perm filename DIALNE.XGP[DIA,JMC] blob
sn#533498 filedate 1980-09-03 generic text, type T, neo UTF8
/LMAR=0/XLINE=3/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#3=GACB25/FONT#4=SUP/FONT#5=NGR20/FONT#6=METLB/FONT#7=MS25/FONT#8=SAIL25/FONT#9=SIGN57
␈↓ ↓H␈↓ ␈↓ ¬≡DIALNET
␈↓ ↓H␈↓ ␈↓ ∧ ␈↓εA Computer Communication Study
␈↓ ↓H␈↓ε␈↓ ¬=␈↓αproposal submitted to
␈↓ ↓H␈↓ε␈↓ ∧WNational Science Foundation
␈↓ ↓H␈↓α␈↓ ¬0Washington, D.C. 20550
␈↓ ↓H␈↓α␈↓ ε?by
␈↓ ↓H␈↓α␈↓ ∧1Stanford Arti≡cial Intelligence Laboratory
␈↓ ↓H␈↓α␈↓ ¬≡␈↓εStanford University
␈↓ ↓H␈↓ε␈↓ ¬≥␈↓αStanford, California 94305
␈↓ ↓H␈↓α␈↓ ¬Z␈↓εOctober 1976
␈↓ ↓H␈↓α␈↓ β_Research Proposal Submitted to the National Science Foundation
␈↓ ↓H␈↓Proposed Amount ␈↓&␈↓λ$95,996␈↓␈↓)αβ Proposed E≥ective Date ␈↓&␈↓λ1 July 1977␈↓␈↓)αβ Proposed Duration ␈↓&␈↓λ18 month␈↓␈↓)αβ
␈↓ ↓H␈↓Title ␈↓&␈↓λDialnet: a Computer Communication Study␈↓␈↓)αβ
␈↓ ↓H␈↓Principal Investigator:␈↓ ¬hSubmitting Institution:
␈↓ ↓H␈↓ ␈↓&␈↓λProfessor John McCarthy␈↓␈↓)αβ␈↓ ¬h ␈↓&␈↓λStanford University␈↓␈↓)αβ
␈↓ ↓H␈↓ Soc. Sec. No. ␈↓&␈↓λ558-30-4793␈↓␈↓)αβ␈↓ ¬h Department ␈↓&␈↓λComputer Science Department␈↓␈↓)αβ
␈↓ ↓H␈↓␈↓ ¬h Branch ␈↓&␈↓λSchool of Humanities and Sciences␈↓␈↓)αβ
␈↓ ↓H␈↓Make grant to ␈↓&␈↓λ Leland Stanford Junior University ␈↓␈↓)αβ
␈↓ ↓H␈↓Endorsements:
␈↓ ↓H␈↓␈↓ αXPrincipal Investigator␈↓ ¬hDepartment Head␈↓ λxInstitutional Admin. O≠ci
␈↓ ↓H␈↓Name␈↓ αX␈↓&␈↓λJohn McCarthy ␈↓␈↓)αβ␈↓ ¬h␈↓&␈↓λEdward A. Feigenbaum ␈↓␈↓)αβ␈↓ λx␈↓&␈↓λ ␈↓␈↓)αβ
␈↓ ↓H␈↓Signature␈↓ αX␈↓&␈↓λ ␈↓␈↓)αβ␈↓ ¬h␈↓&␈↓λ ␈↓␈↓)αβ␈↓ λx␈↓&␈↓λ ␈↓␈↓)αβ
␈↓ ↓H␈↓Title␈↓ αX␈↓&␈↓λ Professor ␈↓␈↓)αβ␈↓ ¬h␈↓&␈↓λProfessor & Chairman ␈↓␈↓)αβ␈↓ λx␈↓&␈↓λ ␈↓␈↓)αβ
␈↓ ↓H␈↓Telephone␈↓ αX␈↓&␈↓λ(415) 497-4430 ␈↓␈↓)αβ␈↓ ¬h␈↓&␈↓λ(415) 497-4878 ␈↓␈↓)αβ␈↓ λx␈↓&␈↓λ ␈↓␈↓)αβ
␈↓ ↓H␈↓Date␈↓ αX␈↓&␈↓λ ␈↓␈↓)αβ␈↓ ¬h␈↓&␈↓λ ␈↓␈↓)αβ␈↓ λx␈↓&␈↓λ ␈↓␈↓)αβ
␈↓ ↓H␈↓α␈↓ ¬VTable of Contents␈↓ Ni
␈↓ ↓H␈↓α␈↓ αλSection␈↓ ¬wPage
␈↓ ↓H␈↓1. Purpose␈↓ ε∃ 2
␈↓ ↓H␈↓2. Scenario␈↓ ε∃ 3
␈↓ ↓H␈↓3. Protocols␈↓ ε∃ 4
␈↓ ↓H␈↓4. Research Issues␈↓ ε∃ 4
␈↓ ↓H␈↓5. Research Plan␈↓ ε∃ 5
␈↓ ↓H␈↓␈↓ β3␈↓αAppendices␈↓
␈↓ ↓H␈↓A. Personnel␈↓ ε∃ 5
␈↓ ↓H␈↓B. Current Support␈↓ ε∃ 7
␈↓ ↓H␈↓C. Relevant Experience␈↓ ε∃ 7
␈↓ ↓H␈↓D. Facilities␈↓ ε∃ 8
␈↓ ↓H␈↓E. Budget␈↓ ε∃ 9
␈↓ ↓H␈↓αDialnet Proposal␈↓ J2
␈↓ ↓H␈↓α␈↓ β71. Purpose ␈↓ εh␈↓ We propose to design protocols that can
␈↓ εh␈↓be implemented at any time-shared computer
␈↓ ↓H␈↓This is a request for a grant to support an␈↓ εh␈↓installation without joining any formal network.
