues (admittedly my local builds are slightly
> hacked up). The problem is just that they aren't published under
> acceptable terms.


I would also like to see binaries built from the openjdk sources proper,
but Sun is under no obligation to do that, and the proprietary binaries
are in any case a valuable service for folks on both sides of the
source code freedom ideological divide (including myself),
so I would like to see them continue to be
announced on openjdk mailing lists.

Martin


> That was probably an oversight. I assume they should
> have been published under the normal GPL terms.
>
> Cheers,
>
> Mark
>
>

[David Herron-2]

Mark,

Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7 the
difference is very small.  The JDK7 builds have some proprietary bits in
them.

It's valuable to the JDK product cycle for JDK builds to have early access
exposure so people can report bugs etc.  Sun started doing very-early-access
releases with JDK6 and the Peabody Project, and early exposure was a purpose
of the <project-name-never-to-be-spoken-again> Regressions Contest which I
ran in early 2006. (See my java.net blog posting of Jan 30, 2006)  I'm sure
you can understand the value, right?

There would also be value to the OpenJDK project for reference OpenJDK
builds to be available.  For example to help those like you who are involved
with packaging OpenJDK-derived builds.  Anybody could do those builds
couldn't they?

I don't think it's correct to say Sun is "pushing proprietary derivatives as
early access OpenJDK builds.." is it?  The name JDK7 is distinguished from
OpenJDK7, right?  Isn't it well known that they are approximately 96% the
same and that there are differences in specific areas?

- David Herron

On Sat, May 30, 2009 at 12:31 PM, Mark Wielaard <mark@...> wrote:

> On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
> > I agree wholeheartedly, but have to say I long ago ceased to be
> > surprised by Sun builds beinge proprietary. Sadly the converse is
> > true; I'd be surprised by a Sun build released under the same terms as
> > our IcedTea builds.
>
> And that is indeed what is sad about this. That it seems OpenJDK builds
> are actually Sun builds, and by extension such things are proprietary.
> And that is what I object to. OpenJDK builds should be just that,
> OpenJDK builds distributed under the (GPL) terms everybody in our
> community adheres to.
>
> If a project wants to publish "early access" builds then they really
> should if they feel people would like to play with the bits. But such
> builds should follow the standard OpenJDK project rules
> (http://openjdk.java.net/legal/) that everybody else also uses.
>
> Going to Sun legal and requesting alternative proprietary terms and then
> publishing the code and binaries under non-free software licenses is
> just bad for creating a community. It is bad enough that the current SCA
> rules around OpenJDK assign all rights to one commercial party, Sun. But
> projects then abusing those rights by pushing proprietary derivatives as
> early access OpenJDK project builds undermines the whole community of
> equals.
>
> You are right that we have IcedTea to fix that. If you get your packages
> through IcedTea (derivatives) you are guaranteed that it truly is Free
> Software. But wouldn't it be better if we could say that about OpenJDK
> itself? Wouldn't that make the community stronger?
>
> Cheers,
>
> Mark
>
>

[Geir Magnusson Jr.-3]

Hi David,

On May 30, 2009, at 7:05 PM, David Herron wrote:

> Mark,
>
> Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7  
> the
> difference is very small.  The JDK7 builds have some proprietary  
> bits in
> them.

Why?  For heaven's sake... why?

>
>
> It's valuable to the JDK product cycle for JDK builds to have early  
> access
> exposure so people can report bugs etc.  Sun started doing very-
> early-access
> releases with JDK6 and the Peabody Project, and early exposure was a  
> purpose
> of the <project-name-never-to-be-spoken-again> Regressions Contest  
> which I
> ran in early 2006. (See my java.net blog posting of Jan 30, 2006)  
> I'm sure
> you can understand the value, right?
>
> There would also be value to the OpenJDK project for reference OpenJDK
> builds to be available.  For example to help those like you who are  
> involved
> with packaging OpenJDK-derived builds.  Anybody could do those builds
> couldn't they?
>
> I don't think it's correct to say Sun is "pushing proprietary  
> derivatives as
> early access OpenJDK builds.." is it?  The name JDK7 is  
> distinguished from
> OpenJDK7, right?  Isn't it well known that they are approximately  
> 96% the
> same and that there are differences in specific areas?

As an interested observer and fan of open and even Free(tm) Java, I  
need to ask why would you want to have this differentiation?

