Re: Symbolics vs stock hardware

[Brad Miller <miller@CS.ROCHESTER.EDU>]

    Date: Sun, 19 Feb 89 14:31 EST
    From: Qobi@ZERMATT.LCS.MIT.EDU (Jeffrey Mark Siskind)

	So, yes, curious as it may sound, the only thing that would make
	Symbolics machines useful to me into the indefinite future would be a
	high-quality Scheme implementation.  I wonder if any other customers
	consider this a priority?

    Jonathan Rees has a Scheme compiler for the 3600. I have never used it
    so I can't comment on its quality. Actually I would very much like
    the plain 3600 Common Lisp compiler to support tail recursion (and
    continuations as well --- though the latter would require some
    extensions to the language.) Many times I have taken to writing loops
    instead of tail recursive calls solely for performance reasons and
    would like to be able to stop that practise.
	    Jeff
    -------

Yes, but this scheme only supports continuations DOWNWARD, when what we need
here are true DYNAMIC/NON-LEXICAL continuations... (i.e. dynamic closures)

INTERLISP allows one to copy the stack into a var and back, which is close
to this, it just doesn't copy the state of all the globals. Even so, this
would be a lot easier to use than stack-groups for this particular
application. Any chance of the stack being opened to such user access?

----
Brad Miller		U. Rochester Comp Sci Dept.
miller@cs.rochester.edu {...allegra!rochester!miller}