␈↓ ↓H␈↓eighteen month study and experimental ␈↓ εh␈↓The hardware cost will be from $1000 to $5000
␈↓ ↓H␈↓implementation of protocols that would permit␈↓ εh␈↓depending on how di≠cult it is to connect
␈↓ ↓H␈↓ARPAnet-like facilities to be provided to any␈↓ εh␈↓devices to the computer. There will be
␈↓ ↓H␈↓time-sharing computer system that implemented␈↓ εh␈↓programs to operate a telephone dialer (rented
␈↓ ↓H␈↓them and equipped itself with telephone dialing␈↓ εh␈↓from the telephone company) and to transmit
␈↓ ↓H␈↓equipment and modems. We call the system ␈↓ εh␈↓signals and information according to the
␈↓ ↓H␈↓Dialnet by analogy, but unlike the ARPAnet, it␈↓ εh␈↓protocols. Any installation implementing the
␈↓ ↓H␈↓requires no administrator to "admit" new ␈↓ εh␈↓protocols will be able to communicate with any
␈↓ ↓H␈↓members. ␈↓ εh␈↓other. The only disadvantage compared with
␈↓ εh␈↓the ARPAnet will be lower speed and higher
␈↓ ↓H␈↓ The ARPAnet connects several hundred␈↓ εh␈↓cost when the volume of information
␈↓ ↓H␈↓computer facilities and allows users of one␈↓ εh␈↓transferred is very high.
␈↓ ↓H␈↓system to log in on others, allows transmission
␈↓ ↓H␈↓of messages between users of di≥erent ␈↓ εh␈↓ Like ARPAnet, Dialnet will be most
␈↓ ↓H␈↓computers, and allows the transfer of ≡les␈↓ εh␈↓useful to ␈↓↓full time-sharing systems␈↓. In such
␈↓ ↓H␈↓between computers. More generally, it allows␈↓ εh␈↓systems, each user has named disk ≡les than are
␈↓ ↓H␈↓interaction among programs in di≥erent ␈↓ εh␈↓kept in the system even when he is absent (and
␈↓ ↓H␈↓computers. ␈↓ εh␈↓therefore remotely accessible), and new ≡les can
␈↓ εh␈↓be created by ≡le transfer from other machines
␈↓ ↓H␈↓ These facilities have proven valuable in␈↓ εh␈↓and on receipt of messages. The usefulness of
␈↓ ↓H␈↓permitting collaboration between computer ␈↓ εh␈↓the message facilities normally requires that
␈↓ ↓H␈↓scientists at di≥erent sites and in permitting␈↓ εh␈↓users habitually log in each working day and
␈↓ ↓H␈↓nationwide access to unique facilities such as␈↓ εh␈↓are most bene≡cial when users have individual
␈↓ ↓H␈↓the MACSYMA system for computing with ␈↓ εh␈↓display terminals in their o≠ces. Further
␈↓ ↓H␈↓algebraic and analytic expressions at M.I.T. It␈↓ εh␈↓bene≡ts accrue when reports are normally
␈↓ ↓H␈↓permits a new form of publication in which␈↓ εh␈↓prepared at terminals and when secretaries use
␈↓ ↓H␈↓documents are kept in the computer, are ␈↓ εh␈↓terminals for letters and messages. However,
␈↓ ↓H␈↓continuously updatable, are immediately ␈↓ εh␈↓many less advanced installations have found
␈↓ ↓H␈↓accessible throughout the country, and in which␈↓ εh␈↓the ARPAnet useful and more and more
␈↓ ↓H␈↓comments from readers are accessible to other␈↓ εh␈↓systems are acquiring economical full time-
␈↓ ↓H␈↓readers. ␈↓ εh␈↓sharing capability.
␈↓ ↓H␈↓ The usefulness of the ARPAnet has ␈↓ εh␈↓ While we expect that the ≡rst users of
␈↓ ↓H␈↓prompted many non-defense installations to try␈↓ εh␈↓Dialnet will be regular computer users, the
␈↓ ↓H␈↓to connect to it, and in some cases this has been␈↓ εh␈↓corresponding ARPAnet facilities have been
␈↓ ↓H␈↓possible, but in other cases the institutional and␈↓ εh␈↓used by non computer people, Dialnet protocols
␈↓ ↓H␈↓≡nancial obstacles have been insuperable. The␈↓ εh␈↓will not require ability to program, and we
␈↓ ↓H␈↓main ≡nancial obstacles are the need for a␈↓ εh␈↓expect increasing use by others as terminals
␈↓ ↓H␈↓dedicated computer called an IMP costing ␈↓ εh␈↓become more widespread.
␈↓ ↓H␈↓about $80,000 at each site and the need for
␈↓ ↓H␈↓dedicated communication lines rented by DoD␈↓ εh␈↓ In order to make the picture more
␈↓ ↓H␈↓at great expense from the telephone companies.␈↓ εh␈↓concrete, here is a scenario of the use of the
␈↓ ↓H␈↓Other networks have been started, but they all␈↓ εh␈↓system.
␈↓ ↓H␈↓have problems of expense and ␈↓↓also of deciding
␈↓ ↓H␈↓↓who should be on them␈↓. Some facilities have
␈↓ ↓H␈↓gone to the expense of joining more than one
␈↓ ↓H␈↓network.
