[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
- To: CL-Cleanup@SAIL.Stanford.EDU
- Subject: Issue: ERROR-CHECKING-IN-NUMBERS-CHAPTER
- From: Kent M Pitman <KMP@STONY-BROOK.SCRC.Symbolics.COM>
- Date: Tue, 4 Apr 89 14:13 EDT
My notes show the following...
The major roadblocks were that GZ (and a few others) were hung up on the
presentation of ``should signal.''
GZ cited the example of (LOCALLY (DECLARE (OPTIMIZE (SAFETY 0))) #'+).
She wanted to know if the resulting function could fail to do error checking.
(RPG and KMP will pursue this.)
JonL really hated the presentation of the boole arguments using #.
RWK said we should definitely do this kind of stuff (error type identification)
now if possible, and not wait for the next standard.
Walter van Roggen was worried that some of this stuff might be controversial,
but I assured him that we would back off to a more vague error type
rather than dispute endlessly about controversial cases. He seemed happy with
Haflich seemed to believe that this was especially important for numbers, so
he was happy to see this chapter done.
Masinter said that with his implementor's hat on, he thought this was a pain,
but that with his user's hat on, he liked it. He was letting his user side
dominate and being very supportive.
There was consensus that we should discuss this (and similar proposals) at
the next meeting.