Subject: Re: CL grammar ambiguities?
From: Erik Naggum <erik@naggum.no>
Date: 17 Feb 1995 18:58:51 UT
Newsgroups: comp.lang.lisp
Message-ID: <19950217T185851Z.enag@naggum.no>

[Jeff Dalton]

|   Of course one can have a standard without absolute precision!
|   Whatever gave you the idea that one couldn't?

wearing my standards-committee member and standards writer hat for a second
or two (it hurts to wear it), a standard with absolute precision must be
about things for which absolute precision makes sense, such as dimensions
of physical objects such as those for mechanical and electrical
engineering.  however, even they will not enumerate everything that will go
wrong if you do something that is outside of the scope of the standard.
defining the scope of a standard is the hardest part of the process,
because it defines what you will and will not talk about.

the Lisp reader as defined does not detail what will happen to any random
character sequences when presented to it as input, but rather what should
be presented as input to that reader when a particular set of results is
expected and desired.  I think this is no different from any other syntax
description, formal or informal.

|   Indeed, can anyone think of a programming language standard that does
|   have absolute precision?

modula-2 comes close.  its formal semantics is defined using VDM.  I also
think Scheme comes close for its semantics.  this hinges on an operational
definition of "absolute precision" that makes it possible for the this term
to make sense, and I don't think we have an absolutely precise definition
of "absolute precision" against which anything could be measured.

#<Erik>
-- 
miracle of miracles.  look what the Net dragged in.