LISP Activities in Japan
Proceedings of the Sixth International Joint Conference on Artificial Intelligence. .
© 1979 International Joint Conferences on Artificial Intelligence
Abstract
Lisp is one of the major programming languages for AI research. This paper surveys the current status of Lisp activities in Japan. First, Lisp systems developed in Japan on commercial machines are reviewed . Next, projects for Lisp machines are reported with commentary. The current problems and future prospects for the Lisp developments are discussed. We propose that the general educational activities and developments of more powerful portable systems for this language are necessary.
- J.Allen: Anatomy of Lisp, M c<: rawllill, 1978.
- I. Takeuchi ed.: Report of the 2nd Lisp contest, Working paper of Kigoshori-Kenkyuukai, IPSJ, Aug. 1978. The report of the 1st contest is included in the Kigoshori Symposium Report, IPSJ, July 1974.
- The followings are on Li s p machin es . Except [3,6,9,) are in Japanese.
- E.Goto et al.: Report of th<' FI.A'I'S pro- ject Vol.!, The Institute of Physi ca l and Chemical Research, Oct. 1978.
- Y.Hibino: A Parallel garbage collection algorithm and its application to Lisp, IECE Working paper, Nov. 1978.
- H. Ida et al.: Performanc e evaluation of ALPS/I Lisp machine, IPS.J Working paper, Oct. 1977.
- Some Topics fr om ALPS, IPS.J Working paper, Mar . 1979.
- Y.Kaneda and K.Taki: An Experimen tal l.isp Hachine, in this 6th I.JCA l Proceedin gs .
- M.Nagao et al.: Design and implementation of Lisp-machine NK-3, Kyoto Univ., 1977.
- Lisp Machine NK3 and It s Performance Evaluation, IPS.J \Jorking papc>r, March 1979. also in this 6th I.JCAI Proccpdings.
- T.Shimada ct al.: A Lisp Ma c hin e and its evaluation, Trans. TECE, Vol.59-D, No.6, .June 1976 pp.406-413.
- K.Taki et al.: lmplenent ation of Lisp Machine, IPSJ Working paper, Aug. 1978.
- Microprogramme d Interpreter of the Experimental LISP Machin e , lECE Workin g paper, Nov. 1978.
- An Experimental Lisp M achine and its Evaluation, IPSJ \-larkin g paper, March 1979 .
- T. Usuki et al.: Implementation of LISP machine on multi-proces so r, IPS. I Conference Proceedings, 1978, pp. 27-2 R.
- LISP Machin e Implementation on Multi-microproc esso r System, paper, Nov. 1978.
- Y.Ya maguchi and '!'.Shimada: Measurements of LISP Programs on Hac hine, Trans. IECE, Vo l.61-ll, 1978. IF:CE Working llynamLc a Virtual No.8, Aug.
References were extracted automatically from the PDF and may contain errors.
Full text extracted automatically from the PDF; figures, tables and equations may differ. The PDF is the version of record.
1. LISP SYSTEMS ON COMMERCIAL ~1ACHINES IN JAPAN
No reports have been available on Japanese Lisp systems, except for the reports on the Lisp contests in Japan [2], one of which was held in April 1978 by the working group on symbol manipulation under Information Processing Society of Japan (Kigoshori Kenkyuukai). In this contest about 20 Lisp systems participated including 3 overseas systems. The primary purpose of the contest was to know the current status of the art of Lisp systems developed in Japan. However, there existed some problems, as Eric Sandewall pointed out [2], that the benchmark programs for the contest were not appropriate especially from the viewpoint of AI
applications, and that some systems were not included in the contest.
The reported items are as follows: system's name, authors, their institution, dates for first and current version installed, primary purposes, its resemblance, its service, the number of users, the number of Lisp system functions, its top level function, its data types available, the number of cells, stack, compiler status, program size, its hardware systems, memory size, the number of general registers , CPU cycle time, time for add (memory to register), operating systems, version-up schedule and other information. The report also contains the execution time of several test programs, which are cited at many places for performance evaluation.