I can understand the need to provide source and/or binaries to  
commercial partners and customers under licenses that aren't the GPL,  
but given your right to relicense the whole thing, the same code  
should be able to be offered under the GPL...

geir

>
>
> - David Herron
>
> On Sat, May 30, 2009 at 12:31 PM, Mark Wielaard <mark@...>  
> wrote:
>
>> On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
>>> I agree wholeheartedly, but have to say I long ago ceased to be
>>> surprised by Sun builds beinge proprietary. Sadly the converse is
>>> true; I'd be surprised by a Sun build released under the same  
>>> terms as
>>> our IcedTea builds.
>>
>> And that is indeed what is sad about this. That it seems OpenJDK  
>> builds
>> are actually Sun builds, and by extension such things are  
>> proprietary.
>> And that is what I object to. OpenJDK builds should be just that,
>> OpenJDK builds distributed under the (GPL) terms everybody in our
>> community adheres to.
>>
>> If a project wants to publish "early access" builds then they really
>> should if they feel people would like to play with the bits. But such
>> builds should follow the standard OpenJDK project rules
>> (http://openjdk.java.net/legal/) that everybody else also uses.
>>
>> Going to Sun legal and requesting alternative proprietary terms and  
>> then
>> publishing the code and binaries under non-free software licenses is
>> just bad for creating a community. It is bad enough that the  
>> current SCA
>> rules around OpenJDK assign all rights to one commercial party,  
>> Sun. But
>> projects then abusing those rights by pushing proprietary  
>> derivatives as
>> early access OpenJDK project builds undermines the whole community of
>> equals.
>>
>> You are right that we have IcedTea to fix that. If you get your  
>> packages
>> through IcedTea (derivatives) you are guaranteed that it truly is  
>> Free
>> Software. But wouldn't it be better if we could say that about  
>> OpenJDK
>> itself? Wouldn't that make the community stronger?
>>
>> Cheers,
>>
>> Mark
>>
>>

[Dmitri Trembovetski]

Geir Magnusson Jr. wrote:

> Hi David,
>
> On May 30, 2009, at 7:05 PM, David Herron wrote:
>
>> Mark,
>>
>> Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7 the
>> difference is very small.  The JDK7 builds have some proprietary bits in
>> them.
>
> Why?  For heaven's sake... why?

   Because the corresponding open source parts aren't good enough yet and we
don't have enough resources to make them on par with the proprietary bits
although this is what we want in the long run.

   Specific parts that I know of are color management, AA shape rasterizer and
font rasterizer.

   You must understand that "passing the TCK" doesn't necessarily mean "has
acceptable performance, fidelity and stability".

   Thanks,
     Dmitri

>
>>
>>
>> It's valuable to the JDK product cycle for JDK builds to have early
>> access
>> exposure so people can report bugs etc.  Sun started doing
>> very-early-access
>> releases with JDK6 and the Peabody Project, and early exposure was a
>> purpose
>> of the <project-name-never-to-be-spoken-again> Regressions Contest
>> which I
>> ran in early 2006. (See my java.net blog posting of Jan 30, 2006)  I'm
>> sure
>> you can understand the value, right?
>>
>> There would also be value to the OpenJDK project for reference OpenJDK
>> builds to be available.  For example to help those like you who are
>> involved
>> with packaging OpenJDK-derived builds.  Anybody could do those builds
>> couldn't they?
>>
>> I don't think it's correct to say Sun is "pushing proprietary
>> derivatives as
>> early access OpenJDK builds.." is it?  The name JDK7 is distinguished
>> from
>> OpenJDK7, right?  Isn't it well known that they are approximately 96% the
>> same and that there are differences in specific areas?
>
> As an interested observer and fan of open and even Free(tm) Java, I need
> to ask why would you want to have this differentiation?
>
> I can understand the need to provide source and/or binaries to
> commercial partners and customers under licenses that aren't the GPL,
> but given your right to relicense the whole thing, the same code should
> be able to be offered under the GPL...
>
> geir
>
>
>>
>>
>> - David Herron
>>
>> On Sat, May 30, 2009 at 12:31 PM, Mark Wielaard <mark@...> wrote:
>>
>>> On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
>>>> I agree wholeheartedly, but have to say I long ago ceased to be
>>>> surprised by Sun builds beinge proprietary. Sadly the converse is
>>>> true; I'd be surprised by a Sun build released under the same terms as
>>>> our IcedTea builds.
>>>
>>> And that is indeed what is sad about this. That it seems OpenJDK builds
>>> are actually Sun builds, and by extension such things are proprietary.
>>> And that is what I object to. OpenJDK builds should be just that,
>>> OpenJDK builds distributed under the (GPL) terms everybody in our
>>> community adheres to.
>>>
>>> If a project wants to publish "early access" builds then they really
>>> should if they feel people would like to play with the bits. But such
>>> builds should follow the standard OpenJDK project rules
>>> (http://openjdk.java.net/legal/) that everybody else also uses.
>>>
>>> Going to Sun legal and requesting alternative proprietary terms and then
>>> publishing the code and binaries under non-free software licenses is
>>> just bad for creating a community. It is bad enough that the current SCA
>>> rules around OpenJDK assign all rights to one commercial party, Sun. But
>>> projects then abusing those rights by pushing proprietary derivatives as
>>> early access OpenJDK project builds undermines the whole community of
>>> equals.
>>>
>>> You are right that we have IcedTea to fix that. If you get your packages
>>> through IcedTea (derivatives) you are guaranteed that it truly is Free
>>> Software. But wouldn't it be better if we could say that about OpenJDK
>>> itself? Wouldn't that make the community stronger?
>>>
>>> Cheers,
>>>
>>> Mark
>>>
>>>
>