␈↓ ↓H␈↓αDialnet Proposal␈↓ J3
␈↓ ↓H␈↓α␈↓ β22. Scenario ␈↓ εh␈↓β␈↓ πλlocate dave@utex-chem3
␈↓ εh␈↓which obtains the following information from
␈↓ ↓H␈↓A user named Smith types on his terminal ␈↓ εh␈↓the speci≡ed computer:
␈↓ ↓H␈↓β␈↓ ↓hmail Organik ␈↓ εh␈↓β␈↓ πλDavid Piranha last logged out at
␈↓ ↓H␈↓β␈↓ ↓hDo you have any active work there on ␈↓ εh␈↓β␈↓ πλ23:47 on 9 May 1976. Plan: I will be
␈↓ ↓H␈↓β␈↓ ↓hhuman red cell carbonic anhydrase B? ␈↓ εh␈↓β␈↓ πλout of touch May 10 through 16. I
␈↓ εh␈↓β␈↓ πλplan to visit Martin Shumway at the
␈↓ ↓H␈↓The system looks up Organik in Smith's ␈↓ εh␈↓β␈↓ πλUniversity of Utah on May 17 and
␈↓ ↓H␈↓correspondent ≡le and discovers that his ␈↓ εh␈↓β␈↓ πλshould return by May 18. Will check
␈↓ ↓H␈↓computer pseudonym is "NAT" at a computer ␈↓ εh␈↓β␈↓ πλmail from Utah.
␈↓ ↓H␈↓called UTEX-CHEM1 that is reached at 512
␈↓ ↓H␈↓471-3221 via a 1200/150 baud asychronous ␈↓ εh␈↓Noting that the current date is May 14, so that
␈↓ ↓H␈↓modem. It selects an outgoing line with a␈↓ εh␈↓there is no point in getting the message there
␈↓ ↓H␈↓matching modem, dials the number and ␈↓ εh␈↓quickly, Smith types
␈↓ ↓H␈↓attempts to transmit the message. If the ␈↓ εh␈↓β␈↓ πλnight mail dave@utex-chem3
␈↓ ↓H␈↓transmitting computer cannot elicit a response␈↓ εh␈↓β␈↓ πλI am interested in your work on
␈↓ ↓H␈↓from the desired recipient, it informs the user␈↓ εh␈↓β␈↓ πλanhydrase B. If possible, give
␈↓ ↓H␈↓that it will try again later and send him a␈↓ εh␈↓β␈↓ πλpointers to online documentation,
␈↓ ↓H␈↓message when the transmission has succeeded.␈↓ εh␈↓β␈↓ πλelse give me a call at 415 497-4430
␈↓ ↓H␈↓If the user's correspondent ≡le did not contain␈↓ εh␈↓β␈↓ πλ(Stanford) or 415 321-7580 (home).
␈↓ ↓H␈↓the telepone number and modem characteristics,␈↓ εh␈↓The "night mail" command causes the message
␈↓ ↓H␈↓the user would have to supply them. ␈↓ εh␈↓transmission to be deferred until inexpensive
␈↓ εh␈↓nighttime telephone rates are in force.
␈↓ ↓H␈↓The identity and location of the sender and
␈↓ ↓H␈↓date and time of the message are automatically␈↓ εh␈↓Additional capabilities of the Dialnet system
␈↓ ↓H␈↓placed at the front of the message. At the␈↓ εh␈↓can be used to follow up on the above inquiry,
␈↓ ↓H␈↓receiving end, if the addressee is logged in on␈↓ εh␈↓as follows.
␈↓ ↓H␈↓the computer, he is immediately informed that
␈↓ ↓H␈↓mail has arrived and from whom. If not ␈↓ εh␈↓␈↓π⊗␈↓ The ability to access remote text ≡les will be
␈↓ ↓H␈↓logged in, he will receive the message the next␈↓ εh␈↓␈↓ πλprovided (with permission of the owners
␈↓ ↓H␈↓time he logs in. In either case, he can use the␈↓ εh␈↓␈↓ πλrequired, of course). This interactive reading
␈↓ ↓H␈↓same facility to respond: ␈↓ εh␈↓␈↓ πλfacility will include the addition of
␈↓ ↓H␈↓β␈↓ ↓hmail Smith ␈↓ εh␈↓␈↓ πλ"footnotes" to various parts of the text.
␈↓ ↓H␈↓β␈↓ ↓hDavid Piranha (DAVE@UTEX-CHEM3) has a ␈↓ εh␈↓␈↓ πλThese footnotes may be declared private (i.e.
␈↓ ↓H␈↓β␈↓ ↓hstudent working on inhibition by ␈↓ εh␈↓␈↓ πλbelonging to the reader) or public (available
␈↓ ↓H␈↓β␈↓ ↓hanions of anhydrase B. ␈↓ εh␈↓␈↓ πλto the author and possibly others).
␈↓ ↓H␈↓Following up on this lead, the user types ␈↓ εh␈↓␈↓π⊗␈↓ It will be possible to run programs on a
␈↓ ↓H␈↓β␈↓ ↓hlink dave@utex-chem3 ␈↓ εh␈↓␈↓ πλremote computer, permitting experiments
␈↓ εh␈↓␈↓ πλwith programs developed in other places.
␈↓ ↓H␈↓A connection is made to the speci≡ed computer␈↓ εh␈↓␈↓ πλThis facility will permit the sharing of
␈↓ ↓H␈↓and, if DAVE is logged in, he immediately ␈↓ εh␈↓␈↓ πλunique specialized capabilities over a
␈↓ ↓H␈↓receives a message saying ␈↓ εh␈↓␈↓ πλgeographically distributed population.
␈↓ ↓H␈↓β␈↓ ↓h** Link request from Smith @SU-CHEM7
␈↓ ↓H␈↓β␈↓ ↓h** ␈↓ εh␈↓␈↓π⊗␈↓ File transfers will be permitted, with suitable
␈↓ ↓H␈↓He could then type "␈↓βlink␈↓" and have his ␈↓ εh␈↓␈↓ πλerror detection and correction features, to
␈↓ ↓H␈↓keyboard and display e≥ectively linked to those␈↓ εh␈↓␈↓ πλpermit sharing of data. The communication
␈↓ ↓H␈↓of the caller, permitting a conversation. ␈↓ εh␈↓␈↓ πλprotocol should be able to adapt to a wide
␈↓ εh␈↓␈↓ πλrange of noise conditions on phone lines.
