Newsgroups: comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!pipex!uunet!sytex!smcl
From: smcl@sytex.com (Scott McLoughlin)
Subject: Re: Where was Dylan as OOPSLA?
Message-ID: <yukJVc1w165w@sytex.com>
Sender: bbs@sytex.com
Organization: Sytex Access Ltd.
References: <creedy-0911941238460001@clreedy-mac.mitre.org>
Date: Wed, 9 Nov 1994 21:14:33 GMT
Lines: 47

creedy@mitre.org (Chris Reedy) writes:

> 
> I was also surprised at the predominance of SmallTalk and the lack of C++,
> especially in the context of business systems.  One comment I heard from
> an attendee from a commercial manufacturing company was that one of the
> reasons his company was interested in SmallTalk was that they had had much
> better experience retraining COBOL programmers as SmallTalk programmers as
> opposed to C++ programmers.
> 

Howdy,
        I wouldn't be so surprised. My background is in the "trenches"
of MIS, and my partner still spends part of his time in a MIS
shop in the U.S. Gov. Our general observation is that THE MIS 
WORLD IS CHOKING ON 4GL'S, VISUAL BASIC AND RDBMS TOOLS. The tools
simply don't have the "expressive power" for the problems they
wish to solve. We saw/see "killer" n>5 table joins and all 
sorts of weird hacks around the lack of reference/pointer
semantics all the time.
        Smalltalk was/is successful in MIS because: (1) It rode
the wave of OO enthusiasm early on, and (2) Digitalk and 
ParcPlace actually seem to _care_ about the MIS market,
in addition to (3) solving the expressiveness problem briefly
characterized above.  Similarly, I read recently (last month
or so) that an OODBMS company in VT or NH is the fastest
growing privately held business in Americal (no, I don't
know how they collect these figures).
        C/C++ would solve some of these problems, but it
is a tough language to learn _well_. I think it's a 
great language for building _tools_ and for delivery
of systems with a big payback, etc.  But it's not a
great language for traditional MIS, rambling incremental
development, high employee turnover, etc. Smalltalk IS
good for these sorts of things. So is Lisp/Scheme/ and
I presume, Dylan, given its heritage.
        Of course, to sell Lisp/Scheme/Dylan to MIS
you gotta go knock on their door and offer them the
right tools at the right price. Because this is not
always _fun_, I think this is better achieved by
hungry companies than research organizations (who 
do a fantastic job for _their_ "market").

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