Newsgroups: comp.lang.dylan,comp.lang.misc,comp.lang.lisp,comp.object,comp.arch
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.mathworks.com!newsfeed.internetmci.com!psgrain!charnel.ecst.csuchico.edu!csusac!csus.edu!netcom.com!NewsWatcher!user
From: hbaker@netcom.com (Henry Baker)
Subject: Re: allocator and GC locality (was Re: cost of malloc)
Message-ID: <hbaker-2807951148460001@192.0.2.1>
Sender: hbaker@netcom19.netcom.com
Organization: nil
References: <9507261647.AA14556@aruba.apple.com> <3v8g7l$cge@jive.cs.utexas.edu> <3vac07$ptf@info.epfl.ch>
Date: Fri, 28 Jul 1995 19:48:46 GMT
Lines: 29
Xref: glinda.oz.cs.cmu.edu comp.lang.dylan:4913 comp.lang.misc:22377 comp.lang.lisp:18467 comp.object:35923 comp.arch:59966

In article <3vac07$ptf@info.epfl.ch>, "Stefan Monnier"
<stefan.monnier@epfl.ch> wrote:

> Too true: current systems are designed mainly to run stack-based programs, so
> it's always safer to try to stick as much as possible to the same paradigm.
> Appel's argument about GC and stack being just as fast (while the GC
approach is
> simpler since there *is* a GC anyway) is not convincing enough: the added
> complexity buys you predictability (and it's a "one-time" complexity cost).
                                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Oh, but it's _not_ a "one-time" complexity cost!  Just look at that thousands
of papers over the past 25 years about how to force everything into the
Algol/PLI/Pascal/C/C++ "everything has to be stack-based" model.

Somehow, we keep paying, and paying, and paying for this 'one-time' cost.

---------

[For those of you who will be confused by this posting because of my
previous postings advocating stacks, you should know that the stacks
I was advocating are quite a bit different from the Algol/PLI/Pascal/C/C++
model.  If you _have_ to program in one of these dinosaurs, you should
definitely consider using the "Cheney on the MTA"/pushy stack buffer
approach:  ftp://ftp.netcom.com/pub/hb/hbaker/CheneyMTA.html  (also .ps.Z).]

-- 
www/ftp directory:
ftp://ftp.netcom.com/pub/hb/hbaker/home.html
