Newsgroups: comp.lang.misc,comp.programming,comp.software-eng,comp.ai
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!satisfied.elf.com!news.mathworks.com!mvb.saic.com!eskimo!usenet
From: hgeorge@eskimo.com (Harry George)
Subject: Re: A new approach to software engineering!
X-Nntp-Posting-Host: tia1.eskimo.com
Message-ID: <D2DI3G.CnB@eskimo.com>
Sender: usenet@eskimo.com (News User Id)
Organization: none so far
X-Newsreader: WinVN 0.93.10
References: <D0wsx8.6zC@prl.philips.nl> <johnson.788537339@hal.cs.uiuc.edu> <3eb0je$dje@jhereg.mdd.comm.mot.com> <D21EF7.BF6@eskimo.com> <3f349i$agl@hawk.ee.port.ac.uk>
Mime-Version: 1.0
Date: Sat, 14 Jan 1995 02:11:39 GMT
Lines: 67
Xref: glinda.oz.cs.cmu.edu comp.lang.misc:19919 comp.programming:13980 comp.software-eng:29504 comp.ai:26437

In article <3f349i$agl@hawk.ee.port.ac.uk>, sgt@sis.port.ac.uk says...
>
>In article <D21EF7.BF6@eskimo.com>, hgeorge@eskimo.com (Harry George) 
says:
>>
>>In article <3eb0je$dje@jhereg.mdd.comm.mot.com>, 
>>dcampbel@jhereg.mdd.comm.mot.com says...
>>>
>>[deleted material]
>
>>I agree 100%.  Several years ago I was managing a data modeling and
>>process modeling organization.  We used all the best CASE tools.
>>The more we tried to accurately (no, let's say _adequately_) model
>>the clients' domains, the worse the tools looked.  I started looking
>>for alternatives and realized that a good, readable OOPL
>>was semantically more powerful and syntactically more elegant for 
>>exactly the same job -- and could be type-checked and executed as well.
>>
>>-- 
>>Harry George
>
>I don't think that I understand! I thought that CASE tools like 
>"softwear thru pictures" and "cohesion" were used to help you construct 
>better, more efficient and larger programs in languages like C++, 
>smalltalk, ADA and modular-3.
>
>Help!
>
>        Si.
>
>

Yup, they are supposed to help you.  But in real life you get bogged down
with configuration management of the CASE models, and arguments over
just how detailed to get, and confusion over the exact syntax to use
for a given semantic concept.  As you put in "guidelines", and
"conventions" and "change boards", the CASE world gets out of hand.  It
is great for a few hand-waving pictures, but it just isn't scalable
to thousands of data structures and related processes.

SO if you want to capture and use a serious collection of domain
models, you find yourself wishing for a good, readable OOPL.  Think
about it:  What does a CASE tool promise?  It promises to draw labeled
DAGS and to associate the nodes and arcs to fairly simple record
structures.  Now think about all the abstraction mechanisms available 
in an OOPL:  Everything the CASE tool provides, plus far richer 
type constructors and type checking.  Plus executability.  Old OOPL's
didn't offer to draw DAGS, but visual programming is attacking that.
(Even so, visual programming is being oversold at bit.  After I get 
a data model with several hundred entities plastered on the wall, 
I need a scorecard -- an ascii listing that looks suspiciously like
a program listing.)

Now, I certainly would use a CASE tool if my alternative was to write 
in C++, smalltalk, Ada, etc.  That's why I stipulated a "good" OOPL.
But if I can get power, elegant syntax, and html based literate 
programming, the CASE tools begin to lose the battle.  I'm saying
Modula-3 seems to have the edge here.
-- 
Harry George
email: hgeorge@eskimo.com
smail: 22608 90th Ave W / Edmonds WA 98026
quote: "The reasonable man adapts himself to the world;
       the unreasonable one persists in trying to adapt
       the world to himself.  Therefore, progress depends 
       on the unreasonable man." G. B. Shaw

