Re: policy to adopt

[Kent Dybvig <dyb%unc.csnet@csnet-relay.arpa>]

I have certainly gotten the impression that everyone WAS after
a Common Scheme of sorts, and not just so that papers use common
syntax.  Why standardize on so many of the function names?  More
importantly, why standardize on the controversial nil/t/boolean
issue?  Why have numbers and streams subcommittees?

I think we need more discussion of the details of the proposal
rather than less.  Standardizing is dangerous if done only half-
heartedly.  We must be absolutely sure of what we standardize on,
since we will likely be stuck with our decisions for several
years.

As for the list-length, list-append issue, I think we ought to take
one of two stances:

	Stance 1:  sequence-type functions such as ref, length and
		   append should be prefixed by "list-" for the
		   list version, "string-" for the string version, 
		   and "vector-" for the vector version.

	Stance 2:  sequence-type functions such as ref, length and
		   append should be prefixed by nothing for the
		   list version, "string-" for the string version,
		   and "vector-" for the vector version.

The first stance is preferable because of the symmetry, the second
because the names for the list versions (probably the most commonly
used) are shorter.  I think that the first stance is less confusing
and less likely to cause trouble.

    Cheers,
	Kent Dybvig