Minutes of the 4th Scheme standarization meeting

[Chris Haynes <chaynes@iuvax.cs.indiana.edu>]

                IEEE/MCS/P1178 Working Group on Scheme

               Unapproved Minutes of the Fourth Meeting
                 January 19, 1990, San Francisco, CA

The meeting was called to order by Christopher Haynes at about 1:45PM.
The attendees were:

        Norman Adams                    Xerox PARC
        Scott Burson                    Blue Space Foundation
        William Clinger                 University of Oregon
        Pavel Curtis                    Xerox PARC
        Ken Dickey                      Tektronix, Inc.
        Kent Dybvig                     Indiana University
        Marc Feeley                     Brandeis University
        Dan Friedman                    Indiana University
        Dick Gabriel                    Stanford University
        John Gateley                    Texas Instruments
        Jed Harris                      Apple Computer
        Christopher Haynes              Indiana University
        Robert Hieb                     Indiana University
        Roger Kirchner                  Carleton College
        William Maddox                  U. of California at Berkeley
        Sidney Marshal                  Xerox
        James Miller                    Brandeis University
        John D. Ramsdell                The MITRE Corporation
        Benjamin Schreiber              Brandeis University
        Guy L. Steele, Jr               Thinking Machines Inc

1. The agenda was amended; the changes are reflected by the minutes.

2. John D. Ramsdell was elected secretary.

3. Minutes of the third meeting were accepted with no changes.

4. Reports on R4RS, macros, and ISO.

   A. Report on R4RS. (Will Clinger)
      R4RS is complete except for the appendix describing macros and the
      bibliography.

   B. Report on the status of the macro committee. (Will Clinger)
      1. Syntactic closures are out of favor.  Chris Hanson attempted to
         implement EXTEND-SYNTAX using syntactic closures.  The attempt
         was not entirely successful because of a previously unrecognized
         technical problem: in some cases an EXTEND-SYNTAX based on
         syntactic closures must know the syntactic role of an identifier
         before macro expansion has made that role apparent.
      2. A proposal based on a descendent of hygienic macro expansion
         has some members of the macro committee confident that a
         resolution to the macro impasse is at hand.   The current
         proposal is due to Jonathan Rees, who modified a proposal by
         Will Clinger, who based his proposal on Eugene Kohlbecker's
         Ph.D thesis work. 

   C. Report on the International Standards Organization WG-16. (Dick Gabriel)
      WG-16 is not currently interested in Scheme as a basis for an 
      international standard.
      
5. Quick review of IEEE standardization procedures. (Chris Haynes)
   Resolved: The chair will post instructions to the Scheme notesfiles on
   how to request membership in the balloting group for the Scheme standard.
   
6. Discuss draft P1178/D3.
   Many of the agenda items were entered by Sidney Marshall.

   A. Consensus: The editors should fix the introduction so that the 
      definitions of "confirming implementation" and "conforming program"
      constrain the implementation to meet the standard.

   B. Consensus: No change is needed in the description on pg. 11 of
      the "evaluates to" symbol.
   
   C. Consensus: An editorial change shall be made to the section on
      disjoint types with "No object satisfies more than one ..."
      changed to "Every object satisfies at most one ...".

   D. Consensus: The restriction on duplicate entries in CASE key
      sets should be removed and the text updated to be consistent
      with the derived semantics.  That is
               (case x ((3.14 3.14159) 'pi) (else '()))
      is fine on machines on which (eqv? 3.14 3.14159) => #t, and
      key lists are examined inorder to find the first match.

   E. Consensus: Add to the text a statement of the form "Assigning
      a new value to a variable defined in the initial environment
      does not affect the behavior of any procedure defined in the
      report."

   F. Consensus: Change reference 20 to the second edition of CLtL.

   G. Consensus: Change the description of remainder to correctly deal
      with the sign of negative zero.

   H. Consensus: References to unpublished works should be changed into
      references to published works.
      Ref 5 --> "The Theory of Numbers," by G H Hardy & E W Wright, 
      Refs 17 --> William Clinger and Jonathan Rees (editors), The revised^4
      report on the algorithmic language Scheme, University of Oregon 
      Technical Report CIS-TR-90-02.
      Ref 22 --> Thinking Machines Technical Report by Guy Steele, Jr.
      Section 1.1 should refer to Steele's master's thesis (MIT AI-TR-474)
      for a definition of "properly tail recursive".
      The bracketed sentence at the end of Appendix C.3 should removed
      and the sentence preceding it changed to refer to William Clinger,
      How to read floating point numbers accurately, University of
      Oregon Technical Report CIS-TR-90-01.

   I. Withdrawn: The word "should" should be eliminated from text.

   J. Withdrawn: Definitions should be treated syntactically as
      expressions. 

   K. Moved and rejected: Eliminate CALL-WITH-CURRENT-CONTINUATION
      from text.
      It was argued that elimination would encourage more people to use
      IEEE Scheme as their extension language, but others felt Scheme
      without CALL-WITH-CURRENT-CONTINUATION is simply not Scheme.
      The vote was three for and eleven against.

   L. Withdrawn: Change text so that "(BEGIN)" is a legal expression.

   M. Moved and rejected: A key set of a CASE form may not be empty.

   N. Moved and rejected: Remove named LET from the text.

   O. Discussed and rejected: All parts of lists generated by a
      quasiquote expression should be mutable.

   P. Moved and accepted: Change the text in the section on symbols to
      qualify some of the examples with the phrase "symbols as defined
      in this report".

   Q. Withdrawn: All conforming implementations must support a minimum
      number range.

6. Moved and unanimously accepted: After amendment to conform with the
   consensus of this meeting, and other changes deemed non-controversial 
   by the editors, draft D3 should be submitted to the MSC for public 
   comment and balloting. 

7. Consensus: An ad hoc subcommittee of this working group will
   respond to public comments and negative ballots.  The members of
   this committee will be Chris Hanson, James S. Miller, and John D.
   Ramsdell.  Another meeting of the working group will be called
   if it is felt that the responses recommended by this subcommittee
   may be controversial.

8. Moved and accepted: This working group will terminate its activity when
   the draft is accepted as a standard, and will reform when necessary to
   revise the standard.

9. Moved and unanimously accepted: The committee thanks Christopher Haynes,
   the IEEE editors, David H. Bartley, Chris Hanson, and James S. Miller,
   and the RnRS editors, Will Clinger and Jonathan Rees.

10. Moved and accepted: Identify the standard draft editors in the text.

The meeting adjourned at 5:45PM.

-- Minutes by John D. Ramsdell, edited by Christopher Haynes.