Tables 1,2,3, and 4 contain the list of the ex isting Lisp systems in Japan. Table 1 lists the portable systems, Table 2 big and medium, machines, Table 3 for small machines and Table 4 shows the micro-processor based svstcms .
The Tables contain only the system's name, au thors, their inst~tutions, hardware systems , and the main features. The import ed Lisp systems such as Interlisp, UCL-Lisp, Standerd-Lisp and Maclisp are not included.
The followin3s should be noted:
- Host of the systems have been recently im plemented (during 1975- 1978.)
- Portable Lisp systems are implemented on
Tables for Lisp Systems on Commercial Machines in Japan
I T.Chikayana IUniv. of TokyojTNTEL 8080
| System name | Author | Institution | Hardware | Features |
|---|---|---|---|---|
| Table 1. Portable Lisp System | ||||
| IILISP LTSP/P TI.ICS | Y.Kanada et al. T.Chlkayama M.Suzuki et al. | Univ. of Tokyo Univ. of Tokyo Tohoku Univ. | HITAC-8800/8700 FACOM-230/75 IBM-370/168 FACOM-230-38 ACOS-6 | Written in Fortran Written in Pascal Written in Fortran |
| 1able 2. L1sp Systems on B1g Mach1nes | ||||
| LISP 1. 5 OLISP PETL TOSBAC-5600 LISP 1. 9 VS-DI.I SP 1. 5 | H.Kumura H.Yasui et al. M.Tsukamoto T.Kurokawa H. Imura | Okayama Univ. Osaka Univ. ETL Toshiba Co. Doshisha Univ. | MELCOM COSM0-700 ACOS-77 System 800 System 900 HELCOM COSH0-700II TOSBAC-5600 ACOS-77 HITAC 8350 | TSS TSS TSS |
| lable 3 Ltsp Systems on Small Mach1nes . | ||||
| FLISP IIITAC-10 LISP KL!SP-ll LLPCI Ll.PX LISP 1.7 LISP 38 NIT LISP-U | N.Abe et al. Y.Nakamura M.Nakanishi et al. I. Takeuchi & II. Okuno K.Hiramatsu et al. Y.Ohta T.Niwa et al. fable 4. | Osaka Univ. Keio Univ. Keio Univ. ECL of NTT Kyoto Univ. Nagoya Univ. NIT L1sp Systems on | IIP2108A HITAC--10 PDP-ll PDP-ll/55 FACOM U-200 FACOM 230-38 FACOM U-200 H1cro-computers | Hardware modified Hardware modified Parallel 4-Cell Hardware modified |
I
mini-computers and is used for large natural lansuage processing programs. ETL(Electro technical Laboratory) [10,16] is trying to make a Lisp machine on a chip. (It will really be on a few chips.) They use LSI microprocessors produced by a national project, PIPS. (Pattern Information Processing System)
A Kobe machine[7,11,12,13,] is rather in expen sive but is very fast. Its interpreter runs as fast as the compiled Lisp programs on a big ma chine, although its address space is rather limited (16 bit). Keio Hachine[l4,15] is a Lisp system on an experimental multi-microprocessors.
3. PROBLEMS AND FUTURE PROSPECTS
The author would like to point out the follow ing two problems: one is for education and the other is for computing facility.
- Education As Lisp is (seemed to be) tied up with the AI programs, those who do not study AI, have a tendency to neglect Lisp for the
ILtSPM
Fortran or on Pascal.
- Many systems arc implemented on small com puters, some with hardware modifications for commercial machines.
2. LISP MACHINE PROJECTS IN JAPAN
At the time of April 1979, there are 6 Lisp ma chine projects reported in Japan. The imple mentors, machines, principal purposes and progress atatus are listed in Table 5.
