[R6RS] my notes on today's conference call (20 July 2006)

[William D Clinger]

Conference call July 20 2006 2:00pm
Four present by 2:01pm:
Kent, Mike, Will, Anton
Matthew arrived about 3:10pm

NOTE CHANGE in day to THURSDAY and time to 2:00pm Eastern Time (this week only)

0. finalize agenda (1 minute)
    (agenda was re-ordered to postpone items until Matthew arrived)

1. Permanent action item (5 mintues)
   - reduce response latency on email discussions (ALL)

2. action items from 7/11/2006 (5 minutes)
   - safe/unsafe mode:
     - (re)post safe/unsafe proposal, address questions in draft (Kent)
     - (re)post safe/unsafe proposal (Will)
     - discuss (All)
        done
   - hash tables (ALL)
     - discuss whether updating table during for-each/fold operators should
       be left unspecified
     - discuss whether to keep hash-table-update!
        no email, will vote today
   - library syntax and semantics
     - add syntax for specifying default phase (Kent and Matthew)
        to be posted
     - shed some light on the dark corners mentioned in Kent's note (Matthew)
        done
     - discuss whether to separate phases (All)
       - if separate, propose syntax for handling phasing in library
         imports and in eval
   - email proposal for new arithmetic procedures (Will)
        done
   - I/O
     - read and discuss on current I/O SRFI (All)
        discussed but not converged
   - plan end game for delivering September 1 draft
     - Mike will lay out critical path
     - discuss implied meeting/email agenda for next 5-6 weeks
        in process
   - carried over:
     - update library srfi (Matthew and Kent)
        in progress
     - update reference implementation for arithmetic SRFI (Will)
        no progress since last week, but was current last week
     - complete unicode reference implementation (Will)
        no progress
     - comment on core/library split (All)
        not much going on

3. hash tables (5 minutes)
   - remove hash-table-update! ?
        Vote: keep, 3-2
        keep: Anton, Mike, Matthew
        remove: Will, Kent
   - remove hash-table-for-each and hash-table->alist ?
        Vote re hash-table-for-each: remove, 1-3
        keep: Mike
        remove: Anton, Kent, Will
        Vote re hash-table->alist:
        keep: 
        remove: Will, Kent, Anton, Mike
   - any other changes?
        hash-table-fold and list version not compatible?
            procedure argument should come first
            Anton, Kent, Will agree
            (Mike temporarily offline)
   - accept (with agreed-upon changes)?
        yes (by the three people present)

7. New arithmetic procedures (10 minutes)
   - see Will's posting
        no objections to including them

6. I/O (10 minutes)
    several open issues
        how to deal with encoding errors
            Matthew and Will have responded, but Kent and Anton haven't
        meaning of EOF
        changes of encoding in mid-stream
            once-only mutation of input and output ports
        changes of buffering in mid-stream
        whither display?
        requiring explicit port arguments?
    need more discussion via email

5. multiple return values (10 minutes)
   - first vote: okay to leave unspecified?
        Vote: yes, 4-1
        yes: Will, Kent, Matthew, Anton
        no:  Mike
     if yes, leave unspecified
     if no, proceed to second vote
   - second vote: accept Will's proposed semantics
     if no, proceed to third vote
   - third vote: accept Kent's proposed semantics
     if no, proceed to fourth vote
   - table for email discussion on hybrid solution, e.g., accepting
     will's suggestion to treat zero values as the unspecified value
     in single-value contexts and treat two or more values as an
     error in single value contexts
        tabled because Matthew isn't here

4. simplified condition type hierarchy (10 minutes)
   - should we go with existing hierarchy?
   - should we simplify?
     - see https://r6rs.scheming.org/node/242#comment-1362
        yes, we should simplify
        the relevant sub-editor (Mike) will propose something concrete

for next week:
set-car!, set-cdr!
safe/unsafe
i/o
body and library-body syntax
library syntax and phasing

8. adjourned about 3:39pm