␈↓ ↓H␈↓Let us suppose, however, that DAVE is not
␈↓ ↓H␈↓logged in and the caller is so informed. He
␈↓ ↓H␈↓then types
␈↓ ↓H␈↓αDialnet Proposal␈↓ G4
␈↓ ↓H␈↓α␈↓ β.3. Protocols ␈↓ εh␈↓communication between computer programs.
␈↓ εh␈↓We will study how to make this mesh with
␈↓ ↓H␈↓In order to make these facilities available,␈↓ εh␈↓communication between computer programs
␈↓ ↓H␈↓suitable protocols must be designed, and in the␈↓ εh␈↓and people.
␈↓ ↓H␈↓course of this, a number of technical problems
␈↓ ↓H␈↓have to be solved. Besides the protocols
␈↓ ↓H␈↓themselves, which are communication ␈↓ εh␈↓α␈↓ λ"4. Research Issues
␈↓ ↓H␈↓procedures and data structures, there will be a
␈↓ ↓H␈↓recommended set of terminal-level commands␈↓ εh␈↓ There are many research issues, and we
␈↓ ↓H␈↓with syntax prompting and standard error ␈↓ εh␈↓don't expect to settle all of them in the time and
␈↓ ↓H␈↓messages. ␈↓ εh␈↓with the resources requested in this proposal.
␈↓ εh␈↓Since we expect many of the issues will be
␈↓ ↓H␈↓We believe that we have the experience to ␈↓ εh␈↓clari≡ed by the initial implementation, we will
␈↓ ↓H␈↓produce a set of workable protocols, and that it␈↓ εh␈↓concentrate on getting a reasonable ≡rst
␈↓ ↓H␈↓is better to start with an implementation than to␈↓ εh␈↓implementation into experimental use.
␈↓ ↓H␈↓standardize something that doesn't exist. The
␈↓ ↓H␈↓latter procedure in recent years has led to gold-␈↓ εh␈↓ Here are some of the issues we will study:
␈↓ ↓H␈↓plating the requirements to the extent that the
␈↓ ↓H␈↓standard is not implementable. ␈↓ εh␈↓1. What error correction facilities are required
␈↓ εh␈↓to make up for the de≡ciencies of telephone
␈↓ ↓H␈↓We propose to devise suitable protocols, test␈↓ εh␈↓lines?
␈↓ ↓H␈↓them at a few sites, publish them, and attempt
␈↓ ↓H␈↓to convince other installations to implement␈↓ εh␈↓2. What is the minimal necessary burden on the
␈↓ ↓H␈↓them. Almost certainly, initial experience will␈↓ εh␈↓time-sharing computers carrying out the
␈↓ ↓H␈↓produce a requirement for changes, and ␈↓ εh␈↓communication? What is the trade-o≥ between
␈↓ ↓H␈↓standardization committees will be formed and␈↓ εh␈↓bu≥er size and compute time?
␈↓ ↓H␈↓set to work. A likely forum for a
␈↓ ↓H␈↓standardization e≥ort would be through the␈↓ εh␈↓3. Can dial-up telephone communication rates
␈↓ ↓H␈↓ACM to the American National Standards ␈↓ εh␈↓meet most of the needs for communication
␈↓ ↓H␈↓Committee. ␈↓ εh␈↓between computers belonging to di≥erent
␈↓ εh␈↓research organizations?
␈↓ ↓H␈↓We propose to allow interaction with ARPAnet
␈↓ ↓H␈↓sites via TIPs and propose to discuss with␈↓ εh␈↓4. What is the best way to handle the fact that
␈↓ ↓H␈↓ARPA and DCA whether this will be allowed.␈↓ εh␈↓di≥erent modem speeds have di≥erent prices?
␈↓ εh␈↓Should one strive for a standard speed or can a
␈↓ ↓H␈↓The most general use of Dialnet involves a␈↓ εh␈↓wide variety be easily accomodated?
␈↓ ↓H␈↓program in one computer "waking up" and
␈↓ ↓H␈↓interacting with a program in another machine.␈↓ εh␈↓5. How will the improved communication a≥ect
␈↓ ↓H␈↓Dialnet protocols will handle human messages␈↓ εh␈↓research? Since changes will be slow, how can
␈↓ ↓H␈↓as a subcase of this, taking into account the fact␈↓ εh␈↓we tell as early as possible what the e≥ects will
␈↓ ↓H␈↓that the subcase will have the most application␈↓ εh␈↓be?
␈↓ ↓H␈↓for a long time to come. Messages about where
␈↓ ↓H␈↓to deliver a message sent by one time-sharing␈↓ εh␈↓6. What style of interaction is convenient for
␈↓ ↓H␈↓system to another will be handled as a special␈↓ εh␈↓both experienced and inexperienced users?
␈↓ ↓H␈↓sort of message that one program may send ␈↓ εh␈↓How can communication programs be made
␈↓ ↓H␈↓another in cases where the two programs are␈↓ εh␈↓self-teaching without being cumbersome?
␈↓ ↓H␈↓not written together, but each must know a
␈↓ ↓H␈↓certain "public" language. Thus we will
␈↓ ↓H␈↓attempt to make a general format for requests,
␈↓ ↓H␈↓questions, and assertions suitable for
␈↓ ↓H␈↓αDialnet Proposal␈↓ I5
␈↓ ↓H␈↓α␈↓ β 5. Research Plan ␈↓ εh␈↓α␈↓ λMAppendix A
␈↓ εh␈↓α␈↓ λ←Personnel
␈↓ ↓H␈↓We plan to undertake this project with rather
␈↓ ↓H␈↓modest sta≠ng. Initial emphasis will be on␈↓ εh␈↓Biography of John McCarthy
␈↓ ↓H␈↓designing and implementing experimental
␈↓ ↓H␈↓protocols using existing computer facilities at␈↓ εh␈↓BORN: September 4, 1927 in Boston,
␈↓ ↓H␈↓Stanford. Speci≡cally, we plan to create an␈↓ εh␈↓␈↓ π(Massachusetts
␈↓ ↓H␈↓experimental link between the computer
␈↓ ↓H␈↓facilities of the Stanford Arti≡cial Intelligence␈↓ εh␈↓EDUCATION: B.S. (Mathematics) California
␈↓ ↓H␈↓Laboratory (SAIL) and the Low Overhead ␈↓ εh␈↓␈↓ π(Institute of Technology, 1948. Ph.D.
