ICOT FREE-SOFTWARE News Letter No.7

                                          Jan 14, 1994

[Table of Contents]

Newly Released ICOT Free Software
Report on Access to IFS
User's Group
About Common ESP
Contacts

[Newly Released ICOT Free Software]

To date, we have released 77 programs of ICOT Free Software (IFS), as was reported in previous issues of ICOT Free Software Newsletters. We are now preparing to release seven further programs as IFS, the KL1 including the KLIC Version system.

The KLIC system is a programming environment for parallel logic programming language for use on Unix-based general-purpose machines. KLIC Version 1 is the first practical system to include a debugging tool and standard manuals. However, its use is limited to single-processor systems such as standard workstations. KL1 programs are therefore executed sequentially as pseudo parallel programs. KLIC Version 2 will support real parallel machines. This version will be released in the second quarter of 1994. The experimental version of the KLIC system, released about six months ago, should be replaced with KLIC Version 1. We recommend that you use this latest version for learning parallel language and discovering the merits of high-level parallel programming.

Using KLIC Version 1, it is now possible for us to port other KL1 programs to Unix-based machines. We are currently preparing to the KLIC version of the knowledge representation language Quixote. The Quixote KLIC version is written in KL1 and is intended to be ported using KLIC Version 1. As the Quixote coding is rather large, we are preparing a condensed version to allow users to quickly try out a deductive and object-oriented programming using Quixote. This will be named micro-Quixote and will provide only representative Quixote functions. However, it is fully written in C and is small in size. We feel it is very well-suited to educational applications.

A similar condensed version is being prepared for the theorem prover MGTP. Two versions of MGTP, both of which are written in Prolog, are being prepared. In addition to these programs, two convenient tools for genome information processing will be released. Currently, the archive files of KLIC, and micro-Quixote have already been installed on the ftp server machine.

1) Newly released ICOT Free Software


  The following lists the new programs:

Symbol Processing: 1
-- ------------------------------------------------------------------
78  KLIC Version 1: An implementation of KL1 for general-purpose
    computers

Knowledge Representation, Knowledge-base, Constraint: 2
-- ------------------------------------------------------------------
79  Knowledge Representation Language: Quixote KLIC Version
80  Knowledge Representation Language: micro-Quixote

Problem Solving, Meta-reasoning, Theorem Proving: 2
-- ------------------------------------------------------------------
81  Parallel Theorem Prover: MGTP/G Prolog version
82  Parallel Theorem Prover: MGTP/N Prolog version

Experimental Application: 2
-- ------------------------------------------------------------------
83  Protein Motif Knowledge-Base and Retriever Based on Quixote
84  Protein Structure Prediction Based on Multi-Level Description

2) Detailed Description

1. KLIC Version 1: A KL1 implementation for general-purpose computers: This system is able to execute programs written in the KL1 parallel logic programming language on commercially available unix machines. Note that this system does not support all of the functions provided by the KL1 language processor running on PIM, instead being intended to run on a single processor system in pseudo parallel mode.; For instance, the functions provided by PIMOS are not supported by the KLIC Version 1. Using KLIC, you can execute KL1 programs on your unix workstations, and furthermore, you can easily combine KL1 programs with programs written in other languages, such as C.

This program consists of the following:

A KL1 to C compiler. Currently, the compiler is written in Prolog. Therefore, you need some kind of Prolog system on your machine.
Some PIMOS functions, including those for debugging, are provided as a run time library of C code. They are linked with your program if you wish to use these functions.

The program has the following features:

You can execute KL1 programs on unix workstations and personal computers supporting a C environment.
The KLIC system compiler translates KL1 programs into equivalent C programs. Then, the resulting C program is compiled by a C compiler and linked with run time libraries to produce a code that can be executed under your Unix operating system.
The KLIC system generates C code instead of native machine code. Thus, it is not affected by differences in machine architectures.
Although the KLIC system uses the C as its intermediate code and C compiler, it attains high performance and a compact object code, comparable to when we employ a compiler that generates native machine code.

We released "Portable KL1 processing system: experimental version June of 1993. The experimental version should be replaced with KLIC Version 1.

The archive file of the KLIC system has been installed in the following ftp server directory:

/ifs/symbolic-proc/unix/klic

2. Knowledge Representation Language: Quixote KLIC Version: This system is a variant of the Quixote system that was released in 1992 as IFS. It is a rewrite of Quixote that allows to be smoothly ported to a Unix environment using the KLIC system.; To install this system, the KLIC system and a C compiler are required, in addition to GNU-Emacs and an X-window system interface.

