Scheme vs. CommonLisp vs. the World
Mon, 12 May 1997 12:33:58 -0700
> non-conforming Common Lisp. For example, full-fledged
> common lisp arrays, with fill-pointers, displacement, etc.
> are likely to be a higher level construct built upon
> single dimensional arrays.
Well, define Silk (my new CL dialect name) to not include them.
I see the job as simplifying CL where it pays dividends
in complexity and performance, relative to the functionality it
provides. I think I can count on my hands how many people think
complex numbers should be retained in a new CL.
BTW, displaced arrays can be very useful, especially when
interfacing efficiently with hardware, since they can be used
instead of doing pointer-arithmetic, as is done in C.
If they aren't part of the language, do we have two different
AREF functions, kernel:aref, which only handles simple kinds of arrays,
and cl:aref which handles both simple and complex?
Yet another BTW, I recall Scheme only has arrays of type "t".
It doesn't have (unsigned byte 8) element arrays,
or (unsigned byte 32) element arrays,
or (unsigned byte 1), all of which are incredible useful for
writing efficient system-level code and interfacing with
hardware and other byte-level internet protocols.