Path: utzoo!mnetor!uunet!husc6!cmcl2!rutgers!sdcsvax!ucbvax!ATHENA.MIT.EDU!martillo
From: martillo@ATHENA.MIT.EDU
Newsgroups: comp.protocols.tcp-ip
Subject: Need TCP/IP on RSTS/E
Message-ID: <8801060423.AA24419@ARES.MIT.EDU>
Date: 6 Jan 88 04:23:13 GMT
References: <8801050739.AA01158@GAAK.LCS.MIT.EDU>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 19


While the display code and the network interface code (which live in
their own separate library modules) are rather dependent on
characteristics of the PC and DOS, I would say the rest of the code is
rather machine and OS independent even though small chunks of assembly
code would have to be rewritten for a given target machine.

I thought the PCIP code was developed under a Darpa grant for the DOS
environment but that the starting point was a package which ran under
PDP11 Unix and there is a comment in the tasking package to this effect.

However, once I asked Jerry Saltzer why the event flag was cleared by the
scheduler and not left for the task to clear so that useful information
could be passed via the event flag to the task and he mentioned in reply
that the tasking piece at least was running on Multics in 1966 (I may have
the year off) so that large chunks of the code have run in at least
3 different environments.  I think I saw that Tim Maroney ported the
code to the Macintosh so that chunks of PCIP have run in at least
4 very different environments.