Myths about Lisp [was RE: strangest systems I've sent email from]

[Rich Alderson]

From: ben
Sent: Wednesday, May 11, 2016 5:42 PM

> On 5/11/2016 5:54 PM, Toby Thain wrote:

>> On 2016-05-11 7:43 PM, Liam Proven wrote:

>>> If we'd had 4 decades of effort aimed at fast Lisp Machines, I think
>>> we'd have them.

>> Compiled Lisp, even on generic hardware, is fast. Fast enough, in fact,
>> that it obviated Symbolics. (More in Richard P. Gabriel's history of
>> Lucid.) See also: The newly open sourced Chez Scheme.

> But List still sequential processing as far as I can see? How do you 
> speed that up?

This is another of the long-standing myths perpetuated by people who
know nothing about the language.

It has literally been decades since lists were the only data structure
available in Lisp.  If you need non-sequential access to process data,
arrays are the ticket, or hashes.  Choose the best data structure for
to problem at hand.

(Similarly, data types other than atoms have been around since the very
 earliest LISP.  They just weren't sexy, and didn't get a lot of press
 since they weren't novel and difficult to understand.  Math code from
 the MACLISP compiler was better than that generated by the F40 FORTRAN
 compiler.)

>> The myths around garbage collection are also thick, but gc doesn't
>> impede efficiency except under conditions of insufficient headroom (long
>> documented by research old and new).

> Well GC is every Tuesday here. :)

You joke, but in one of the visionary papers on GC from the early 70s, a
tongue-in-cheek scenario was proposed in which GC was done by a portable
system which had sufficient memory would visit large facilities to do
background GC for them on, say, a monthly basis.

                                                                Rich

Rich Alderson
Vintage Computing Sr. Systems Engineer
Living Computer Museum
2245 1st Avenue S
Seattle, WA 98134

mailto:RichA at LivingComputerMuseum.org

http://www.LivingComputerMuseum.org/