␈↓ ↓H␈↓Timesharing System (LOTS). The only ␈↓ εh␈↓␈↓ π((Mathematics) Princeton University, 1951.
␈↓ ↓H␈↓additional equipment needed will be a
␈↓ ↓H␈↓telephone port with autodial capability for the␈↓ εh␈↓HONORS AND SOCIETIES: American
␈↓ ↓H␈↓LOTS computer. We expect this initial ␈↓ εh␈↓␈↓ π(Mathematical Society, Association for
␈↓ ↓H␈↓development phase will take about 6 months.␈↓ εh␈↓␈↓ π(Computing Machinery, Sigma Xi, Sloan
␈↓ εh␈↓␈↓ π(Fellow in Physical Science (1957-59), ACM
␈↓ ↓H␈↓Once a working set of protocols is developed,␈↓ εh␈↓␈↓ π(National Lecturer (1961), IEEE, A.M.
␈↓ ↓H␈↓implementation will be undertaken in ␈↓ εh␈↓␈↓ π(Turing Award from Association for
␈↓ ↓H␈↓cooperation with at least one outside university␈↓ εh␈↓␈↓ π(Computing Machinery (1971).
␈↓ ↓H␈↓research group. Professor Harry Huskey of the
␈↓ ↓H␈↓University of California at Santa Cruz ␈↓ εh␈↓PROFESSIONAL EXPERIENCE: Proctor
␈↓ ↓H␈↓Information Sciences Department has agreed to␈↓ εh␈↓␈↓ π(Fellow, Princeton University (1950-51),
␈↓ ↓H␈↓having it on one of the machines there, ␈↓ εh␈↓␈↓ π(Higgins Research Instructor in
␈↓ ↓H␈↓probably a PDP-11/45 or a Burroughs B5000.␈↓ εh␈↓␈↓ π(Mathematics, Princeton University (1951-
␈↓ ↓H␈↓We propose to install a telephone dialout port␈↓ εh␈↓␈↓ π(53), Acting Assistant Professor of
␈↓ ↓H␈↓there and hire a UC Santa Cruz student to do␈↓ εh␈↓␈↓ π(Mathematics, Stanford University (1953-
␈↓ ↓H␈↓the necessary system programming and testing.␈↓ εh␈↓␈↓ π(55), Assistant Professor of Mathematics,
␈↓ ↓H␈↓The implementation should take about six ␈↓ εh␈↓␈↓ π(Dartmouth College (1955-58), Assistant
␈↓ ↓H␈↓months. ␈↓ εh␈↓␈↓ π(Professor of Communication Science,
␈↓ εh␈↓␈↓ π(M.I.T. (1958-61), Associate Professor of
␈↓ ↓H␈↓In the following six months, we plan to test,␈↓ εh␈↓␈↓ π(Communication Science, M.I.T. (1961-62),
␈↓ ↓H␈↓evaluate, and modify the protocols. During the␈↓ εh␈↓␈↓ π(Professor of Computer Science Stanford
␈↓ ↓H␈↓latter part of this period, we plan to publish the␈↓ εh␈↓␈↓ π(University (1962 - present).
␈↓ ↓H␈↓protocols and encourage additional groups to
␈↓ ↓H␈↓join the Dialnet community. ␈↓ εh␈↓PROFESSIONAL RESPONSIBILITIES
␈↓ εh␈↓␈↓ π(AND SCIENTIFIC INTERESTS: With
␈↓ εh␈↓␈↓ π(Marvin Minsky organized and directed
␈↓ εh␈↓␈↓ π(the Arti≡cial Intelligence Project at M.I.T.
␈↓ εh␈↓Organized and directs Stanford Arti≡cial
␈↓ εh␈↓␈↓ π(Intelligence Project
␈↓ εh␈↓Developed the LISP programming system for
␈↓ εh␈↓␈↓ π(computing with symbolic expressions,
␈↓ εh␈↓␈↓ π(participated in the development of the
␈↓ εh␈↓␈↓ π(ALGOL 58 and the ALGOL 60
␈↓ εh␈↓␈↓ π(languages.
␈↓ εh␈↓Present scienti≡c work is in the ≡elds of
␈↓ εh␈↓␈↓ π(Arti≡cial Intelligence, Computation with
␈↓ ↓H␈↓αDialnet Proposal␈↓ J6
␈↓ ↓H␈↓␈↓ αλSymbolic Expressions, Mathematical ␈↓ εh␈↓[11] (with James Painter) "Correctness of a
␈↓ ↓H␈↓␈↓ αλTheory of Computation, Time-Sharing ␈↓ εh␈↓␈↓ π(Compiler for Arithmetic Expressions",
␈↓ ↓H␈↓␈↓ αλcomputer systems. ␈↓ εh␈↓␈↓ π(Amer. Math. Soc., ␈↓↓Proc. Symposia in
␈↓ εh␈↓↓␈↓ π(Applied Math., Math. Aspects of Computer
␈↓ ↓H␈↓PUBLICATIONS: ␈↓ εh␈↓↓␈↓ π(Science␈↓, New York, 1967.
␈↓ ↓H␈↓[1] "Towards a Mathematical Theory of ␈↓ εh␈↓[12] "Programs with Common Sense", in
␈↓ ↓H␈↓␈↓ αλComputation", in ␈↓↓Proc. IFIP Congress 62␈↓, ␈↓ εh␈↓␈↓ π(Marvin Minsky (ed.), ␈↓↓Semantic Information
␈↓ ↓H␈↓␈↓ αλNorth-Holland, Amsterdam, 1963. ␈↓ εh␈↓↓␈↓ π(Processing␈↓, MIT Press, Cambridge, 1968.
