Newsgroups: comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!satisfied.elf.com!news.mathworks.com!uunet!news-in1.UU.NET!sytex!smcl
From: smcl@sytex.com (Scott McLoughlin)
Subject: Re: DIM, CLIM or DK?!? (Was: Re: purify vs. gc)
Message-ID: <D858Zc1w165w@sytex.com>
Sender: bbs@sytex.com
Organization: Sytex Access Ltd.
References: <24464.9502081609@subnode.aiai.ed.ac.uk>
Date: Wed, 8 Feb 1995 21:50:48 GMT
Lines: 46

jeff@aiai.edinburgh.ac.uk (Jeff Dalton) writes:

> (as well as users) who can't afford to license a GUI.  Other factors
> may also act to prevent all implementations from having a common GUI.
> The result is a common Language, and some common libraries, but
> not a common GUI -- which is exacrtly how the Lisp community
> managed to throw away their lead in GUIs.

I think the precise nature of "Other factors" will have alot
to do with the success/failure of Dylan.  If optimizing code 
is a "black art" that requires extensive knowledge of the
implementation or constant dipping down into various system
specific modules, then Dylan will "fail" as a mass market
general purpose language.  Lots and lots of companies make
X-platform GUI's for C++.  It's enough work juggling the
feature/API differences between Windows/Mac/Win32/Motif/etc.
and mapping these to a common code interface without 
having to screw around with language implementation specific
hacks. The "basic idioms" of the language are going to have
to compile to rather fast code so that a good old fashion
"tower of abstractions" approach makes X-platform GUI's
commercially viable. In the past few years, various Dr. Dobbs
articles have described "porting" X-platform C++ GUI's to
other OS's/language implementations, and if memory serves,
I've read of levels of effort ranging from two person
weeks (a simple Win3.1 -> Win32) to some number of person
months.

> 
> I note that you're unable to make your case without insulting people.
> 

Sorry if I insulted anyone. I myself am a "lab nerd" type.
I love the freely available tools and have written a few
minor ones myself years ago. When it comes to compilers,
though, I spend $$. I want _very_ timely updates to support
whatever the latest whim of the machine/os vendors happens
to be. I want language implementations tailored to the
production of software for the "sweet spot" majoritarian
(if often brain dead) hardware/software platforms in use
today throughout offices in the U.S.

=============================================
Scott McLoughlin
Conscious Computing
=============================================