[Geir Magnusson Jr.-3]

On May 30, 2009, at 7:20 PM, Dmitri Trembovetski wrote:

>
>
> Geir Magnusson Jr. wrote:
>> Hi David,
>> On May 30, 2009, at 7:05 PM, David Herron wrote:
>>> Mark,
>>>
>>> Please recall that JDK<n> != OpenJDK<n> though for values of n >=  
>>> 7 the
>>> difference is very small.  The JDK7 builds have some proprietary  
>>> bits in
>>> them.
>> Why?  For heaven's sake... why?
>
>  Because the corresponding open source parts aren't good enough yet  
> and we don't have enough resources to make them on par with the  
> proprietary bits although this is what we want in the long run.
>
>  Specific parts that I know of are color management, AA shape  
> rasterizer and font rasterizer.

It's been how many years that you've had to re-write?

>
>
>  You must understand that "passing the TCK" doesn't necessarily mean  
> "has acceptable performance, fidelity and stability".

Oh, I understand that.  Of course, I'm still in the "getting the TCK"  
phase...

     http://www.apache.org/jcp/sunopenletter.html

;)

geir

>
>
>  Thanks,
>    Dmitri
>
>>>
>>>
>>> It's valuable to the JDK product cycle for JDK builds to have  
>>> early access
>>> exposure so people can report bugs etc.  Sun started doing very-
>>> early-access
>>> releases with JDK6 and the Peabody Project, and early exposure was  
>>> a purpose
>>> of the <project-name-never-to-be-spoken-again> Regressions Contest  
>>> which I
>>> ran in early 2006. (See my java.net blog posting of Jan 30, 2006)  
>>> I'm sure
>>> you can understand the value, right?
>>>
>>> There would also be value to the OpenJDK project for reference  
>>> OpenJDK
>>> builds to be available.  For example to help those like you who  
>>> are involved
>>> with packaging OpenJDK-derived builds.  Anybody could do those  
>>> builds
>>> couldn't they?
>>>
>>> I don't think it's correct to say Sun is "pushing proprietary  
>>> derivatives as
>>> early access OpenJDK builds.." is it?  The name JDK7 is  
>>> distinguished from
>>> OpenJDK7, right?  Isn't it well known that they are approximately  
>>> 96% the
>>> same and that there are differences in specific areas?
>> As an interested observer and fan of open and even Free(tm) Java, I  
>> need to ask why would you want to have this differentiation?
>> I can understand the need to provide source and/or binaries to  
>> commercial partners and customers under licenses that aren't the  
>> GPL, but given your right to relicense the whole thing, the same  
>> code should be able to be offered under the GPL...
>> geir
>>>
>>>
>>> - David Herron
>>>
>>> On Sat, May 30, 2009 at 12:31 PM, Mark Wielaard <mark@...>  
>>> wrote:
>>>
>>>> On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
>>>>> I agree wholeheartedly, but have to say I long ago ceased to be
>>>>> surprised by Sun builds beinge proprietary. Sadly the converse is
>>>>> true; I'd be surprised by a Sun build released under the same  
>>>>> terms as
>>>>> our IcedTea builds.
>>>>
>>>> And that is indeed what is sad about this. That it seems OpenJDK  
>>>> builds
>>>> are actually Sun builds, and by extension such things are  
>>>> proprietary.
>>>> And that is what I object to. OpenJDK builds should be just that,
>>>> OpenJDK builds distributed under the (GPL) terms everybody in our
>>>> community adheres to.
>>>>
>>>> If a project wants to publish "early access" builds then they  
>>>> really
>>>> should if they feel people would like to play with the bits. But  
>>>> such
>>>> builds should follow the standard OpenJDK project rules
>>>> (http://openjdk.java.net/legal/) that everybody else also uses.
>>>>
>>>> Going to Sun legal and requesting alternative proprietary terms  
>>>> and then
>>>> publishing the code and binaries under non-free software licenses  
>>>> is
>>>> just bad for creating a community. It is bad enough that the  
>>>> current SCA
>>>> rules around OpenJDK assign all rights to one commercial party,  
>>>> Sun. But
>>>> projects then abusing those rights by pushing proprietary  
>>>> derivatives as
>>>> early access OpenJDK project builds undermines the whole  
>>>> community of
>>>> equals.
>>>>
>>>> You are right that we have IcedTea to fix that. If you get your  
>>>> packages
>>>> through IcedTea (derivatives) you are guaranteed that it truly is  
>>>> Free
>>>> Software. But wouldn't it be better if we could say that about  
>>>> OpenJDK
>>>> itself? Wouldn't that make the community stronger?
>>>>
>>>> Cheers,
>>>>
>>>> Mark
>>>>
>>>>