␈↓ ↓H␈↓[2] "A Basis for a Mathematical Theory of␈↓ εh␈↓[13] (with Lester Earnest, D. Raj. Reddy,
␈↓ ↓H␈↓␈↓ αλComputation", in P. Bia≥ort and D. ␈↓ εh␈↓␈↓ π(Pierre Vicens) "A Computer with Hands,
␈↓ ↓H␈↓␈↓ αλHershberg (eds.), ␈↓↓Computer Programming ␈↓ εh␈↓␈↓ π(Eyes, and Ears", ␈↓↓Proc. AFIPS Conf.␈↓
␈↓ ↓H␈↓↓␈↓ αλand Formal Systems␈↓, North-Holland, ␈↓ εh␈↓␈↓ π((FJCC), 1968.
␈↓ ↓H␈↓␈↓ αλAmsterdam, 1963.
␈↓ εh␈↓[14] (with Patrick Hayes) "Some Philosophical
␈↓ ↓H␈↓[3] (with S. Boilen, E. Fredkin, J.C.R. ␈↓ εh␈↓␈↓ π(Problems from the Standpoint of Arti≡cial
␈↓ ↓H␈↓␈↓ αλLicklider) "A Time-Sharing Debugging ␈↓ εh␈↓␈↓ π(Intelligence", in Donald Michie (ed.),
␈↓ ↓H␈↓␈↓ αλSystem for a Small Computer", ␈↓↓Proc. ␈↓ εh␈↓␈↓ π(␈↓↓Machine Intelligence 4␈↓, American Elsevier,
␈↓ ↓H␈↓↓␈↓ αλAFIPS Conf.␈↓ (SJCC), Vol. 23, 1963. ␈↓ εh␈↓␈↓ π(New York, 1969.
␈↓ ↓H␈↓[4] (with F. Corbato, M. Daggett) "The ␈↓ εh␈↓[15] "The Home Information Terminal", ␈↓↓Man
␈↓ ↓H␈↓␈↓ αλLinking Segment Subprogram Language ␈↓ εh␈↓↓␈↓ π(and Computer, Proc. Int. Conf., Bordeaux,
␈↓ ↓H␈↓␈↓ αλand Linking Loader Programming ␈↓ εh␈↓↓␈↓ π(1970␈↓, S. Karger, New York, 1972.
␈↓ ↓H␈↓␈↓ αλLanguages", ␈↓↓Comm. ACM␈↓, July 1963.
␈↓ ↓H␈↓[5] "Problems in the Theory of Computation",
␈↓ ↓H␈↓␈↓ αλ␈↓↓Proc. IFIP Congress 1965␈↓.
␈↓ ↓H␈↓[6] "Time-Sharing Computer Systems", in W.
␈↓ ↓H␈↓␈↓ αλOrr (ed.), ␈↓↓Conversational Computers␈↓, Wiley,
␈↓ ↓H␈↓␈↓ αλ1966.
␈↓ ↓H␈↓[7] "A Formal Description of a Subset of
␈↓ ↓H␈↓␈↓ αλAlgol", in T. Steele (ed.), ␈↓↓Formal
␈↓ ↓H␈↓↓␈↓ αλLanguage Description Languages for
␈↓ ↓H␈↓↓␈↓ αλComputer Programming␈↓, North-Holland,
␈↓ ↓H␈↓␈↓ αλAmsterdam, 1966.
␈↓ ↓H␈↓[8] "Information", ␈↓↓Scienti≡c American␈↓,
␈↓ ↓H␈↓␈↓ αλSeptember 1966.
␈↓ ↓H␈↓[9] "Computer Control of a Hand and Eye", in
␈↓ ↓H␈↓␈↓ αλ␈↓↓Proc. Third All-Union Conference on
␈↓ ↓H␈↓↓␈↓ αλAutomatic Control (Technical Cybernetics)␈↓,
␈↓ ↓H␈↓␈↓ αλNauka, Moscow, 1967 (Russian).
␈↓ ↓H␈↓[10] (with D. Brian, G. Feldman, and J. Allen)
␈↓ ↓H␈↓␈↓ αλ"THOR ¬ A Display Based Time-Sharing
␈↓ ↓H␈↓␈↓ αλSystem", ␈↓↓Proc. AFIPS Conf.␈↓ (FJCC), Vol.
␈↓ ↓H␈↓␈↓ αλ30, Thompson, Washington, D.C., 1967.
␈↓ ↓H␈↓αDialnet Proposal␈↓ J7
␈↓ ↓H␈↓α␈↓ β.Appendix B ␈↓ εh␈↓α␈↓ λMAppendix C
␈↓ ↓H␈↓α␈↓ β⊃Current Support ␈↓ εh␈↓α␈↓ λ⊗Relevant Experience
␈↓ ↓H␈↓Prof. McCarthy is currently being supported by␈↓ εh␈↓The Stanford Arti≡cial Intelligence Laboratory
␈↓ ↓H␈↓the Advanced Research Projects Agency under␈↓ εh␈↓has been active in the development of remote
␈↓ ↓H␈↓Contract MDA903-76-C-0206 (1 January 1976 -␈↓ εh␈↓terminal and data communication systems for a
␈↓ ↓H␈↓30 June 1977, $916,000/year) and by the ␈↓ εh␈↓number of years. SAIL sta≥ members
␈↓ ↓H␈↓National Science Foundation for research in␈↓ εh␈↓participated in the planning of the ARPA
␈↓ ↓H␈↓Veri≡cation Oriented Programming under ␈↓ εh␈↓Network and the SAIL computer facility
␈↓ ↓H␈↓Grant Number MCS76-00327 (June 1976 - ␈↓ εh␈↓(Appendix B) was among the early systems in
␈↓ ↓H␈↓June 1978, $160,000/year). Prof. McCarthy also␈↓ εh␈↓which the ARPAnet protocol was implemented
␈↓ ↓H␈↓oversees (but recieves no personal support␈↓ εh␈↓(1972). Many of the capabilities that we
␈↓ ↓H␈↓from) an NSF research contract on Exploratory␈↓ εh␈↓propose to make available through Dialnet are
␈↓ ↓H␈↓Study of Computer Integrated Assembly ␈↓ εh␈↓available in some form in the ARPAnet. A
␈↓ ↓H␈↓Systems under Contract NSF APR74-01390 ␈↓ εh␈↓similar capability has been developed over a
␈↓ ↓H␈↓A04 (April 1976 - April 1978, $225,000/year).␈↓ εh␈↓leased telephone line between SAIL and the
␈↓ εh␈↓Institute for Mathematical Studies in the Social
␈↓ ↓H␈↓α␈↓ αhPending Applicatiions ␈↓ εh␈↓Sciences, located several miles away on the
␈↓ εh␈↓Stanford campus.
