[TUHS] Re-implementations/Clean-Rooms et al.

Larry McVoy lm at mcvoy.com
Fri Sep 9 10:17:50 AEST 2022


On Thu, Sep 08, 2022 at 08:05:44PM -0400, Steve Nickolas wrote:
> On Thu, 8 Sep 2022, Warner Losh wrote:
> 
> >On Thu, Sep 8, 2022 at 5:29 PM Steve Nickolas <usotsuki at buric.co> wrote:
> >
> >>On Thu, 8 Sep 2022, Warner Losh wrote:
> >>
> >>>But it likely didn't matter, since 32v likely lost its copyright
> >>>protection due to AT&T distributing too many copies without the required
> >>>copyright markings. At least that was the preliminary ruling that caused
> >>>the suit to be settled... AT&T didn't want it finalized, though the cat
> >>>was somewhat out of the bag at this point...
> >>
> >>It would be nice if that were an absolute rather than a probably, because
> >>then the status for 32V wouldn't be clouded.
> >>
> >
> >It would be nice. At this late date, one wonders what would happen if it
> >were litigated again...  I suspect that nobody would bother given the
> >small possible gain and the huge expense... But it would also reduce
> >shareholder values to explicitly say there's no copyright here or to
> >clarify that the ancient licenses are valid. So we're in this state where
> >it's basically free and clear, treated like it's free and clear, but
> >really isn't free and clear.
> >
> >Warner
> 
> I'm probably the only one brazen enough to put it to the test.
> 
> For some years, I've wanted to create a free implementation of System V, and
> then move on from there.  (I know there's limited utility for such a thing,
> because of the BSDs.)

Why?  Have you booted 32V?  Run in it for a while?  No VM, no networking,
very basic system.  Other than historical, I don't understand the point.

> A few things actually hinge on this.  If it were considered a fact, and not
> a mere opinion, that 32V was PD, then I could be sure that certain things
> were safe to use, rather than having to rewrite (including some particularly
> tricky stuff the BSDs never fully reimplemented, like diff(1)).

I'm a source management guy, I've written a couple of systems.  I live and
breath diff and diff(1) is not in the slightest way hard.  I wrote my own
version of SCCS in a way that you could get as many different versions of
the history as you wanted in one pass.  That's a lot harder than diff(1).

But maybe I don't understand what you think is tricky about diff, you 
may have some insight I'm missing, care to share?


More information about the TUHS mailing list