🚀 go-pugleaf

csa@sw.stratus.com (Chris Arthur) writes:
> In article <1991Aug30.195229.4202@craycos.com>, jrbd@craycos.com (James Davies) writes:

> > For example, IBM has a patent
> > on register assignment via graph coloring -- how in the world
> > could an outsider detect that a compiler was using this?
> 
> Based on what I recall happening here at Stratus, that's not the way
> they go about it; basically, IBM said "Hey, you!  Prove that your compiler
> doesn't use our patented technique!" and one of our senior compiler
> engineers was forced to sit down with some IBM $%*&!@ and go through
> how our compiler does register allocation.  Nice, huh?  So much for
> innocent until proven guilty.

Now, just a moment: that means that IBM can have a look at the proprietary
internals of any vendor's compiler technology so that they can check that
the vendors don't use an IBM patent? This looks crass.

Of course if IBM sues then they get to see the internals of the compiler
anyhow thanks to discovery, but odds are that the judge will ask IBM to
prove that there is reasonable suspicion. Well, I hope.
-- 
Piercarlo Grandi                   | ARPA: pcg%uk.ac.aber@nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth        | UUCP: ...!mcsun!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@aber.ac.uk