Networking and Communications 5P41 RROO785-11 The EtherTIP software has undergone further enhancements in the past year. Portions of this work was done in conjunction with the Stanford Computer Science Department. Among those enhancements are the following: a. It now accepts incoming connections to line printer ports, and for remote system diagnosis. b. It can simulate the "old" Stanford EtherTIP for users who have not yet made the transition to the new environment. c. The user interface is more flexible to suit the needs of an increasingly diverse user community. The EtherTIP software has developed into a very stable system, and one enjoying good use within the SUMEX community. 5. 10 MB/SEC Ethernet Development -- SUMEX made a major move this past year to begin our transfer to a 10 megabit/sec network. While the current 3 megabit/sec network continues to serve us well, many new workstations and printers are coming on the market with only 10 MB/SEC interfaces, and in addition, since 3 MB/SEC networks were only used a very few selected settings. it is becoming increasingly difficult to find replacement parts when failures do occur. Therefore, this past year saw several efforts involved in installing and supporting the SUMEX 10MB/SEC Ethernet ; a. Reworking the entire Ethernet system software to handle both 3 and 10 megabit link level standards, i.e., addressing and encapsulation are transparent to the user levels. We similarly made the network link level protocols transparent to the the user level software. In this way one can communicated using PUP protocols on a 1OMB/SEC ethernet and the user software does not have to change. b. Adding address resolution protocols for PUP and IP so that the 3MB/SEC byte addresses can be translated to 10 MB/SEC hardware addresses for the link level. This enables one to communicate using PUP or IP between 3 and 10 megabit hosts. c. Integrating XNS and IP into the PUP routing mechanism. d. Solving some rather subtle software/hardware integration problems in order to simulate “ethernet” on the HPP/Welch Road "twisted pair" ethernet. e. Bringing up the 3 MB/SEC EtherTIP on the 10 MB/SEC network was a proof that the above worked. It was done without any changes to the TIP software itself by simply relinking it with the 10 MB/SEC system software. This required only one additional piece of logic. When a 10 MB/SEC host wants to communicate using PUP which is a 3 MB/SEC protocol, then it must find its PUP address from some host on the 10 MB/SEC network. The gateway maintains a translation table, and listens for such requests, thus translating the 10 MB/SEC hardware address into a “soft PUP address," and replying to the requesting host. 6. HPP-SUMEX Communication link - The Heuristic Programing Project E. A. Feigenbaum 26 5P41 RROO7&5-11 Networking and Communications (HPP) relocated from its campus location to 701 Welch Road, adjacent to the Stanford campus. Since this group is a primary user of the SUMEX computer facility and the principal focus for core AI research, a communication link between the new location and SUMEX machine room was imperative. Several communication schemes for establishing a reliable and relatively fast link were considered, namely ; microwave, laser, infrared, direct ethernet (by trenching and placing a direct ethernet cable), ATT’s T1 service and others. All of the above schemes would have necessitated large budgetary outlays and some would have imposed lengthy time delays (getting permits and the like) due to jurisdictional boundaries. The idea of using bare copper telephone pair already in place looked very attractive especially if reasonable speed and reliability could be achieved. The wire distance between the above mentioned locations is approximately 2000 ft. A design goal was established to try to develop a communication link with Ethernet type speed ( 3MB/SEC ) between these two locations. Utilizing high driving capacity drivers (differential) and ultra high speed, high sensitivity receivers a transceiver was designed and tested for maximum transmission speed with maximum reliability. The final configuration resulted in a half duplex transmission over a bare copper twisted pair in each direction utilizing Manchester coding at a reliable transmission speed of 1.25MBs/sec. each direction for