Network Working Group                                            M. Wilber
Request for Comments #152                                        10 May 71
NIC #6756                                                        SRAI
Category: G.3
Obsoletes: None
Updates: None

Response to RFC #116


                SRI ARTIFICIAL INTELLIGENCE STATUS REPORT
	

     The Stanford Research Institute's Artificial Intelligence Group (SRAI
in the four-letter abbreviations) expects connection to the ARPA net as a
research center after conversion this summer to a TENEX from our current
PDP-10. Our connection will be through the IMP already at SRI for the
Network Information Center and through a PDP-15 serving our PDP-10 as a
controller of peripherals. Our hardware interface to the IMP is currently
in the bidding process, and we intend to use as much as possible of the
TENEX network software. (Probably all we will need to change is the lowest
level, most strongly hardware-oriented part of the NCP.)

     The most optimistic estimate we can give for functional connection to
the network is mid-July 1971. We are currently devoting the energies of
our system support group to the accomodation of various hardware and design
changes, and so our contact with the Network Working Group has been only
minimal and passive. It is entirely conceivable that we may find our par-
ticipation tending to strength and activeness as we cross our other bridges.

     We can project our participation in the network into the first few
months of our connection. We can support several simultaneous outside
users on a system nominally up during the business day and often up at
other times. Lapses in continuity of system operation are usually due to
scheduled maintenance or hardware failures and tend to occur at intervals
of either an hour or several days with remarkable consistency. The prin-
cipal service we offer to other network participants is the availability
of various parts of our own research software. The most notable examples
are QA3.6, a first-order resolution theorem prover;  STRIPS, an extra-
logical problem solver;  and possibly QA4, a language oriented toward
problem-solving strategies. The services we can anticipate requesting
of the network are of two kinds: We could conceivably use other people's
artificial intelligence programs on a trial basis; and we might use the
network to make occasional contact with other people in the network.

        [ This RFC was put into machine readable form for entry ]
        [ into the online RFC archives by BBN Corp. under the   ]
        [ direction of Alex McKenzie.                   12/96   ]