␈↓ ↓H␈↓Prof. McCarthy has recently submitted two
␈↓ ↓H␈↓proposals (involving no personal support) to␈↓ εh␈↓Since 1974, the SAIL timesharing system has
␈↓ ↓H␈↓the National Science Foundation. "A Uni≡ed␈↓ εh␈↓been ≡tted with automatic dialing equipment in
␈↓ ↓H␈↓Approach to Automatic Programming" (MCS- ␈↓ εh␈↓conjunction with Dataphone service, to permit
␈↓ ↓H␈↓7683655) would run two years at a level of␈↓ εh␈↓computer-initiated data communications with
␈↓ ↓H␈↓$96,000/year, while "Veri≡cation of Operating␈↓ εh␈↓other computers. This permits the SAIL
␈↓ ↓H␈↓Systems written in Concurrent Pascal" would␈↓ εh␈↓computer to simulate the functions of
␈↓ ↓H␈↓run two years at a level of $37,000/year. ␈↓ εh␈↓teletypewriter terminals, so as to access
␈↓ εh␈↓programs and data ≡les in external machines.
␈↓ εh␈↓This is the principal communication technique
␈↓ εh␈↓that will be employed in Dialnet, although the
␈↓ εh␈↓protocols will be substantially di≥erent for the
␈↓ εh␈↓sake of e≠ciency and reliability.
␈↓ εh␈↓The general problem of utilizing dial networks
␈↓ εh␈↓for computer communication has been under
␈↓ εh␈↓study in the Laboratory for some time (e.g. [1]).
␈↓ εh␈↓[1] Rubin, Je≥, "Computer Communication via
␈↓ εh␈↓␈↓ π8the Dial-up Network", ␈↓↓Minutes of the
␈↓ εh␈↓↓␈↓ π8DECsystem-10 Spring-75 DECUS
␈↓ εh␈↓↓␈↓ π8Meeting␈↓, Digital Equipment Computer
␈↓ εh␈↓␈↓ π8Users Society, Maynard, Mass., 1975.
␈↓ ↓H␈↓αDialnet Proposal␈↓ I8
␈↓ ↓H␈↓α␈↓ β,Appendix D ␈↓ εh␈↓The computer facilities of the Stanford Low
␈↓ ↓H␈↓α␈↓ βCFacilities ␈↓ εh␈↓Overhead Timesharing System will consist of
␈↓ εh␈↓the following. In order to participate in this
␈↓ ↓H␈↓The computer facilities of the Stanford␈↓ εh␈↓project, a telephone data set with autodialer
␈↓ ↓H␈↓Arti≡cial Intelligence Laboratory include the␈↓ εh␈↓and computer interface will be needed.
␈↓ ↓H␈↓following equipment, most of it purchased with
␈↓ ↓H␈↓U.S. Government research funds. At present,␈↓ εh␈↓Central processor: Digital Equipment
␈↓ ↓H␈↓we anticipate the need of additional equipment␈↓ εh␈↓␈↓ λλCorporation KL20 (36 bit
␈↓ ↓H␈↓only at other experimental sites. ␈↓ εh␈↓␈↓ λλmachine).
␈↓ ↓H␈↓Central processors: Digital Equipment ␈↓ εh␈↓Primary store: 262k words (36 bit) of 1
␈↓ ↓H␈↓␈↓ αhCorporation KL10 and KA10. ␈↓ εh␈↓␈↓ λλmicrosecond DEC core.
␈↓ ↓H␈↓Primary store: 512k words (36 bit) of 1 to␈↓ εh␈↓File store: DEC RP06 disc drive (capacity: 1.6
␈↓ ↓H␈↓␈↓ αh1.6 microsecond core (DEC and ␈↓ εh␈↓␈↓ λλx 10↑9 bits).
␈↓ ↓H␈↓␈↓ αhAmpex)
␈↓ εh␈↓Peripherals: DEC 9-track tape drive, Printronix
␈↓ ↓H␈↓Swapping store: Librascope disk (5 million␈↓ εh␈↓␈↓ λλline printer, 48 port line scanner.
␈↓ ↓H␈↓␈↓ αhwords).
␈↓ εh␈↓Terminals: 50 ADM3 displays.
␈↓ ↓H␈↓File store: Ampex disc ≡le (3330-11 type),
␈↓ ↓H␈↓␈↓ αh6 spindles (capacity: 7.8 x 10␈↓∧9␈↓ bits).