[gnu_andrew]

2009/5/31 Geir Magnusson Jr. <geir@...>:

>
> On May 30, 2009, at 7:20 PM, Dmitri Trembovetski wrote:
>
>>
>>
>> Geir Magnusson Jr. wrote:
>>>
>>> Hi David,
>>> On May 30, 2009, at 7:05 PM, David Herron wrote:
>>>>
>>>> Mark,
>>>>
>>>> Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7 the
>>>> difference is very small.  The JDK7 builds have some proprietary bits in
>>>> them.
>>>
>>> Why?  For heaven's sake... why?
>>
>>  Because the corresponding open source parts aren't good enough yet and we
>> don't have enough resources to make them on par with the proprietary bits
>> although this is what we want in the long run.
>>
>>  Specific parts that I know of are color management, AA shape rasterizer
>> and font rasterizer.
>
> It's been how many years that you've had to re-write?
>

>From my perspective at least, if Sun aren't shipping it, they have no
motivation to fix it.  If they were shipping OpenJDK builds and
getting bugs filed against them, then the priority of fixing these
issues would be much higher.  As is, the next JDK release or previous
u bump is always going to take precedence and it will simply never get
done.

As a first step, it at least needs to be narrowed down to specific
issues.  Just saying 'it isn't as good' is of no use to anyone wanting
to work on this.  This same little quote has been thrown about for two
years now and nothing has changed much in the last year to suggest
we're going to see any progress ever.

