25 Jan
2021
25 Jan
'21
9:55 p.m.
A couple of updates.
At a Usenix (Portland?) dmr had a keynote and he interviewed himself.Q:
"What do you think of X?" A: "Sometimes when you fill a vacuum it still
sucks." Maybe Rob said it earlier, but that was the first time I heard it
publically.
Osterhout's Tk was beyond amazing. I was at OSF and gave several demo's
(the "windowing ksh" was also available). The idea that you could
"send"
to another GUI program and add buttons, etc., instantaneously! John had
already implemented the Motif look and feel (from the spec), mostly, and
was willing to do anything if OSF would take it and call it Motif. The
Motif leads (I was in DCE) weren't interested, although maybe they'd put it
on the "extra's" tape at the end. This was circa 1992. I once told this
story to some Microsoft PM's, and they agreed it would have completely
killed Visual Basic. Ah well.
Apollo's had two 68K processors, the second one watched for page faults and
patched things up since it wasn't until 68020 that the faulting stuff
worked properly. The Apollo source control team left after HP and formed
Clearcase.
DCE RPC was based on Apollo NCS which was a very elegant RPC system built
on UDP. It had no XDR because it was "reader makes it right" and datatypes
were tagged. I don't recall details of the tagging. Digital added TCP
transport. Microsoft took the DCE RPC spec (we had a name for them, I
forget what it was) and used it to implement DCOM's RPC. There's an IETF
draft, https://tools.ietf.org/html/draft-brown-dcom-v1-spec-01, but it
never progressed beyond that.
Hope this is interesting to some.