␈↓ ↓H␈↓Peripherals: 4 Dectape drives, 2 mag tape
␈↓ ↓H␈↓␈↓ αhdrives (7 channel), line printer,
␈↓ ↓H␈↓␈↓ αhCalcomp plotter, Xerox Graphics
␈↓ ↓H␈↓␈↓ αhPrinter
␈↓ ↓H␈↓Terminals: 58 Data Disc displays, 6 III
␈↓ ↓H␈↓␈↓ αhdisplays, 4 IMLAC displays, 10
␈↓ ↓H␈↓␈↓ αhDatamedia displays, 15 Teletype
␈↓ ↓H␈↓␈↓ αhterminals, 5 TI terminals
␈↓ ↓H␈↓Realtime processors: DEC PDP-11/45 and SPS-
␈↓ ↓H␈↓␈↓ αh41 with 8k words (16 bit) of core
␈↓ ↓H␈↓␈↓ αhand 197k words of Intel MOS
␈↓ ↓H␈↓␈↓ αhmemory.
␈↓ ↓H␈↓Communications processor: BBN IMP
␈↓ ↓H␈↓␈↓ αh(Honeywell DDP-516) connected to
␈↓ ↓H␈↓␈↓ αhthe ARPA Network.
␈↓ ↓H␈↓Special equipment: Audio input and output
␈↓ ↓H␈↓␈↓ αhsystems, hand-eye equipment (4 TV
␈↓ ↓H␈↓␈↓ αhcameras, 2 arms), remote-controlled
␈↓ ↓H␈↓␈↓ αhcart.
␈↓ ↓H␈↓αDialnet Proposal␈↓ J9
␈↓ ↓H␈↓α␈↓ ¬}Appendix E
␈↓ ↓H␈↓α␈↓ ε!Budget
␈↓ ↓H␈↓Eighteen months beginning 1 July 1977.
␈↓ ↓H␈↓β␈↓ α_ Man 1 Jul.'77 to 1 Jul.'78 to
␈↓ ↓H␈↓β␈↓ α_ Months 30 Jun.'78 31 Dec.'78
␈↓ ↓H␈↓β␈↓ α_A. SALARIES AND WAGES
␈↓ ↓H␈↓β␈↓ α_ 1.Senior Personnel
␈↓ ↓H␈↓β␈↓ α_ Professor John McCarthy 1 - - - - - -
␈↓ ↓H␈↓β␈↓ α_ Principal Investigator (5%)
␈↓ ↓H␈↓β␈↓ α_ Lester Earnest (10%) 2 $3,689 $1,955
␈↓ ↓H␈↓β␈↓ α_ Res. Scientist & Lecturer
␈↓ ↓H␈↓β␈↓ α_ 2..Other Personnel
␈↓ ↓H␈↓β␈↓ α_ a.Programmer
␈↓ ↓H␈↓β␈↓ α_ Martin Frost (50%) 7,806 4,137
␈↓ ↓H␈↓β␈↓ α_ b.Student Research Assistants
␈↓ ↓H␈↓β␈↓ α_ Acad. yr. 50%, Sum. 100% 7,155 4,550
␈↓ ↓H␈↓β␈↓ α_ Acad. yr. 50%, Sum. 100% 4,293 4,550
␈↓ ↓H␈↓β␈↓ α_ (1 Jan. '78 on)
␈↓ ↓H␈↓β␈↓ α_ c.Support Personnel
␈↓ ↓H␈↓β␈↓ α_ Secretary (10%) 1,160 615
␈↓ ↓H␈↓β␈↓ α_ Electronic Technician (15%) 2,226 1,180
␈↓ ↓H␈↓β␈↓ α_ ______ ______
␈↓ ↓H␈↓β␈↓ α_ Total Salaries and Wages 26,329 16,987
␈↓ ↓H␈↓β␈↓ α_B. STAFF BENEFITS 5,221 3,488
␈↓ ↓H␈↓β␈↓ α_ 9/1/76 to 8/31/77:19%
␈↓ ↓H␈↓β␈↓ α_ 9/1/77 to 8/31/78:20%
␈↓ ↓H␈↓β␈↓ α_ 9/1/78 to 8/31/79:20.8% ______ ______
␈↓ ↓H␈↓β␈↓ α_C. TOTAL SALARIES, WAGES,
␈↓ ↓H␈↓β␈↓ α_ AND STAFF BENEFITS 31,550 20,475
␈↓ ↓H␈↓β␈↓ α_D. PERMANENT EQUIPMENT
␈↓ ↓H␈↓β␈↓ α_ 2 computer-Dataphone interfaces 4,000
␈↓ ↓H␈↓β␈↓ α_E. EXPENDABLE SUPPLIES 1,000 500
␈↓ ↓H␈↓β␈↓ α_ & EQUIPMENT(e.g., copying,
␈↓ ↓H␈↓β␈↓ α_ office supplies,postage,
␈↓ ↓H␈↓β␈↓ α_ freight,consulting,honoraria)
␈↓ ↓H␈↓β␈↓ α_F. TRAVEL 1,300 700
␈↓ ↓H␈↓β␈↓ α_G. PUBLICATION COSTS 500 500
␈↓ ↓H␈↓αDialnet Proposal␈↓ :10
␈↓ ↓H␈↓β␈↓ α_H. COMPUTER COSTS - - - - - -
␈↓ ↓H␈↓β␈↓ α_I. OTHER COSTS
␈↓ ↓H␈↓β␈↓ α_ Telephone 200 100
␈↓ ↓H␈↓β␈↓ α_ 2 computer phone lines with autodial 920 480
␈↓ ↓H␈↓β␈↓ α_ ______ ______
␈↓ ↓H␈↓β␈↓ α_J. TOTAL DIRECT COSTS (C through I) 39,470 22,755
␈↓ ↓H␈↓β␈↓ α_K. INDIRECT COSTS 20,573 13,198
␈↓ ↓H␈↓β␈↓ α_ 58% of (J less D) ------ ------
␈↓ ↓H␈↓β␈↓ α_L. TOTAL COSTS (J plus K) $60,043 $35,953
␈↓ ↓H␈↓β␈↓ α_ ------- -------
␈↓ ↓H␈↓β␈↓ α_Total Budget (eighteen months) $95,996
␈↓ ↓H␈↓¬The source file of this document is DIALNE.PRO[E76,JMC]@SU-AI.