>>
>>
>>  You must understand that "passing the TCK" doesn't necessarily mean "has
>> acceptable performance, fidelity and stability".
>
> Oh, I understand that.  Of course, I'm still in the "getting the TCK"
> phase...
>
>    http://www.apache.org/jcp/sunopenletter.html
>
> ;)
>
> geir
>
>>
>>
>>  Thanks,
>>   Dmitri
>>
>>>>
>>>>
>>>> It's valuable to the JDK product cycle for JDK builds to have early
>>>> access
>>>> exposure so people can report bugs etc.  Sun started doing
>>>> very-early-access
>>>> releases with JDK6 and the Peabody Project, and early exposure was a
>>>> purpose
>>>> of the <project-name-never-to-be-spoken-again> Regressions Contest which
>>>> I
>>>> ran in early 2006. (See my java.net blog posting of Jan 30, 2006)  I'm
>>>> sure
>>>> you can understand the value, right?
>>>>
>>>> There would also be value to the OpenJDK project for reference OpenJDK
>>>> builds to be available.  For example to help those like you who are
>>>> involved
>>>> with packaging OpenJDK-derived builds.  Anybody could do those builds
>>>> couldn't they?
>>>>
>>>> I don't think it's correct to say Sun is "pushing proprietary
>>>> derivatives as
>>>> early access OpenJDK builds.." is it?  The name JDK7 is distinguished
>>>> from
>>>> OpenJDK7, right?  Isn't it well known that they are approximately 96%
>>>> the
>>>> same and that there are differences in specific areas?
>>>
>>> As an interested observer and fan of open and even Free(tm) Java, I need
>>> to ask why would you want to have this differentiation?
>>> I can understand the need to provide source and/or binaries to commercial
>>> partners and customers under licenses that aren't the GPL, but given your
>>> right to relicense the whole thing, the same code should be able to be
>>> offered under the GPL...
>>> geir
>>>>
>>>>
>>>> - David Herron
>>>>
>>>> On Sat, May 30, 2009 at 12:31 PM, Mark Wielaard <mark@...> wrote:
>>>>
>>>>> On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
>>>>>>
>>>>>> I agree wholeheartedly, but have to say I long ago ceased to be
>>>>>> surprised by Sun builds beinge proprietary. Sadly the converse is
>>>>>> true; I'd be surprised by a Sun build released under the same terms as
>>>>>> our IcedTea builds.
>>>>>
>>>>> And that is indeed what is sad about this. That it seems OpenJDK builds
>>>>> are actually Sun builds, and by extension such things are proprietary.
>>>>> And that is what I object to. OpenJDK builds should be just that,
>>>>> OpenJDK builds distributed under the (GPL) terms everybody in our
>>>>> community adheres to.
>>>>>
>>>>> If a project wants to publish "early access" builds then they really
>>>>> should if they feel people would like to play with the bits. But such
>>>>> builds should follow the standard OpenJDK project rules
>>>>> (http://openjdk.java.net/legal/) that everybody else also uses.
>>>>>
>>>>> Going to Sun legal and requesting alternative proprietary terms and
>>>>> then
>>>>> publishing the code and binaries under non-free software licenses is
>>>>> just bad for creating a community. It is bad enough that the current
>>>>> SCA
>>>>> rules around OpenJDK assign all rights to one commercial party, Sun.
>>>>> But
>>>>> projects then abusing those rights by pushing proprietary derivatives
>>>>> as
>>>>> early access OpenJDK project builds undermines the whole community of
>>>>> equals.
>>>>>
>>>>> You are right that we have IcedTea to fix that. If you get your
>>>>> packages
>>>>> through IcedTea (derivatives) you are guaranteed that it truly is Free
>>>>> Software. But wouldn't it be better if we could say that about OpenJDK
>>>>> itself? Wouldn't that make the community stronger?
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Mark
>>>>>
>>>>>
>
>

--
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8

[Mark Wielaard]

On Sat, 2009-05-30 at 16:05 -0700, David Herron wrote:
> Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7
> the difference is very small.  The JDK7 builds have some proprietary
> bits in them.  

Sure, that is why we have the Assembly Exception, and the Designated
Exception Modules, in case someone really needs to make such a release.
http://openjdk.java.net/legal/

> It's valuable to the JDK product cycle for JDK builds to have early
> access exposure so people can report bugs etc.  Sun started doing
> very-early-access releases with JDK6 and the Peabody Project, and
> early exposure was a purpose of the
> <project-name-never-to-be-spoken-again> Regressions Contest which I
> ran in early 2006. (See my java.net blog posting of Jan 30, 2006)  I'm
> sure you can understand the value, right?

Sure. If done under a license that would actually allow sharing of
information and help the OpenJDK community and not just Sun.

> There would also be value to the OpenJDK project for reference OpenJDK
> builds to be available.  For example to help those like you who are
> involved with packaging OpenJDK-derived builds.  Anybody could do
> those builds couldn't they?

Yes, and most distros now ship those bits. Maybe we should publicize
that better. It is just that some OpenJDK projects suddenly turn into
Sun builds for their early access builds that I think is not done.

Especially not since the special license on them is much more draconian
than most proprietary licenses I have ever seen. It prevents anybody
from sharing any information or use them for OpenJDK collaboration
(because that would be Confidential Information that would Irreparable
Harm Sun if released).

Cheers,

Mark

[Mark Wielaard]

Hi Dmitri,

On Sat, 2009-05-30 at 16:20 -0700, Dmitri Trembovetski wrote:

> Geir Magnusson Jr. wrote:
> > On May 30, 2009, at 7:05 PM, David Herron wrote:
> >> Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7 the
> >> difference is very small.  The JDK7 builds have some proprietary bits in
> >> them.
> >
> > Why?  For heaven's sake... why?
>
>    Because the corresponding open source parts aren't good enough yet and we
> don't have enough resources to make them on par with the proprietary bits
> although this is what we want in the long run.
>
>    Specific parts that I know of are color management, AA shape rasterizer and
> font rasterizer.