There arc many kinds of approaches for Lisp ma chines projects: FLATS by Goto[3] is a special purpose machine (Formula manipulations). Fortran and Algol can al~o run on that machine. ECL(Electro Communication Lab. of NTT) machine [4] is made to experiment the parallel garbage collection. Although the feasible result was not obtained, it has provided Lisp engineers wllh the valuablt· information. ALPS/L[5,6] is ont· of the (·;nlit·st Lisp machine in the world, and the sip,nificant revision is now considered. NK3 made by Kyoto Univ. [8,9] utilizes 32-bit programming instruction. From the author's viewpoint, however, we should promote the Lisp programming instructions not only for the wide activities of AI, but for the better under standing of the computer science. We also need a good organization for Lisp instructions.
| Implementor (Institution) | Machine | Purpose | Status (April 1979) |
|---|---|---|---|
| E.Goto et al. (Institute of Physical and Chemical Research) | FLATS | Formula manipulation | Design |
| Y.llibino (ECL of NTT) | Tosbac-40L | Parallel GC | |
| H. Ida et al. (Aoyama Gakuin Univ.) | ALPS/I using Intel 8080 | General (esp. formula manipulation) | Working from 1977 |
| ~!.Nagao et al. (Kyoto Univ.) | NK-3 using Interdata 8/32 E~·- - | AI esp. for natural language processing -~--~-~-~~~~~~-·~~~~!~P~I~P~S~~~~~~~~~~~~~~~~~~~t~D~=e-~s~i~-g~~- | Working/Data Gathered |
| ~; -~~~ - ~- Y.~~~~g~~hi -~l~i~:~~--- | ~f~,TI~i~.C~ . | - ~F~i~n~i~s~h~e~d~~~~ | |
| ~K.Ta-ki et al. (Kobe Univ.) | using Am 2903 | General | Working/Data Gathered |
| r.r.·us~kie-t al. (Keio Univ.) | Multi-micro conputers | General | ~Iorking |
- ii) Computing facility-Interactive programming and cooperative program development are two vital elements for the present software devel opment, especially on AI. Utilization of time -sharing services have been the easiest way to experience them. But TSS are rare in Japan even today. In fact, some of the Lisp machine projects are motivated in order to have an in teractive facility for Lisp. Even if you are doing research by yourself in your own style, you should promote to have a center machine with time-sharing services and should provide a network facility to communicate with others.
The author would like to discuss some future prospects for Lisp developments in Japan.
- Although many Lisp machines have been developed good portable Lisp systems are in dispensable to meet a wide variety of users and their applications.
- ii) ~1emory-neck problems will be solved by the recent 32-bit architectures. A portable Lisp on 32-bits hardware(eg. Interdata 8/32, Tosbac 7/70, and DEC VAX-11) and a special purpose Lisp machine will be good competitors.
- iii) Database facilities are necessary for large AT applications.
- iv) Parallel processing facilities are needed, since the techniques for parallel processings are now accumulated. The AI applications such as many worlds hypothesis are utilizing paral lelisms.
- For the AI applications in Japan, most users need the Japanese language interfaces. Thus, Kana and Kan.ii characters should be handled by Lisp. The general pattern handling facilities are also preferable .
- vi) Modularity of Lisp programs encourages to involv~ special features useful for AI, eg. Pattern matching, cont<'Xl swit ching, etc.
- v il) R~cently Lisp has becoml' more popular than before. It is studied now in the software engint-er l.ng fields as well as AI and the mathe matical computer science. However, it is still a research language(at least they say), and its application and education are rather limited.
As AI applications are widely recongni zed as feasible, Lisp implementors should not<' the practical aspects of the Lisp languap,e. From the author's viewpoint, Lisp may he regarded as a system language for high-level application programs.
@inproceedings{ijcai1979p112,
title = {LISP Activities in Japan},
author = {Kurokawa, Toshiaki},
booktitle = {Proceedings of the Sixth International Joint Conference on
Artificial Intelligence, {IJCAI-79}},
publisher = {International Joint Conferences on Artificial Intelligence},
editor = {Bruce G. Buchanan},
year = {1979},
month = {8},
url = {https://ijcai.org/proceedings/1979/112},
}
TY - CPAPER TI - LISP Activities in Japan AU - Kurokawa, Toshiaki T2 - Proceedings of the Sixth International Joint Conference on Artificial Intelligence, IJCAI-79 PB - International Joint Conferences on Artificial Intelligence PY - 1979 UR - https://ijcai.org/proceedings/1979/112 ER -Copied
Machine-readable versions of this paper, for tools and AI agents:
