On Wed, 17 Feb 1999, Michael Emmel wrote:
> > Ah... just being safe. There is nothing in the license agreement
> > that I can see that stops someone who signed it from doing a
> > clean room implementation as long as the Sun source is not
> > referenced. Correct?
> >
> > ..darcy
> >
>
> I recently posted a suggestion to javalobby that Sun make periodice releases
> of thre source
> "free" to clean room implementation. This would require that toe clean romm
> implemenetation must
> pass the JCK but remove gratious imcompatibilities caues by inconsitten
> interpetation of the specicifation.
> Thi implemnetaton may be a green threads full interpeter imp but it must
> implement the spec.
> I think this is a exclelent Idea. Sun distributes a correct but slow refrence
> imp.
> Everybody else write better VM. This is esp important for the Class libs.
> Classpath is a good project but the spece is not enough for compatibility
> hence the JCK ???
> Either the spec is enough and the JCK is not needed or Sun needs to allow a
> path for compativble clena room imps.
> Class libs should be enough imho.
>
> Mike
That would be nice, but it does not look like Sun is going to release the
JCK for testing clean room implementations. By adding terms to the SCSL
that require passing the JCK and then not providing the JCK without
payment to Sun, they are making the source "not free". There is just
no way around this. Sun needs to license it's implementation under
a free license or keep it to themselves. This "we will give you half"
deal is not good for anyone in the long run (except Sun, as they stand
to make lots of money off testing fees and JVM implementors will
have to follow whatever Sun says is "java").
Mo DeJong
dejong at cs.umn.edu
