Re: Issue: FUNCTION-TYPE (Version 8)

[Masinter.pa@Xerox.COM]

I believe this issue may well come to a vote -- i.e., not decided by
consensus. There are fairly strong feelings on both sides of the
coercion issue. 

I don't think it is ready for a vote until it is clear that the folks
voting understand the issue and its ramification. There were enough
questions and incorrect assumptions at X3J13 that I believe we need to
work harder to explain the issues.

There were several folks who believed that removing coercion removed
late binding in several circumstances.

One example:

(mapcar 'frob my-list)

if, during the course of execution of frob, one should hit a breakpoint
and redefine frob, would subsequent iterations get the new definition?

More examples of the coercion are called for, I think.

Frankly, I didn't detect any large amount of enthusiasm for introducing
a new name PROCEDURE in order to avoid stepping on the current
definition of FUNCTION. Do any of you have any additional comments on
this, one way or another?

Someone afterward said that Kent had put up an example of why coercing
was important ... this seems important to add to the proposal.