[a37078%ccut.u-tokyo.junet%utokyo-relay.csnet: Lisp standardization]

["Scott E. Fahlman" <Fahlman@C.CS.CMU.EDU>]

Here is Ida's response to the note I sent him.  It is a bit hard to
parse, as you will see, but it does contain lots of good information
about the structure of the effort in Japan.  I'm still trying to sort it
all out, but it looks like Ida is certainly one of the people we want,
maybe on the steering committee as he suggests.  Suggestions are welcome
at this point.

-- Scott
---------------------------------------------------------------------------
Date: Sat, 14 Jun 86 11:57:40+0900
 From: Masayuki Ida <a37078%ccut.u-tokyo.junet%utokyo-relay.csnet at CSNET-RELAY.ARPA>
To:   fahlman, ida at UTOKYO-RELAY.CSNET
Re:   Lisp standardization

Professor Fahlman,

Thank you for your quick responce to my mail.

Here are my views/opinions/descriptions about the things you wrote to me.

1) My visit to Lisp conference and AAAI conference.
I will visit USA with several persons of our committee as a team.
Scheduled dates are one week stay at Boston from Aug.3 to 10,
and another one week stay at AAAI site, roughly.

1') I will join you at the meeting you suggested. and I am pleased to know
I will be welcomed to discuss and tell the situation in japan.

1'') As I stated in 1), I will be with several colleagues at hotel.
We will have a meeting at the hotel.
So, I want to arrange an invitation to our meeting.
Can you spare your time to join us  at Boston ?

2)General steps toward the standardization of computer languages in japan.
There is only one committee for MITI whose name is JIS programming language
standardization or so. All the computer languages are defined by this 
committee. But the actual working is not carried by the committee, like
SC22 forms WG for each language. The membership of this top committee
are not opened and I am not the member of this top committee.
I think almost all the members are very senior persons.
  From my experinece and my knowledge, the standardization is initially
directed by MITI. This task force is undertaken by Jeida or IPSJ.
After the actual works by Jeida or IPSJ finish, MITI calls the members for
the top JIS committee and ask them to guarantee it. The draft which appear to
the top committee is not actually discused, when the language spec is parallely
defined to ANSI.
Jeida standardization team is carried by Prof. Yoneda (u-tokyo).
IPSJ standardization team is carried by prof. Nakata (u-Tsukuba).
Prof. Nakata is an official member of ISO TC22 as a representative of Japan.
(he gave me a copy of Bob Mathis's proposal of Ad Hoc Group on the preparation
of NWI on Prolog and LISP to ISO/TC97/SC22)
The documents appeared at ISO are send to MITI, then forwarded to several 
persons, including Prof. Nakata at least.
He has, currently, a role to catch up the standardization of Fortran, Cobol,...
On the other hand, Prof. Yoneda has a role to establish a standard 
for more "fresh" languages, like C, Ada, Lisp,...
MITI select and decide which team is more suitable for any computer languages.

Last tuesday, June 10, I was called by the staff of Prof. Yoneda's committee 
 at Jeida.
He told me that MITI suggest to start the working committee for Lisp
standardization, and that the committee is under Prof. Yoneda's committee 
and I should be the chair of the JIS committee also.
Then I will start the working committee to make a JIS draft with 13 members.
the number of members is prior assigned and given to me.
The scheduled dates of this year is one-a-two-month.
The first meeting will be in July.

Prof.Yoneda (and Prof. Nakata) is  very senior person.
I think the formal process in Japan is going just like you mentioned.
I mean I agree your suggestion of your mail. i.e. ANY offical activities for
standardization in Japan need senior persons who have responsibilities to 
totally control the whole process, even though he has only a basic knowledge 
about the language, and he can not understand the details 
or he have no time to spare to learn the language details.

This JIS committee is different from the Common Lisp committee I have been 
talking about. But, they are all in Jeida.
And they will be gathered to form a one large committee, I think.

I think it is very usual to form a JIS working committee which is parallel to
ANSI/ISO committee and the JIS committee will communicate with ANSI committee.

3) Member to ANSI committee from Japan
I think the chair of the JIS committee should be a member of ANSI committee for
this case. The reason is to avoid the separation of two standardizations 
and to keep ANSI committee can totally control the whole thing.
I think the latter property is important, considering an unfortunate condition.
Further, I even think it might be the best that someone in ANSI committee 
will attend Jeida committee.

My current Jeida Common Lisp committee have totaly , over 40 members from
26 different organizations.
They have more knowledge and skill than before.
But KCL persons have a great role and a great infruence to Common Lisp 
in Japan. So, technically, Mr. Yuasa or Mr. Hagiya is the best person to 
technical committee. I think I am the next to them for the technical matter.
For steering and standardization, I think I am the best person to join.

Last april AI Association of Japan was formed. in my view, this organization
has a relation to Common Lisp as a user organization. 
There is also a standardization committee. I was asked by them to be a chair 
for the standardization committee. But it is not formed yet. I do not know
whether this group will survive or no now. But as the result of our 
questionnair shows, AI application in Japan is 60% based on Lisp or more,
and the most actually used Common Lisp implemetation is VAXlisp,then 
Symbolics Lisp, then KCL. We should watch what the user society want.

4) Object oriented facility for Common Lisp.
I have been following CommonLoops.
Last year, before PCL appeared, I designed an interpretation of CommonLoops 
and presented a paper on it.
Last Feb, I stayed at Xerox PARC for one week.
I have experienced and somewhat assisted them to improve PCL.
Xerox lawyer permitted me to carry back PCL sources.
Now, I play a role of re-distributor of PCL in Japan.
I am now researching with PCL.
Since 1984 fall, I followed the discussion of o-o-bboard with the permission of
Ken Kahn. Since last summer, I directed some members of my committee to catch 
up which is the best for Common Lisp among CommonLoops, objectLisp,...
I think we, japanese are ready to accept the function call generalization
of message sending and defstruct based flavor and so on.
Several researchers in japan, to my surprise, already presented papers for 
present Prolog like facility upon CommonLoops-like object oriented facility.

5) Subset.
Personally, I have an opinion that there should be two levels of Lisp 
language specification.
Our subset working group will present their polish up of my proposal at July 
8th in Japan. This group will make a pilot version of the subset.
This subset can not self-compile itself.
BUt this subset is intended to be fully compatible to the super.

I have several friends in Gold Hill computers.
Last March, I presented our subset draft to Dr. Jerry Barber at his office.
And discussed with him for one or two hours.
I also informed Professor Pat. Winston at his office.

I hope we can make peace with them.

As to EuLisp, here is Dr. Jean Peer Briot from Paris who is now visiting 
Japan as a visiting scientist, who has a relation to Chaillox of LeLisp.
I was phoned by him and we met in tokyo.
He told me that he was directed to meet me by Chaillox.
I gave several staffs to him.

6) kanji and character/string
We will polish up our discussion.
I formed a working group for it. members are from ETL, HItachi, NEC, Fujitsu, 
Xerox, Symbolics, I and some othe rpersons.
I think we can have some more concrete opinion before I will goto USA.
I do not think it will be integrated into the one and only one opinion 
immediately. I think it may need a voting or take a year or so.



Masayuki Ida