Re: a (revised) draft of R5RS

[Matthias Blume <blume@CS.Princeton.EDU>]

From: "R. Kent Dybvig" <dyb@cs.indiana.edu>
Subject: Re: a (revised) draft of R5RS
Date: Thu, 8 Jan 1998 19:32:53 -0500 (EST)

> It is a fact that Scheme requires proper treatment of tail recursion.
> I and I'm sure others would prefer if we were to limit discussion on this
> issue in this forum to the merits of the proposed wording for the report,
> rather than on why proper treatment of tail recursion is or is not a good
> idea, at least until R5RS is complete.  I agree that we should fix any
> phrasing that implies generality of proper treatment of tail recursion
> beyond Scheme, but I don't see any such phrasing that a word or two fix
> by the authors of the passage wouldn't cure.

It is also a fact that the requirement for proper tail-recursion does
not mean much as long as the behavior of other language constructs
with regard to space have not been specified.  This has always been
the case, and no matter how much extra wording is added, if it falls
short of a _complete_ characterization of the space requirements of
all features of the language, then it will remain hollow.

As far as I am concerned, the wording in R4RS was just fine because
there is no such thing as being "almost precise".

-Matthias