Newsgroups: comp.lang.c++,comp.lang.smalltalk,comp.object
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!uunet!psinntp!shellgate!camo!rgh
From: Richard G. Hash <rgh@shell.com>
Subject: Re: C++ Productivity
Message-ID: <rgh.792885220@camo>
Sender: usenet@shellgate.shell.com (USENET News System)
Reply-To: rgh@shell.com
Organization: Shell Bellaire Research Center
References: <1995Feb4.210947.718@mole-end.matawan.nj.us> <1995Feb6.132028@di.epfl.ch> <3hdugu$l2q@calvin.st-and.ac.uk> <D3rCwL.D8o@research.att.com>
Date: Wed, 15 Feb 1995 21:53:40 GMT
Lines: 19
Xref: glinda.oz.cs.cmu.edu comp.lang.c++:113110 comp.lang.smalltalk:20902 comp.object:26844

In <D3rCwL.D8o@research.att.com> ark@research.att.com (Andrew Koenig) writes:
> Of course, the choice between expanding out the B-tree for each type
> and making it work in terms of an underlying pointer-based data structure
> is a classic time-space tradeoff.  Compilers are not generally in
> a position to second-guess how their users want to make such tradeoffs,
> which means that if you care about the outcome then you have some
> responsibility to care about the choices you make.

Interestingly enough, certain Ada compilers have had code-sharing
generics ("templates") for many, many years now. You do not end up
with N*codesize for N instantiations of a particular generic.
I can't see why C++ compilers cannot mature to this level as well.
I'll agree templates are pretty much the pits right now.

--
Richard G. Hash                                      email: rgh@shell.com
Shell Development Company, Bellaire Research Center  phone: (713) 245-7311
Member Team Ada                Free Ada95 compilers: cs.nyu.edu:/pub/gnat
Distributed, Full-OO, Multithreading, all built in. And it's free.
