Bennett Todd <[EMAIL PROTECTED]> writes:
> If perl6 were to want to try and rise to the level of jwz's design
> aesthetics, I'd say we oughta save ourselves a lot of work and abandon
> it now. jwz (thanks for xemacs and netscape) is solidly in the "no
> program can be gigantic enough, or slow enough" school of software
> design.
That's really unfair. jwz wasn't responsible for Netscape 4 at all and
xemacs (which I should point out that I use and find quite usable) both
isn't that bloated for the sheer amount of capabilities that it has and
had a lot of its basic architecture already fairly firmly set by emacs.
jwz is a LISP programmer and I'd venture to say that he has a lot more
experience with solid GC than a lot of folks on this list. On the other
hand, it's worth remembering that he may not consider some things
important that Perl folks think are important (like order of destruction
or predictability of when GC happens). There's something to be said for
writing code so that those things *don't matter* and to have a good enough
GC that you just don't have to care (running it in another thread seems to
be a technique that's growing in popularity among the commercial LISP
systems).
Please also note that he's criticizing the emacs GC about as much as he's
criticizing the Perl GC.
--
Russ Allbery ([EMAIL PROTECTED]) <http://www.eyrie.org/~eagle/>
