Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!elroy.jpl.nasa.gov!decwrl!netcomsv!jls
From: jls@netcom.COM (Jim Showalter)
Newsgroups: comp.software-eng
Subject: Re: Software Through Pictures
Message-ID: <1991May22.223228.5483@netcom.COM>
Date: 22 May 91 22:32:28 GMT
References: <136@mishima.mishima.mn.org>
Organization: Netcom - Online Communication Services  UNIX System {408 241-9760 guest}
Lines: 51

I'm not a big fan of any of the pictorial CASE tools. I think they impeded
a good engineer, and they give a bad engineer a false sense of security
("I can't have a bad design--just look at these DIAGRAMS!"). For pictorial
design work, the all-time greatest tool is a whiteboard in a well-lit
room. Indeed, one project I'm familiar with has a room full of whiteboards
on which they draw diagrams, and when they're done they have a clerk come
in and transcribe what they drew into a pictorial CASE tool so that they
have a nice printed copy--they never touch the CASE tool themselves because
they find it hard to manipulate, tedious, and irritating. Before deciding
to hire the clerk, they were seriously considering just PHOTOGRAPHING
the boards. Clearly, the people on this project don't regard the tool as
a design tool, only as a design capture tool--a sort of glorified MacDraw.
The design takes place earlier, when the designers are scribbling stuff on
the walls, backs of envelopes, etc.

People always seem to be looking for something that will "automate" the
design process. There is no magic bullet. If design was easy, anybody
could do it: the truth of the matter is that lots of people CAN'T do
it, and there's no reason to expect them to. You want a good design,
hire good designers (and get out of their way to the maximum extent
possible).

But enough philosophy: there is a quite practical reason I'm dissatisfied
with pictorial CASE tools--they don't scale. I've used structured analysis
tools from a prominent vendor, and any time the complexity got very high
(e.g. lots of bubbles on a screen and/or lots of levels of diagrams), the
usefulness of the tool rapidly approached zero. I'm familiar with several
project leads who had similar experiences: they started out all jazzed about
the tools, but eventually put them back on the shelf when they turned out
to be inadequate for large complex projects. I concede that for simple
small projects the tools are adequate, but, ironically, for projects of
such diminutive scope the need for a tools is far less acute: in short,
I've only seen such tools used successfully on projects where they weren't
necessary in the first place.

None of this is to say that I hold out no hope that CASE tools can live up
to their billing, just that I don't think it has happened yet. When the
tools are able to truly guide a designer toward producing a good design,
and when they are able to tackle extremely large and complex projects (which
presupposes that someone comes up with a notation for architectural level
design), then I'll be more interested in using them.

Note, also, the careful distinction made concerning PICTORIAL CASE tools.
There are all sorts of CASE tools that work great: editors, compilers,
linkers, debuggers, performance and coverage analyzers, CM systems, etc.
It's just the front-end design stuff that I'm not happy with.
-- 
**************** JIM SHOWALTER, jls@netcom.com, (408) 243-0630 ****************
*Proven solutions to software problems. Consulting and training on all aspects*
*of software development. Management/process/methodology. Architecture/design/*
*reuse. Quality/productivity. Risk reduction. EFFECTIVE OO usage. Ada/C++.    *