Steve Langasek wrote:
> On Mon, Jun 05, 2006 at 01:46:29AM -0700, Walter Landry wrote:
>> Steve Langasek <[EMAIL PROTECTED]> wrote:
>>> On Sun, Jun 04, 2006 at 08:16:24AM -0700, Walter Landry wrote:
>>>> This can be fixed by making Jython conflict with sun-java, although I
>>>> suspect that there may be other packages that also cause problems.
> 
>> Another fix for this particular problem is to make sun-java not
>> Provide:java-runtime etc. at all.

Good point.  That would tend to decrease the utility of sun-java, but it
should avoid most of the issues.

>>> There have been various clarifications that the intent of this clause is to
>>> prohibit distributing sun-java5 in a configuration that mixes and matches
>>> parts of Sun's implementation with other Java implementations.

I don't think any of the clarifications have limited the clause to that
sole purpose, and furthermore I don't think even that much of a
clarification addresses the question of what constitutes such a
configuration.  See my message referenced at the end of this mail for
examples that seem to me to fall under such a description.

>>> Why are these clarifications not sufficient when we regularly accept
>>> clarifications of this nature from other copyright holders?
> 
>> There are a few problems here:
> 
>> 1) Clarifications from other copyright holders do not come with a big
>>    disclaimer stating that the clarification has no legal weight.
>>    Debian relies on the clarification having legal weight.
> 
> Well, this seems to have just changed in the case of the DLJ.

I agree; the FAQ now seems sufficiently binding to meet the standards of
other license clarifications Debian has accepted, in my opinion.  I also
think the answers in the FAQ address many of the other concerns raised
by debian-legal: the issue of whether distributing in non-free (which
Debian does not consider part of the distribution) counts as
distributing with the OS, the issue of whether users can develop
software for non-Debian platforms using the JDK (not a freeness issue
but an important question), and the issue of whether removal affects
mirrors or stable releases.

>> 3) There are only two clarifications I know of.  One is in the
>>    non-binding DLJ FAQ, and the other is from Tom Marble [1].  Neither
>>    of these clarifications limit the restrictions to Java
>>    implementations.  For example, in Tom's clarification he says [2]
> 
>>      In a similar way please don't take bits from the Java platform
>>      and use them as part of or to complete alternate technologies
>>      (e.g. plugin.jar).
> 
>>   So using Java's plugin mechanism to implement python plugins for
>>   Jython is not allowed.  And this makes sense.  Sun would definitely
>>   be unhappy if the JDK were used to implement J++ or C#, especially
>>   because they are not Java.
> 
> That's not actually clear to me; "alternate technologies" is very vague.
> I'm also not convinced that dropping the java-runtime provide from the Sun
> Java packages is the sort of thing Sun intends with this, either. 
> Basically, it seems to me that the cited clarification from Tom happens to
> not clarify much after all...

Same with the clarification in the FAQ.  It seems to answer some
questions (we don't need to worry about non-Java software qualifying as
"similar functionality"), but not others; see my message referenced
below for examples of what I don't think it addresses.

> BTW, if jython is the only package that poses a problem for this clause, I'm
> not sure there's much to worry about anyway here: jython is RC-buggy and
> likely to be dropped from etch, because it build-depends on (and implements)
> python2.1, which is 3 stable releases behind where we expect the main python
> implementation to be for etch.

Other packages raise similar concerns; please see my message
<[EMAIL PROTECTED]>, available at
<http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=370295;msg=10>, for
other examples, as well as suggested fixes for the DLJ (or the FAQ now
that it looks like a sufficiently binding clarification) which would
address those concerns.

The bug probably needs a new title. :)

I see no problem with the idea of not letting other Java-related
software use pieces of Sun Java.  My primary concern lies in avoiding
any possibility that a package in main will need to change to satisfy
the DLJ.

- Josh Triplett

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to