Sure, I might not like that there are still proprietary bits around even
though we have free replacements, but I understand that they are still
around. That doesn't mean you need to use a different license from the
rest of us. http://openjdk.java.net/legal/ clearly spells out the rules.
You may ship these proprietary blobs together with the larger GPLed
parts thanks to the Assembly Exception. That means you can just ship
everything under free terms except for these remaining blobs if you
really want to. No need to get some special terms different from the
rest of the community.

Also I agree with Andrew, if you keep only shipping the proprietary
blobs and not the free replacements you will have less incentive to help
us all improve the free bits.

Lastly this is not really relevant for most of the OpenJDK stack except
some parts of the graphics stack, why would the nio2 project for
example, ship early access builds under a completely proprietary
license?

Cheers,

Mark

[Andrew Haley]

Dmitri Trembovetski wrote:

>
>
> Geir Magnusson Jr. wrote:
>> Hi David,
>>
>> On May 30, 2009, at 7:05 PM, David Herron wrote:
>>
>>> Mark,
>>>
>>> Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7 the
>>> difference is very small.  The JDK7 builds have some proprietary bits in
>>> them.
>>
>> Why?  For heaven's sake... why?
>
>   Because the corresponding open source parts aren't good enough yet and
> we don't have enough resources to make them on par with the proprietary
> bits although this is what we want in the long run.
>
>   Specific parts that I know of are color management, AA shape
> rasterizer and font rasterizer.

Understood, but that doesn't apply in all the cases that Mark is talking
about, nio2 being a good example.

Andrew.

[kelly.ohair]

Geir Magnusson Jr. wrote:

>
> On May 30, 2009, at 7:20 PM, Dmitri Trembovetski wrote:
>
>>
>>
>> Geir Magnusson Jr. wrote:
>>> Hi David,
>>> On May 30, 2009, at 7:05 PM, David Herron wrote:
>>>> Mark,
>>>>
>>>> Please recall that JDK<n> != OpenJDK<n> though for values of n >= 7 the
>>>> difference is very small.  The JDK7 builds have some proprietary
>>>> bits in
>>>> them.
>>> Why?  For heaven's sake... why?
>>
>>  Because the corresponding open source parts aren't good enough yet
>> and we don't have enough resources to make them on par with the
>> proprietary bits although this is what we want in the long run.
>>
>>  Specific parts that I know of are color management, AA shape
>> rasterizer and font rasterizer.
>
> It's been how many years that you've had to re-write?

                                 ^^^^^^
                                 we have

Seems like I am reading too much "them vs. us" in these emails.

