Re: What are tainted developers allowed to work on?

[Dalibor Topic]

Chris Gray wrote:
> On Sunday 09 January 2005 20:25, Dalibor Topic wrote:
>
> Interesting ... I started down the same road myself, but got sidetracked by 
> the usual end-of-year rush (which ain't over yet). Time to take this up 
> again. There's sufficient difference between our two projects that it will 
> probably be interesting to compare the results.

Nice! I expect the test suite setup and similar things to be a pain the 
rear, given that it is a non-free piece of software, and those tend to 
have horrible build systems. At least I've heard war stories about the 
J2EE test suite being labour-intensive to set up. So I guess we can 
share some knowledge there. :)

I guess you'll be going for the 1.5 TCK scholarship as well? Or are 
there other Java-whatever-edition versions that Sun has started removing 
their legal cruft to allow compatible open source implementations? 
Afaik, with the TCKs <1.5 the problem was that Sun tied in a lot of SCSL 
covered code in the TCK package, requiring implementations to propagate 
it, thereby killing any chance of an implementation being compatible 
with their test suite and open source at the same time.

> In the meantime I did manage to elicit this statement from a Sun engineer,
> while talking about something other than JCK:
>
>         "Regarding independant VM's, people are free to use the Java
> Specifications for R&D non-commercial use but as soon as they make use
> of the code for commercial use then they have to take a license from
> Sun.  The so called "CleanRoom" implementations of VM's are in fact not
> clean as they are based on the Java Specification which is Sun's IP.
> Therefore before making use of this commercially they should license
> from Sun the test suites to ensure compliance."

That sounds like an attempt at 'Carrot And Stick'. :)

I doubt that Sun running amok in SCO style would make a lot of sense. 
People wanting their implementations to be compatible with Sun is a 
no-brainer, even without a stick, as otherwise one couldn't run the 
existing code out there.

> > Part of that non-free software legal mess are trade marks, which is why
> > I prefer not to use the term "Free Java" or similar idioms. I'd rather
> > see Kaffe (and other free software runtimes) being referred to as free
> > runtimes, to avoid the connotations that go with the term 'JVM' or 'Java'.
>
>
> Isn't that kind of like referring to gcc as a "free compiler", without naming 
> a language? For those for whom there is only One True Language that might 
> seem unambiguous, but the rest of us might be confused and think you were 
> working on a runtime for Ada or Occam or something.

I'll add those to my list of code to merge into Kaffe, thanks :)

> Best wishes for 2005,

All the best to you, too!

cheers,
dalibor topic