an aggregate speed of 2.5MBs/sec. This communication link has been in operation for about six months now without any appreciable down time or noticeable error rate or data delays. Many HPP researchers are utilizing this link to communicate with SUMEX and the University Ethernet network. In addition, various Lisp machines and printers located in the HPP facility and connected to a local network there are also able to communicate with the University network. INTERNET SOFTWARE One major issue we face at SUMEX-AIM in support of our network environment is the lack of standardization in network protocols among various vendors. Currently, many vendors are adding support to their products for the Internet (IP/TCP) protocols. SUMEX continues to support the IP/TCP protocols on the DEC2060, and we are currently alpha-testing a release of Interlisp-D which also supports IP/TCP protocols. In addition, we sucessfully adapted the IP/TCP software to our VAX systems running UNIX 4.2BSD. This Vax TCP adaptation involved provisions for subnet routing, 3 MB/SEC byte swap problems, encapsulation problems and 10 MB/SEC debugging with our gateways. 27 E. A. Feigenbaum Progress - Progress in Core Research 5P41 RROO785-11 I.A.2.6. Progress in Core Research Over the past year we have continued to support several core research activities aimed at developing information resources, basic AI research, and tools of general interest to the SUMEX-AIM community. SUMEX is providing only partial support for these projects, with complementary funding coming from ARPA, ONR, NLM and NSF contracts and grants to the Stanford Heuristic Programming Project. Core Research Core Research at SUMEX-AIM focuses on understanding the roles of knowledge in symbolic problem solving systems, its representation in software and hardware, its use for inference, and its acquisition. We are continuing to develop new tools for system builders and to improve old ones. The research crosses a number of application domains, as reflected in the subprojects discussed earler, but the main issues that we are addressing in this research are those fundamental to all aspects of AI. We believe this core research is broadening and deepening the groundwork for the design and construction of even more capable and effective biomedical systems. AS mentioned above, although our style of research is largely empirical, the questions we are addressing are fundamental. The three major research issues in AI have, since its beginning, been knowledge representation, control of inference (search), and learning. Within these topics, we will be asking the following kinds of questions. As our work progresses, we hope to leave behind several prototype systems that can be developed by others in the medical community. 1. Knowledge Representation -- How can we represent causal models and structural information? What are the relative benefits of logic-based, rule- based, and frame-based systems? How can we represent temporal relations and events so that reasoning over time is efficient? to . Knowledge Acquisition -- How can an expert system acquire new knowledge without consuming substantial time from experts? Can we improve the knowledge engineering paradigm enough to make a difference? Can automatic learning programs be designed that will work across many disciplines? Will cooperative man-machine systems be able to open the communication channel between expert and expert system? 3. Knowledge Utilization -- By what inference methods can a variety of sources of knowledge of diverse types be made to contribute jointly and efficiently toward solutions? What is the nature of strategy and control information? Plans for the Coming Year Several systems have been developed in recent years to serve as vehicles for knowledge engineering and research on knowledge representation and its use. Knowledge acquisition (including machine learning) and advanced architectures for AI will be the two areas of most new activity in the coming year. Research on these topics obviously must draw on on-going work in representation and control. In particular, we will focus on e Inductive learning of MYCIN-like rules from case data in the domain of diagnosing disorders where the chief complaint is jaundice; e Learning from experience in domains where the means for interpreting new BE. A. Feigenbaum 28 §P41 RR0OO785-11 Progress - Progress in Core Research data are largely contained in the emerging (and thus incomplete and not wholly correct) theory; eLearning by watching a medical expert diagnose cases presented by NEOMYCIN; e Investigating complex signal understanding systems for ways to exploit and represent concurrency with a view toward hardware and software architectures that may be capable of several orders of magnitude improvement in performance. Further information on the core research at SUMEX-AIM and the Heuristic Programming Project can be found in the Projects section starting on page 89. 