-kto

>
>>
>>
>>  You must understand that "passing the TCK" doesn't necessarily mean
>> "has acceptable performance, fidelity and stability".
>
> Oh, I understand that.  Of course, I'm still in the "getting the TCK"
> phase...
>
>     http://www.apache.org/jcp/sunopenletter.html
>
> ;)
>
> geir
>
>>
>>
>>  Thanks,
>>    Dmitri
>>
>>>>
>>>>
>>>> It's valuable to the JDK product cycle for JDK builds to have early
>>>> access
>>>> exposure so people can report bugs etc.  Sun started doing
>>>> very-early-access
>>>> releases with JDK6 and the Peabody Project, and early exposure was a
>>>> purpose
>>>> of the <project-name-never-to-be-spoken-again> Regressions Contest
>>>> which I
>>>> ran in early 2006. (See my java.net blog posting of Jan 30, 2006)  
>>>> I'm sure
>>>> you can understand the value, right?
>>>>
>>>> There would also be value to the OpenJDK project for reference OpenJDK
>>>> builds to be available.  For example to help those like you who are
>>>> involved
>>>> with packaging OpenJDK-derived builds.  Anybody could do those builds
>>>> couldn't they?
>>>>
>>>> I don't think it's correct to say Sun is "pushing proprietary
>>>> derivatives as
>>>> early access OpenJDK builds.." is it?  The name JDK7 is
>>>> distinguished from
>>>> OpenJDK7, right?  Isn't it well known that they are approximately
>>>> 96% the
>>>> same and that there are differences in specific areas?
>>> As an interested observer and fan of open and even Free(tm) Java, I
>>> need to ask why would you want to have this differentiation?
>>> I can understand the need to provide source and/or binaries to
>>> commercial partners and customers under licenses that aren't the GPL,
>>> but given your right to relicense the whole thing, the same code
>>> should be able to be offered under the GPL...
>>> geir
>>>>
>>>>
>>>> - David Herron
>>>>
>>>> On Sat, May 30, 2009 at 12:31 PM, Mark Wielaard <mark@...> wrote:
>>>>
>>>>> On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
>>>>>> I agree wholeheartedly, but have to say I long ago ceased to be
>>>>>> surprised by Sun builds beinge proprietary. Sadly the converse is
>>>>>> true; I'd be surprised by a Sun build released under the same
>>>>>> terms as
>>>>>> our IcedTea builds.
>>>>>
>>>>> And that is indeed what is sad about this. That it seems OpenJDK
>>>>> builds
>>>>> are actually Sun builds, and by extension such things are proprietary.
>>>>> And that is what I object to. OpenJDK builds should be just that,
>>>>> OpenJDK builds distributed under the (GPL) terms everybody in our
>>>>> community adheres to.
>>>>>
>>>>> If a project wants to publish "early access" builds then they really
>>>>> should if they feel people would like to play with the bits. But such
>>>>> builds should follow the standard OpenJDK project rules
>>>>> (http://openjdk.java.net/legal/) that everybody else also uses.
>>>>>
>>>>> Going to Sun legal and requesting alternative proprietary terms and
>>>>> then
>>>>> publishing the code and binaries under non-free software licenses is
>>>>> just bad for creating a community. It is bad enough that the
>>>>> current SCA
>>>>> rules around OpenJDK assign all rights to one commercial party,
>>>>> Sun. But
>>>>> projects then abusing those rights by pushing proprietary
>>>>> derivatives as
>>>>> early access OpenJDK project builds undermines the whole community of
>>>>> equals.
>>>>>
>>>>> You are right that we have IcedTea to fix that. If you get your
>>>>> packages
>>>>> through IcedTea (derivatives) you are guaranteed that it truly is Free
>>>>> Software. But wouldn't it be better if we could say that about OpenJDK
>>>>> itself? Wouldn't that make the community stronger?
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Mark
>>>>>
>>>>>
>

[Geir Magnusson Jr.-3]

On May 31, 2009, at 1:55 PM, Kelly O'Hair wrote:

>
>
> Geir Magnusson Jr. wrote:
>> On May 30, 2009, at 7:20 PM, Dmitri Trembovetski wrote:
>>>
>>>
>>> Geir Magnusson Jr. wrote:
>>>> Hi David,
>>>> On May 30, 2009, at 7:05 PM, David Herron wrote:
>>>>> Mark,
>>>>>
>>>>> Please recall that JDK<n> != OpenJDK<n> though for values of n  
>>>>> >= 7 the
>>>>> difference is very small.  The JDK7 builds have some proprietary  
>>>>> bits in
>>>>> them.
>>>> Why?  For heaven's sake... why?
>>>
>>> Because the corresponding open source parts aren't good enough yet  
>>> and we don't have enough resources to make them on par with the  
>>> proprietary bits although this is what we want in the long run.
>>>
>>> Specific parts that I know of are color management, AA shape  
>>> rasterizer and font rasterizer.
>> It's been how many years that you've had to re-write?
>                                ^^^^^^
>                                we have
>
> Seems like I am reading too much "them vs. us" in these emails.

Oh, come on.  I don't know where to begin here.

1)  I'm not a "you" :)  I'm really happy OpenJDK exists, but as one of  
the founder's of Apache Harmony, I think it's good that there are many  
free/open/libre Java communities.   I'm very interested in floss Java,  
which is why I pay attention to this community.

2) This whole thread is about members of the OpenJDK community  
complaining about *you* publishing proprietary builds.  They don't  
seem to feel like a part of "us".