This system has the following features:

It consist of two parts; one is a client part providing the user interface and I/O, written in C. Another is a server part, providing a deductive database and written in KL1.
The knowledge representation language Quixote is based on a "Deductive Object-Oriented Database" which combines the concepts of logic, object-orientation, and constraints.
Compared to other knowledge representation languages, it provides more natural knowledge description in a more compact program size.

Currently, we are still preparing for the release of this system. We expect it to be ready in a couple of months.
3. Knowledge Representation Language: micro-Quixote: micro-Quixote is a condensed version of the Quixote system that was released as IFS in 1992. To make it compact, only representative functions are provided and it is implemented in C to improve portability.; Some Emacs programs are needed for its user interface. However, it will run on most workstations and personal computers that support a C environment.; The following functions have either been omitted or reduced in scope, relative to the original Quixote system.

Merging of more than two solutions
Database updating
Some property inheritance functions

Even with this condensed version, you can make use of inheritance based on a lattice relation, constraint based on a subsumption relation, hypothesis generation, hypothetical reasoning, and a module facility.

The archive file of the micro-Quixote system has been installed in the following ftp server directory:

/ifs/kbms-clp/unix/

4. Parallel Theorem Prover: MGTP/G Prolog version

This system is a variant of the MGTP parallel theorem prover that was released as IFS in 1992 and which produced very high levels of performance on the parallel inference machine PIM.

To provide a condensed version of MGTP that can easily be executed on a normal workstations, all MGTP systems are rewritten using Prolog. Thus, this condensed version is not capable of parallel processing and thus can not attain high performance. However, you can use the prover for educational purposes or as an introductory step prior to moving to the parallel version.

This MGTP/G prover is called the "MGTP ground version" since the mathematical system handled by this system uses axioms, all of which can be represented by ground terms. In this system, theorems to be proved are compiled into Prolog predicates, and proof is done by executing Prolog programs using forward reasoning.

We are now preparing to release this system as IFS. We expect its release in the very near future.

5. Parallel Theorem Prover: MGTP/N Prolog version

This prover is another variant of the parallel theorem prover MGTP, being called "MGTP non ground version." It is written in Prolog and is a condensed version of the original MGTP that can be used on normal workstations.

This MGTP/N prover has some functions that were not featured by the original MGTP system and the previous MGTP/G prover. It can prove theorems in mathematical systems whose axioms contain variables. In this system, theorems to be proved are compiled into Prolog predicates, and proof is carried by executing a Prolog program.

We are now preparing to release this system as IFS. We expect its release in the very near future.

6. Protein Motif Knowledge-Base and Retriever Based on Quixote

Both of these, the first being an experimental knowledge base, and the second its retriever, are implemented using the knowledge representation language Quixote. They can be used to extract and analyze DNA sequence patterns, or protein sequences when used together with the "Multiple Sequence Alignment by Parallel Iterative Aligner" and the "Intelligent Refiner for Multiple Sequence Alignment", both of which were released as IFS in March of 1993.

In some DNA sequences and some protein sequences, there are parts called "motifs" that are closely related to important biological features. We can find some special correlations between some motifs. We can describe the characteristics of these motifs as well as their correlations in Quixote, and store them in a knowledge base. These knowledge descriptions can be useful for producing better analyses of DNA and protein sequences to detect important biological features.

We are currently preparing to release this system as IFS.

7. Protein Structure Prediction System based on Multi-Level Descriptions of Protein Sequences

This system is intended for the prediction of the higher order structures of proteins. It is important to know the functions of higher order structures of proteins. Although several methods of predicting higher order structures have been proposed their prediction accuracies have not been satisfactory, hence more accurate prediction methods are required.

We analyzed proteins whose structures are well known, and set up sample relations between specific protein sequences and their expected structures when the size of the proteins is changed. This system uses this table of hierarchical sample relations to predict unknown protein structures. It gives us a more accurate prediction than is possible with existing systems for some proteins.

To execute this system, the KL1 programming environment, running on the parallel inference machine PIM, a C processor, and a Silicon Graphics workstation are required.

We are currently preparing to release this system as IFS.

For detail of these seven programs, refer to the "ICOT Free Software Catalogue: Supplementary". The "ICOT Free Software Catalogue" has recently been published. Its includes the contents of all catalogues that we have published to date. Anyone wishing to obtain these catalogues is invited to contact the IFS desk via e-mail, mail or fax at the address given at the end of this newsletter.