29 E. A. Feigenbaum Progress - Resource Operations Statistics 5Pi1 RROO785-11 J.A.2.7. Resource Operations Statistics The following data give an overview of various aspects of SUMEX-AIM resource usage. There are 5 subsections containing data respectively for: 1. Overall resource loading data (page 31). to . Relative system loading by community (page 33). 3. Individual project and community usage (page 36). 4, Network usage data (page 44). 5. System reliability data (page 44). For the most part, the data used for these plots covers the entire span of the SUMEX-AIM project. This includes data from both the TENEX KI10 system and the current DECsystem 2060. At the point where the SUMEX-AIM community switched over to the 2060 (February, 1983), you will notice severe changes in most of the graphs. This is due to many reasons which I will mentioned briefly here : 1. Even though the Tenex operating system used on the KI10 was a forerunner of the current Tops20 operating system, the Tops20 system is still different from Tenex is many ways. Tops20 uses a radically different job scheduling mechanism, different methods for computing monitor statistics, different I/O routines, etc. In general, it can not be assumed that statistics measured on the Tenex system correlate one to one with similar statistics under Tops20. ho . The KL10 processor on the 2060 is a faster processor than the KI10 processor used previously. Hence, a job running on the KL10 will use less CPU time than the same job running on the KI10. This aspect is further complicated by the fact that the SUMEX KI10 system was a dual processor system. 3. The SUMEX-AIM Community was changing during the time of the transfer to the 2060. The usage of the GENET community on SUMEX had just been phased out. This part of the community accounted for much of the CPU time used by the AIM community. Since the purchase of the 2060 was partially funded by the Heuristic Programming Project (HPP), an additional number of HPP Core Research Projects started using the 2060, increasing the Stanford communities usage of the machine. And finally, the move to the 2060 occurred during a pivotal time in the community when more and more projects were either moving to their own local timesharing machines, or onto specialized Lisp workstations. It also was the time for the closure of many long time SUMEX- AIM projects, like Dendral and Puff/VM. Any conclusions reached by comparing the data before and after February, 1983 should be done with caution. The data is included in this years annual report mostly for casual comparison. Starting next year, only data from the 2060 will be recorded on the annual report. Readers will be referred to previous annual reports (such as this one) for data from the KI10 Tenex system. BE. A. Feigenbaum 30 5P41 RROO785-11 Progress - Resource Operations Statistics i Overall Resource Loading Data The following plots display several different aspects of system loading over the life of the project. This data includes usage of the Tenex KI10 system and the current DECsystem 2060. These plots include total CPU time delivered per month, the peak number of jobs logged in, and the peak load average. The monthly "peak" value of a given variable is the average of the daily peak values for that variable during the month. Thus, these "peak" values are representative of average monthly loading maxima and do not reflect the largest excursions seen on individual days, which are much higher. 800F total CPU Usage Hours/Month 600}; 400} 200} Oo 1 1 1 h. Sweet 1 1 1. 1 1 i 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Figure 7: Total CPU Time Consumed by Month 31 E. A. Feigenbaum Progress - Resource Operations Statistics 5P41 RROO785-11 60, SOF peak Daily Jobs 40} 30} 20} 10 Oo £ 1 I 2 1 rl 1 i 1 1 1 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Figure 8: Peak Number of Jobs by Month T Peak Daily | Load Average yO © A GT DA N & 0 wh i 4. 1 i 1 1 1 L 1 J 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Figure 9: Peak Load Average by Month E. A. Feigenbaum 32 5P41 RRO00785-11 Progress - Resource Operations Statistics Relative System Loading by Community The SUMEX resource is divided, for administrative purposes, into three major communities: user projects based at the Stanford Medical School (Stanford Projects), user projects based outside of Stanford (Nattonal AIM Projects), and common system development efforts (System Staff). As defined in the resource management plan approved by the BRP at the start of the project, the available system CPU capacity and file space resources are divided between these communities as follows: Stanford 40% AIM 40% Staff 20% The “available" resources to be divided up in this way are those remaining after various monitor and community-wide functions are accounted for. These include such things as job scheduling, overhead, network service, file space for subsystems, documentation, etc. The monthly usage of CPU resources and terminal connect time for each of these three communities relative to their respective aliquots is shown in the plots in Figure 10 and Figure 11. As mentioned on page 30, these plots include both KILO and 2060 usage data. 33 E. A. Feigenbaum Progress - Resource Operations Statistics 5P41 RRO0785-11 40 % CPU Time National Projects 30 20 . paneer ee oO 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1964 7985 40 % CPU Used Stanford Projects 30 20 10 i i 4. 4 1 dew: a L 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1. 4 1984 1985 40 % CPU Used System Staff 30 20 70} afar oO i 1974 1975 1976 1977 1978 19 79 1980 1981 1932 1883 1984 1935 Figure 10: Monthly CPU Usage by Community E. A. Feigenbaum 34 5P41 RROO785-11 Progress - Resource Operations Statistics 6600 Connect Time National Projects 5000} Hours/Month 4000 3000 2000 1000 1 L 1 Oo de 1 4 4 1 1 ‘ 4 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 8000 Connect Time Stanford Projects 7600F Hours/Month 6000 5000 4000 Any hh 3000 2000 1000 L A. —t. 1 1 4 i re 4 ‘ 4 O 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 habit Connect Time System Staff 5000+ Hours/Month 4000 3000 2000 10C0 1 I L 1. 4. 1 i 1 4. t 4 oO 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Figure 11: Monthly Terminal Connect Time by Community 35 E. A. Feigenbaum Progress - Resource Operations Statistics 5P41 RROO785-11 Individual Project and Community Usage The following histogram and table show cumulative resource usage by collaborative project and community during the past grant year. The histogram displays the project distribution of the total CPU time consumed between May 1, 1983 and April 30, 1984, on the SUMEX-AIM DECsystem2060 system. In the table following, entries include a text summary of the funding sources (outside of SUMEX-supplied computing resources) for currently active projects, total CPU consumption by project (Hours), total terminal connect time by project (Hours), and average file space in use by project (Pages, 1 page = 512 computer words). These data were accumulated for each project for the months between May, 1983 and May, 1984. Several of the projects admitted to the National AIM community use the Rutgers- AIM resource as their home base. We do not explicitly list these projects in this annual report covering the Stanford SUMEX-AIM resource. We do record information about the Rutgers resource itself, however, and note its separate resource status with the flag "(Rutgers-AIM]". E. A. Feigenbaum 36 5P41 RROO785-11 AIM Administration AIM Pilots AIM Users ACT Caduceus SECS Modeis of Human Cag Solver Pulf-VM Rutgers AGE Al Handbook DENORAL EXPEX Guidon Hpp Research HPP Assoc Med Info Sci MOLGEN Oncocin Protein Structure RX Stanford Pilots Stanford Assoc Staff Staff Assoc Figure 12: UU Progress - Resource Operations Statistics National AIM (24.4% Total) cs Stanford (59.6% Total) TT | Staff (16.0% Total) a 5 10 15 20 25 Percent of Total CPU Used 37 Cumulative CPU Usage Histogram by Project and Community E. A. Feigenbaum Progress - Resource Operations Statistics 5P41 RRO00785-11 Resource Use by Individual Project - 5/83 through 4/84 National AIM Community 1) ACT Project “Acquisition of Cognitive Procedures" John R. Anderson, Ph.D. Carnegie-Mellon Univ. NSF IST-80-15357 2/81-2/84 $186,000 2) CADUCEUS “Clinical Decision Systems Research Resource" Jack D. Myers, M.D. Harry E. Pople, Jr., Ph.D. University of Pittsburgh NIH RR-01101-07 7/80-6/85 $1,607,717 7/83-6/84 $369,484 NLM LM03710-04 7/80-6/85 $817,884 7/83-6/84 $196,710 NLM New Invest LM03889-02 Gordon E. Banks, M.D. 4/82-3/85 $107,675 4/83-3/84 $35,975 4/84-3/85 $35,975 3) CLIPR Project “Hierarchical Models of Human Cognition" Walter Kintsch, Ph.D. Peter G. Polson, Ph.D. University of Colorado NIMH MH-15872-14-16 (Kintsch) 7/81-6/84 $281,085 7/83-5/84 $69,878 NSF (Kintsch) 8/'83-7/86 $200,000 IBM (Polson) David Kieras University of Arizona 1/82-12/84 $364,000 1/84-12/84 $145,000 E. A. Feigenbaum CPU Connect (Hours) (Hours) 0.37 33.88 58.15 895.52 1.38 209.34 © 38 File Space (**)} (Pages) 2866 6852 750 5P41 RROO785-11 4) 5) 8) PUFF-VM Project 0.65 61.20 "Biomedical Knowledge Engineering in Clinical Medicine" John J. Osborn, M.D. Med. Research Inst., San Francisco Edward H. Shortliffe, M.D.,Ph.D. Stanford University Johnson & Johnson 1 year $50,000 (*) SECS Project 264.61 9877.34 "Simulation & Evaluation of Chemical Synthesis" W. Todd Wipke, Ph.D. U. California, Santa Cruz NIHEHS ES02845-02 4/82-3/85 $257,801 4/84-3/85 $89,140 Evans & Sutherland Corp. Equipment gift Value $95,000 Stauffer Chemical Co. $6,000 SOLVER Project 5.76 356.23 "Problem Solving Expertise" Paul E. Johnson, Ph.D. William B. Thompson, Ph.D. Control Data Corp. (Johnson) 1983-85 $90,000 Microelect. and Info. Ctr. Univ. of MN (Plus Two Colleagues) 1984-1987 $800,000 39 Progress - Resource Operations Statistics 303 10500 492 E. A. Feigenbaum Progress - Resource Operations Statistics 7) *** (Rutgers-AIM] *** Rutgers Research Resource “Computers in Biomedicine" Saul Amarel, D.Sc. Casimir Kulikowski, Ph.D. Sholom Weiss, Ph.D Rutgers U., New Brunswick NIH RR-00643-12 (Amarel, Kulikowski) 12/82-11/83 $405,304 NIH RR-02230-01 (Kulikowski, Weiss) 12/83-11/87 $3,198,075 12/83-11/84 $989,276 8) AIM Pilot Projects 9) AIM Administration 10) AIM Users Community Totals E. A. Feigenbaum 40 0.52 65.85 38.59 17654.52 5P41 RROO785-11 1117 5P41 RROO785-11 Stan ford Community 1) AGE Project (Core) “Attempt to Generalize“ Edward A. Feigenbaum, Ph.D. Dept. Computer Science ARPA MDA903-80-C-0107 (***) (partial support) 2) AI Handbook Project (Core) Edward A. Feigenbaum, Ph.D. Dept. Computer Science ARPA MDA903-80-C-0107 (**) (partial support) 3) DENDRAL Project “Resource Related Research: Carl Djerassi, Ph.D. Dennis H. Smith, Ph.D. Dept. Chemistry NIH RR-00612-13 5/82-4/83 $170,710 4) 5) Computers in Chemistry" EXPEX Project "Expert Explanation" Edward H. Shortliffe, M. Dept. Medicine ONR NR 049-479 1/81-12/83 $456,622 ONR NRO0O49-479 D.,Ph.D. Michael Genesereth 1/84-12/86 $312,070 NSF [ST83-12148 Bruce G. Buchanan $330,000 (*) $99,410 (*) 3/84-2/87 3/84-2/85 GUIDON-NEOMYCIN Project “Exploration of Tutoring & Problern-solving Strategies" Bruce G. Buchanan, Ph.D. William J. Clancey, Ph.D. Dept. Computer Science ONR/ARI NO00014-79-C-0302 3/79-3/85 $683,892 Progress - Resource Operations Statistics CPU (Hours) 11.80 11.03 3.72 53.75 45.44 41 Connect (Hours) 845.30 980.94 183.81 2391.40 4418.68 File Space (Pages) 4076 4425 2980 4920 5967 E. A. Feigenbaum Progress - Resource Operations Statistics 6) 7) 8) MOLGEN Project 106.92 "Applications of Artificial Intelligence to Molecular Biology" Edward A. Feigenbaum, Ph.D. Peter Friedland, Ph.D. Charles Yanofsky, Ph.D. Depts. Computer Science/ Biology NSF MCS-8310236 (Feigenbaum, Yanofsky) 11/83-10/84 $139,215 (*) ONCOCIN Project 239.97 “Knowledge Engineering for Med. Consultation" Edward H. Shortliffe, M.D.,Ph.D. Dept. Medicine NLM LM-03395 (Shortliffe/ONCOCIN) Edward A. Feigenbaum, Ph.D. 7/79-6/84 $497,420 7/83-6/84 $95,424 NLM LM-00048 7/79-6/84 $196,425 7/33-6/84 $39,502 ONR NR 049-479 1/81-12/83 $456,622 (*) NIH RR-01613 7/83-6/86 $624,455 7/83-6/84 $220,371 NLM LM-04136 8/83-7/86 $211,851 8/83-7/84 $60,517 H.J. Kaiser Family Fdn. 7/83-6/86 $150,000 7,/83-6/84 $50,000 ONR N00014-81-K-0004 Michael R. Genesereth (Shortliffe) 