[R6RS] my notes on today's conference call (6 December 2006)

William D Clinger will at ccs.neu.edu
Wed Dec 6 09:24:16 EST 2006

Conference call December 6 2006 8:00am-9:00am
Present by 8:06am:
Kent, Anton, Matthew, Mike, Will

0. finalize agenda (1 minute)
    move library phasing to follow scripts

1. action items (1 minute)
   - finish recording draft responses in svn tickets directory (All)
        not all done yet
   - draft response to ticket #88 (improve port i/o) (Mike)
        Mike did not volunteer for this
   - read ticket #88 (improve port i/o) and Mike's response for discussion
     TODAY (All)
   - prepare proposal to replace create ports directly from input sources
     and output sinks as a replacement for readers and writers (addresses
     ticket 89) (Mike)
   - check new grammar to make sure inf.0 and nan.0 are only symbols (Will)
   - update #97 response to add timeline: 3/15 deadline for formal comments,
     4/15 deadline for editor responses, 5/15 deadline for new R6RS draft;
     mention that after that the process is up to the steering committee (Mike)
   - update #136 response to note that no top level is specified and that
     implementations who supply one (and probably most will) can provide
     a mechanism, possibly through imports, for addressing this comment (Kent)
        not done
   - prepare proposal for numeric libraries: "incremental improvement" of
     discarding exact-only, inexact-only, fixnum libraries while moving
     fixnum carry operators into the fx library and establishing a new
     bitwise library for the displaced exact-only bitwise operators
     (addresses tickets 27, 40, 52) (Will)
   - establish svn-based voting mechanism for accepting formal-comment
     responses (Kent)
        should split routine tickets into separate files (Mike)
   - update response to rationalize the various iteration procedures [78] (Anton)
     - flush hash-table-fold
     - add hash-table->vector
     - vector-for-each, vector-map (same interface as for-each and map)
     - call/cc can be used for premature termination
     - (vector-map p (hash-table->vector ht)) and
       (vector-for-each p (hash-table->vector ht)) can be recognized by
       optimizer if implementor feels optimization is useful
     - problem: how does hash-table->vector return both keys and values?
        Anton will discuss later
   - split the report into base and libraries (Mike)
   - follow up with steering committee re: front page (Kent)
        the steering committee considers the current draft to be misleading
            Findler wasn't an author of R5RS
            the listed authors didn't actually write R5RS
            the list doesn't include some who actually did write parts
   - insert readability and other guiding principles from
     status reports into draft R6RS (Mike)
        not done
   - note: tickets 1-3 are test tickets and can be ignored
   - note: tickets 135 and up came in after 2006/11/15 and can be deferred

2. scripts
   - tickets 51, 105, 106
   - concrete proposal given in response to ticket #51
        vote on Anton's draft response: 4-1 (Will voted no)

10. library phasing
   - tickets 92, 109, 110, 112, 123
   - need concrete proposal
        Matthew will write something

3. add #!case-sensitive and #!case-insensitive syntax?
   - related to ticket 107
        general feeling: this could go into a non-binding appendix

4. port/transcoder changes
   - ticket 88
        Mike and Will still disagree on some details
        Anton, Kent, and Matthew should consider this issue

5. should bodies be more widely permitted?
   - ticket 117
   - leave begin alone but use "implicit let" (i.e., body) instead of
     "implicit begin" in case, cond, when, unless, etc.?
        seems radical but sound
        Kent and Will will search the Chez and Larceny code bases
            to see how often this would have been useful

We didn't have time to consider the following issues:

6. should we flush compound library names?
   - ticket 84
   - proposal in response: keep compound names
   - may need better rationale

7. expansion process
   - tickets 61, 62

8. reduce overspecification as well as underspecification
   - ticket 87
   - need a concrete proposal

9. revisiting list operation exception-raising requirements
   - tickets 26, 42, 41, 36, 48, 49
   - need a concrete proposal

11. rationalize hash-table/vector/list procedures
   - ticket 78
   - concrete proposal pending

12. record-type-length procedure
   - ticket 72

13. multiple versions of one library
   - ticket 130

14. enumerations
   - ticket 70

15. adjourned about 9:20am

More information about the R6RS mailing list