Re: [PROPOSAL] identifiers are like identifiers in programs

[ramsdell@linus.mitre.org (John D. Ramsdell)]

Erik Ostrom <eostrom@research.att.com> writes:

> Note to implementers: One effect of this is to separate "ident"
> (used in message and argument names) from "unquoted string" (used in
> data tags, auth keys, and argument values).  Is this the right
> division?  Should some of those unquoted strings be idents instead?
> Should we give up and just quote all strings?  I don't care.
> --Erik

I think unquoted string values should be a distinct syntactic category
and not tied to the syntax we choose for identifiers.  It seems silly
to me to have to quote the string value 0.  The extra code required to
maintain a distinction between unquoted string values and identifiers
is small.  Let's have two categories.

John