geir



geir

>
>
> -kto
>
>>>
>>>
>>> You must understand that "passing the TCK" doesn't necessarily  
>>> mean "has acceptable performance, fidelity and stability".
>> Oh, I understand that.  Of course, I'm still in the "getting the  
>> TCK" phase...
>>    http://www.apache.org/jcp/sunopenletter.html
>> ;)
>> geir
>>>
>>>
>>> Thanks,
>>>   Dmitri
>>>
>>>>>
>>>>>
>>>>> It's valuable to the JDK product cycle for JDK builds to have  
>>>>> early access
>>>>> exposure so people can report bugs etc.  Sun started doing very-
>>>>> early-access
>>>>> releases with JDK6 and the Peabody Project, and early exposure  
>>>>> was a purpose
>>>>> of the <project-name-never-to-be-spoken-again> Regressions  
>>>>> Contest which I
>>>>> ran in early 2006. (See my java.net blog posting of Jan 30,  
>>>>> 2006)  I'm sure
>>>>> you can understand the value, right?
>>>>>
>>>>> There would also be value to the OpenJDK project for reference  
>>>>> OpenJDK
>>>>> builds to be available.  For example to help those like you who  
>>>>> are involved
>>>>> with packaging OpenJDK-derived builds.  Anybody could do those  
>>>>> builds
>>>>> couldn't they?
>>>>>
>>>>> I don't think it's correct to say Sun is "pushing proprietary  
>>>>> derivatives as
>>>>> early access OpenJDK builds.." is it?  The name JDK7 is  
>>>>> distinguished from
>>>>> OpenJDK7, right?  Isn't it well known that they are  
>>>>> approximately 96% the
>>>>> same and that there are differences in specific areas?
>>>> As an interested observer and fan of open and even Free(tm) Java,  
>>>> I need to ask why would you want to have this differentiation?
>>>> I can understand the need to provide source and/or binaries to  
>>>> commercial partners and customers under licenses that aren't the  
>>>> GPL, but given your right to relicense the whole thing, the same  
>>>> code should be able to be offered under the GPL...
>>>> geir
>>>>>
>>>>>
>>>>> - David Herron
>>>>>
>>>>> On Sat, May 30, 2009 at 12:31 PM, Mark Wielaard <mark@...>  
>>>>> wrote:
>>>>>
>>>>>> On Fri, 2009-05-29 at 22:10 +0100, Andrew John Hughes wrote:
>>>>>>> I agree wholeheartedly, but have to say I long ago ceased to be
>>>>>>> surprised by Sun builds beinge proprietary. Sadly the converse  
>>>>>>> is
>>>>>>> true; I'd be surprised by a Sun build released under the same  
>>>>>>> terms as
>>>>>>> our IcedTea builds.
>>>>>>
>>>>>> And that is indeed what is sad about this. That it seems  
>>>>>> OpenJDK builds
>>>>>> are actually Sun builds, and by extension such things are  
>>>>>> proprietary.
>>>>>> And that is what I object to. OpenJDK builds should be just that,
>>>>>> OpenJDK builds distributed under the (GPL) terms everybody in our
>>>>>> community adheres to.
>>>>>>
>>>>>> If a project wants to publish "early access" builds then they  
>>>>>> really
>>>>>> should if they feel people would like to play with the bits.  
>>>>>> But such
>>>>>> builds should follow the standard OpenJDK project rules
>>>>>> (http://openjdk.java.net/legal/) that everybody else also uses.
>>>>>>
>>>>>> Going to Sun legal and requesting alternative proprietary terms  
>>>>>> and then
>>>>>> publishing the code and binaries under non-free software  
>>>>>> licenses is
>>>>>> just bad for creating a community. It is bad enough that the  
>>>>>> current SCA
>>>>>> rules around OpenJDK assign all rights to one commercial party,  
>>>>>> Sun. But
>>>>>> projects then abusing those rights by pushing proprietary  
>>>>>> derivatives as
>>>>>> early access OpenJDK project builds undermines the whole  
>>>>>> community of
>>>>>> equals.
>>>>>>
>>>>>> You are right that we have IcedTea to fix that. If you get your  
>>>>>> packages
>>>>>> through IcedTea (derivatives) you are guaranteed that it truly  
>>>>>> is Free
>>>>>> Software. But wouldn't it be better if we could say that about  
>>>>>> OpenJDK
>>>>>> itself? Wouldn't that make the community stronger?
>>>>>>
>>>>>> Cheers,
>>>>>>
>>>>>> Mark
>>>>>>
>>>>>>

[Mark Wielaard]

Hi Kelly,

On Sun, 2009-05-31 at 10:55 -0700, Kelly O'Hair wrote:
> Geir Magnusson Jr. wrote:
> >
> > It's been how many years that you've had to re-write?
>                                  ^^^^^^
>                                  we have
>
> Seems like I am reading too much "them vs. us" in these emails.

Don't feed the trolls :)