1/84-12/86 $512,070 (*) NSF IST83-12148 Bruce G. Buchanan (Shortliffe) 3/84-2/87 $330,000 (*) 3/84-2/85 $99,410 (*) PROTKEKIN Project 4.79 “Heuristic Comp. Applied to Prot. Crystallog." Edward A. Feigenbaum, Ph.D. Dept. Computer Science NSF MCS-81-17330 1/82-1/83 $28,976 E. A. Feigenbaum 42 7734.34 14404.62 635.43 5P41 RROO785-11 10448 14389 1296 5P41 RROO785-11 9) RADIX Project Progress - Resource Operations Statistics 79.44 "Deriving Medical Knowledge from Time- Oriented Clinical Databases“ Robert L. Blum, M.D. Gio C.M. Wiederhold, Ph.D. Depts. Computer Science/ Electrical Engrg. NSF IST-8317858 (Blum) 3/84-3/86 $89597 (*) NLM (Wiederhold) 5/84-11/86 $291,192 10) Stanford Pilot Projects 11) HPP Core AI Research 12) HPP Associates 13) Stanford Associates 14) Medical Information Sciences Community Totals SUMEX Staff 1} Staff 2) System Associates Community Totals System Operations 1) Operations Resource Totals (*) Award includes indirect costs. 61.55 383.07 57.37 27.01 5.62 1091.46 CPU (Hours) 288.21 16.65 CPU 530.54 2382.43 3140.27 4115.02 29073.96 1600.31 1016.59 1315.64 71856.29 Connect (Hours) 17591.82 1983.43 19575.25 Connect (Hours) 67375.43 176461.50 8777 6097 42202 2997 1681 587 110842 File Space (Pages) 23292 7847 31139 File Space (Pages) 167863 345520 (**) Supported by a larger ARPA contract MDA-903-80-C-0107 awarded to the Stanford Computer Science Department: 43 E. A. Feigenbaum Progress - Resource Operations Statistics 5P41 RROO785-11 System Reliability System reliability for the DECsystem 2060 has been much better than with our previous KI10 system. We have had very few periods of particular hardware or software problems. The data below covers the entire period in which the SUMEX-AIM community has used the 2060. The actual downtime was rounded to the nearest hour. 7 18 1 Feb Mar Apr Table 1: System Downtime Hours per Month - February 83 through Apr 83 11 11 1 2 6 0 11 15 26 13 16 28 May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr Table 2: System Downtime Hours per Month - May 83 through Apr 84 Reporting period : 462 days, 23 hours, 41 minutes, and 42 seconds Total Up Time : 454 days, 5 hours, 16 minutes, and 57 seconds PM Downtime : 1 days, 14 hours, 2 minutes, and 55 seconds Actual Downtime : 7 days, 4 hours, 21 minutes, and 50 seconds Total Downtime : & days, 18 hours, 24 minutes, and 45 seconds Mtbf : 2 days, 16 hours, 30 minutes, and 16 seconds Uptime Percentage : 98.45 Network Usage Statistics The plots in Figure 13 and Figure 14 show the monthly network terminal connect time for the TYMNET and the INTERNET usage. The INTERNET is a broader term for what was previously referred to as Arpanet usage. Since many vendors now support the INTERNET protocols (IP/TCP) in addition to the Arpanet, which converted to IP/TCP in January of 1983, it is no longer possible to distinguish between Arpanet usage and Internet usage on our 2060 system. E. A. Feigenbaum 44 5P41 RROO785-11 Progress - Resource Operations Statistics 14007 TVMNET Connect Time Hours/Month 1200 T 1000 800} 600 400 200F oO 2 1 4 L 4 i 2. 2 1 L 4 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Figure 13: TYMNET Terminal Connect Time 12007 aRpaAnet Connect Time Hours/Month 1000 800+ 600} 400; 200} 1 4 1 LL L 1 oO i A L 1 i 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Figure 14; ARPANET Terminal Connect Time 45 E. A. Feigenbaum Progress - SUMEX Staff Publications 1.A.2.8. SUMEX Staff Publications program developments. reports (see Section II on page 69). The following are publications for the SUMEX staff and include papers describing the SUMEX-AIM resource and on-going research as well as documentation of system and Many of the publications documenting SUMEX-AIM community research are from the individual collaborative projects and are detailed in their respective Publications for the AGE and AI Handbook core research projects are given there. 