Language standards for LispOS

Henry G. Baker
Wed, 21 May 1997 15:46:53 -0700 (PDT)

> BTW, I would really hate to start attaching semantic significance
> to the innards of symbol names, since I was sorta imagining a day
> where we don't use strings as symbols, and no code breaks except
> that that deals with textual source code.

Actually, this is an excellent point.  The original Lisp systems had
the property that, if the reader & printer were expunged, one could
kill all the pnames as well, and the system would still run!  (Well,
assuming that one didn't try to execute 'intern'.)  I have tried to
'understand' pnames, to the extent that I have proposed putting off
assigning pnames to gensym'd symbols _until they are first printed_.
This saves a great deal of time for the large number of gensym'd
symbols that are never printed out, and also saves space.

Getting back to your original point: there is no reason why a pname
couldn't be a Kanji character, or a bitmap, or a tiff image, or whatever.
There does need to be a default/canonical method for comparing two of
these pnames for equality, however.

Henry Baker
www/ftp directory URL: