Re: Issue: DESCRIBE-INTERACTIVE (Version 3)

[David A. Moon <Moon@STONY-BROOK.SCRC.Symbolics.COM>]

    Date: Mon, 31 Oct 88 13:21:15 EST
    From: Dan L. Pierson <pierson%mist@multimax.ARPA>

	I oppose DESCRIBE-INTERACTIVE:NO on the grounds that it creates
	extra work for implementors and users of at least one implementation,
	for no compelling reason.
    
    I have to disagree that there is "no compelling reason".  The problem
    centers on whether on not you believe that portable programs should be
    able to describe objects to users (and how compelling you feel that
    need is).

I believe they should, but I don't feel that the DESCRIBE-INTERACTIVE:NO
proposal is necessary to allow them to do so.

	Instead of eliminating functionality from DESCRIBE, it would be better
	to suggest that DESCRIBE methods (and any other programs that ask
	questions of the user) should call the recently added
	STREAM-INTERACTIVE-P function (I'm not sure I've got the right name) so
	as to avoid asking questions when there is no one to answer them.  This
	would address the stated goal "to call DESCRIBE in a batch applications
	without hanging the application" without requiring incompatible changes
	to current practice.

    Unfortunately an implementation suggestion does not license portable
    programs to rely on it.  While your proposal might increase the
    probability that a portable program that called DESCRIBE wouldn't
    break, it doesn't eliminate it.

But this is simply an argument that no portable program can call any
function that it did not define itself, because that function might
have a bug in it!

    In general, I don't see what the world gains from having DESCRIBE be a
    restricted INSPECT instead of an output-only function.  On the other
    hand, I don't want to waste a lot of everyone's time arguing this.

I'll try not to send any more mail on this subject.