1. Carhart, R.E., Johnson, S.M., Smith, D.H., Buchanan, B.G., Dromey, R.G., and Lederberg, J., Networking and a Collaborative Research Community: A Case Study Using the DENDRAL Programs. IN P.Lykos (Ed.), COMPUTER NETWORKING AND CHEMISTRY, ACS Symposium Series, No. 19, 1975. . Levinthal, E.C., Carhart, R.E., Johnson, S.M., and Lederberg, J.: When Computers Talk to Computers. Industrial Research, November, 1975. . Wilcox, C.R., MAINSAIL - A Machine-Independent Programming System. Proc. DECUS Symposium 2(4), Spring, 1976. . Wilcox, C.R.: The MAINSAIL Project: Developing Tools for Software Portabiltty. Proc. SCAMC, October, 1977, pp. 76-83. . Lederberg, J.L.: Digital Communications and the Conduct of Science: The New Literacy. Proc. IEEE 66(11), November, 1978. . Wilcox, C.R., Jirak, G.A., and Dageforde, M.L.: MAINSAIL - Language Manual. Stanford University Computer Science Report STAN-CS-80-791, 1980. . Wilcox, C.R., Jirak, G.A., and Dageforde, M.L.: MAINSAIL - Implementation Overview. Stanford University Computer Science Report STAN-CS-80-792, 1980. In addition, a substantial continuing effort has gone into developing, upgrading, and extending documentation about the SUMEX-AIM resource. These efforts include user guides, help files, and introductory notes, an ARPANET Resource Handbook entry, and policy guidelines. E. A. Feigenbaum 46 5P41 RRO0785-11 5P41 RROO785-11 Progress - Future Plans 1.A.2.9. Future Plans Our plans for the next grant year are based on the Council-approved plans for our 5-year renewal that began in August, 1980. In addition to the specific plans for the next grant year, we include a summary of the overall objectives for this 5-year period to serve as a background. Near- and long-term objectives and plans for individual collaborative projects are discussed in Section II beginning on page 69. Overall Goals The goals of the SUMEX-AIM resource are long-term in supporting basic research in artificial intelligence, applying these techniques to a broad range of biomedical problems, experimenting with communication technologies to promote scientific interchange, and developing better tools and facilities to carry on this research. Just as the tone of our renewal proposal derives from the continuing long-term research objectives of the SUMEX-AIM community, our approach derives from the methods and philosophy already established for the resource. We will continue to develop useful knowledge-based software tools for biomedical research based on innovative, yet accessible computing technologies. For us it is important to make systems that work and are exportable. Hence, our approach is to integrate available state-of-the-art hardware technology as a basis for the underlying software research and development necessary to support the AI work. SUMEX-AIM will retain its broad community orientation in choosing and implementing its resources. We will draw upon the expertise of on-going research efforts where possible and build on these where extensions or innovations are necessary. This orientation has proved to be an effective way to build the current facility and community. We have built ties to a broad computer science community; have brought the results of their work to the AIM users; and have exported results of our own work. This broader community is particularly active in developing technological tools in the form of new machine architectures, language support, and interactive modalities. Toward a More Distributed Resource The initial model for SUMEX as a centralized resource was based on the high cost of powerful computing facilities, which were not readily duplicated. This role is evolving, though, with the introduction of more compact and inexpensive computing technology. Our future goals are guided by community needs for more computing capacity and improved tools to build more effective expert systems, and to test operational versions of AI programs in real-world settings. in order to meet these needs, we must take advantage of a range of newly-developing machine architectures and systems. As a result, SUMEX- AIM will become a more distributed community resource with heterogeneous computing facilities tethered to each other through communications media. Many of these machines will be located physically near the projects or biomedical scientists using them. The Continuing Role of SUMEX-Central Even with more distributed computing resources, the central resource will continue to play an important role as a communications crossroad, as a research group devoted to integrating the new software and hardware technologies to meet the needs of medical AI applications, as a spawning ground for new application projects, and as a base for local AI projects.