Re: Hierarchical lists, again

[hanche@ams.sunysb.edu (Harald Hanche-Olsen)]

>>>>> On Mon, 24 May 1993 19:17:11 -0400, haahr@mv.us.adobe.com (Paul
>>>>> Haahr) said:

Paul> a truly brilliant insight.

Thank you.  You're being much too kind...  :-)

Paul> now, should Harald's % function (perhaps with some other name)
Paul> be added as a builtin to the language?  (i.e., should it be
Paul> defined in initial.es?)  my feeling is that it should be, but
Paul> with an alphabetic name.

I chose the funny name because it is such a fundamental operation and
it looks almost like syntax (which it isn't, of course).  Also, the
unwrapping operation `<=' doesn't have an alphabetic name either...
But I will not insist on my name.

Paul> 	values		# pretty unclear, but sort of resembles lisp

I think someone else suggested that `result' should have that name.  A
motion I would second, but again I am not fanatic about it.  Note
that the `values' primitive as it is suggested for Scheme is sort of a
weird construct, in that you cannot assign its result to a variable:
Essentially the only thing you're allowed to do with it is to pass it
on to call-with-values.  Using the name here will only cause
confusion, I think.

Paul> 	thunk		# an unevaluated routine; the describes implementation
Paul> 	wrap		# i kind of like this because it's a verb

Either is good.  I vote for wrap, if I can't have my non-alphabetic
name :-).  Except that really calls for an unwrap function to
complement it :-(.  Maybe thunk is better after all.

Paul> one minor observation.  [...]
Paul>         result <={cmd}
Paul> is the same thing as
Paul>         cmd

Ouch.  Yes, I am still getting used to es, and tend to be needlessly
convoluted in my writing sometimes.

- Harald