[Report on Access to IFS]


   As of January 1, 1994, access to IFS can be summed up as follows:
  
   =====================================================
                  Number of Persons     Number of Files
    Country       Accessing             Transferred
   =====================================================
    Japan                  502              3,172
    U.S.A.                 385              2,658
    France                  40                381
    German                  68                372
    South Korea             15                317
    U.K.                    70                277
    Canada                  56                230
    Australia               47                138
    Sweden                  27                110
    Netherlands             22                 62
    Others*                204                722
   =====================================================
    Total                1,436              8,439
   =====================================================

*"Others" include Portugal, Taiwan, Austria, Finland, Italy, Singapore, South Africa, Belgium, Switzerland, Hong Kong, Denmark, Spain, Norway, Turkey, Brazil, Thailand, Greece, New Zealand, Poland, Mexico, Czech Republic, Israel, India, Hungary, Venezuela, Soviet Union, Latvia, Lithuania, Slovakia, Slovenia, and Cyprus. The total number of countries and regions is 41.

The number of persons accessing in the above table include those who transferred only "readme" files that are not archive files for software. Excluding readme file accesses gives the following figures.

   =====================================================
                  Number of Persons     Number of Files
    Country       Accessing             Transferred
   =====================================================
    Japan                  371              1,638
    U.S.A.                 226              1,429
    South Korea             10                239
    France                  29                229
    German                  44                209
    U.K.                    40                113
    Canada                  29                 95
    Sweden                  10                 48
    Portugal                 4                 36
    Australia               18                 27
    Others                 118                284
   =====================================================
    Total                  899              4,347
   =====================================================

The above figures show that more than 60% of all the persons accessing the server actually transferred some software files, with about 52% of the total transferred files being such software files.

Among the 899 persons who transferred files, 597 were from Japan or the U.S. By looking at the organizations to which these 597 persons belong, we can draw up the following table that shows the types of organizations to which these persons belong.

  ==============================================================
                              Japan                U.S.A.
                        ----------------------------------------
                         No. of    No. of     No. of    No. of
   Organization          Persons    Files     Persons    Files
  ==============================================================
   Educational(ac/edu)      300    1,294         132      705
   Commercial(co/com)        46      279          74      577
   Nonprofit(or/org)         16       37           7       71
   Government(go/gov)         7       21          10       67
   Others                     2        7           3        9
  ==============================================================
    Total                   371    1,638         226    1,429
  ==============================================================

[User's Group]

Anyone interested in organizing a user's group to carry out research on any particular item of ICOT Free Software, for the purposes of revision or improvement, is invited to contact the IFS-desk via e-mail at the address given at the end of this newsletter.

We hope to feature some of your proposals in the next issue of the newsletter. Reactions to your proposals shall be forwarded to you.

Revised programs can be stored on the ftp server at ICOT, if you feel that your revision would be useful to other users.

1) User's Group for cu-Prolog

Following its release as ICOT Free Software, the constraint logic programming language cu-Prolog, including Prof. Shirai's Macintosh/MS-DOS version, has been used by many students, engineers, and CLP/NLP researchers.

Now, encourage the exchange of comments and information relating to cu-Prolog, constraint-based grammar formalisms, and CLP, we have organized a cu-Prolog user's group. The address is

cup@icot.or.jp

To join the mailing list, send e-mail to

cup-request@icot.or.jp

Currently, 15 domestic and 11 overseas researchers have joined the ML.

[About Common ESP]

Some ICOT Free Software programs can be executed under Common ESP(CESP). CESP is not part of the ICOT Free Software, instead being available from MITSUBISHI ELECTRIC COMPUTER SYSTEMS(TOKYO) CORP. (MTC). The address of MTC is as follows. To contact MTC with any questions related to Common ESP, use the following address.


        2-nd Application Systems Develpment Dept.
        Software Products Promotion Group

        87-1, Kawakamicho, Totsukaku, Yokohama,  Kanagawa, 244 Japan

        TEL:  +81-45-825-5679(Japanese Only)
        FAX:  +81-45-825-5822

[Contacts]

For information on IFS, use e-mail to access

ifs@icot.or.jp

If you receive a paper edition of this newsletter, please let us know your e-mail address and we will send you the electronic edition. If you do not have an e-mail facility, contact the address below.

All available IFS is listed in the "ICOT Free Software Catalogue" and "ICOT Free Software Catalogue II" . If you do not have a copy of either catalog, please let the IFS-desk have postal address and we shall arrange to send a copy to you.

If any of your colleagues or acquaintances are interested in IFS, let us know their names and both their e-mail and postal addresses, and we shall send them both the newsletter and catalog.


        ICOT Free Software desk

        Institute for New Generation Computer Technology
        21st Floor, Mita Kokusai Bldg.
        4-28, Mita 1-chome
        Minato-ku, Tokyo 108
        Japan

        FAX:   +81-3-3456-1618

www-admin@icot.or.jp