Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!wuarchive!brutus.cs.uiuc.edu!apple!voder!pyramid!athertn!hemlock!mcgregor
From: mcgregor@hemlock.Atherton.COM (Scott McGregor)
Newsgroups: comp.software-eng
Subject: CASE, the Little Red Hen, and Stone Soup
Message-ID: <28596@athertn.Atherton.COM>
Date: 10 Aug 90 18:25:17 GMT
Sender: news@athertn.Atherton.COM
Reply-To: mcgregor@hemlock.Atherton.COM (Scott McGregor)
Organization: Atherton Technology -- Sunnyvale, CA
Lines: 120


 
CASE, the Little Red Hen, and Stone Soup
(An CASE industry analysis in fairy tale terms)

Copyright 1990 by Scott L. McGregor

I am concerned about where our industry is going, or rather where it isn't.
I am afraid that we are creating lots of capabilities and possibilities and
few realities and actualities.  I am afraid that there will be a backlash
against what we have done by people who seek for solutions, hear that
CASE has the answer, look past the possibilities in search of the solutions
and disappointed turn away, only to criticize the industry as having
inflated claims.  

I believe that many people want applications that will improve their
ability to develop software--with little effort in understanding their
current situation, or where they want to go.  Of course, not everyone
is this way.  Some people DO know their current situation and where they
want to go, and they chose CASE tools to help them get there.  These
are the early adopters of our industry.  But most people don't have the
time, the technical background or the inclination to find out the answers
to these questions--they want the answers to be provided to them.  

In some ways, this is like the AI/Expert Systems situation a few years ago.
Expert Systems provided a capability for creating great advisory systems.
People were excited.  Everyone wanted these powerful, accurate and
(hopefully inexpensive) advise systems.  But no one wanted to build
them.  Like in the story of the Little Red Hen, everyone wanted to
eat the bread, but no one wanted to bake it.  It is a lot of trouble
to build and expert system, and you have to be one to build one.
But if you are one, you need one less than someone else who is not.
Of course a generic advice system could be used by anyone, a specialized
one by only a few. How many are those few? Are there enough to pay for
the effort of building such a system? Will they pay enough? Will you be
able to find them?  Will they be able to find you?  It is so problematic
that people pushed the general purpose engines and few specialized systems
that really solved problems for people were sold.

Several years ago, when PCs were brand new, a fellow I used to know 
brought home a PC from work and went to show it off to his wife.
They sat down, booted the machine from a floppy, and then he proceeded
to show her how you could start the basic compiler, write a program,
and have it read something such as a name and then have it
print out something.  She typed in her name. He had it print out "Hello,
Fiona!"
She was non-plussed.  "Have it tell me something new!" she said.
He  had type in another name and it printed that out.  "No," she said
"have it say something NEW. Something you didn't tell it."  He explained
to here how with programming you have to know what you want up front
and then the computer does it for you.  He showed her how it could
multiply large numbers."  She was still non-plussed.   "What good is it
if you have to know everything already when you start?" she stated as
she left the room.

I am afraid that this is the situation for all too many software development
managers out there.  They have too much to worry about to have time to
spend on figuring out what they want.  "Do you want to use the
Yourdon methodology? Jackson? James Martin...?"  someone asks.
"I don't give a hoot about any of them--which one will make us
instantly productive."  Meanwhile we develop more systems in the
vein of X-windows: policy free systems. And we suffer from the
same problems of Xinitis--configuration is complex because so many
decisions are left to the user.   But look at how all the
clamour is over added policy, in the form of Motif and Open Look and
style guides.  Companies regard it as an unfortunate burden that they
have to have some expert define and build the company standard configuration
file that everyone else will get.

For many people CASE is stone soup.  In the story of Stone Soup, some
soldiers wander through a small village and ask everyone for some food.
But everyone is suspicious of strangers, especially soldiers who once they
find out what good food they are keeping might steal it by force of arms.
So everyone says they have no food to give them.  Finally, the soldiers
give up and build a great fire in the central square.  They borrow an old
cauldron and fill it with water and put it on the fire.  Then one of
the pulls out some "magic stones" out of his knapsack and puts them in
the pot.  The villagers crowd around to see what the soldiers are doing
and the soldiers say "we are making stone soup."

"I've never heard of stone soup" says a villager, "is it any good?"

"Well," says one of the soldiers, "we'd be glad to let you taste,
but we have so little.  But if you could contribute something to the pot, a
carrot, an onion, a potatoe, a soup bone or a bit of beef, we'd be
glad to share with us, besides, the contribution might improve the
taste." 

The villager agrees and runs home and quickly returns with some carrots.
Other villagers become curious as well and each runs home and returns
with something to contribute. Soon the pot is full of vegetables and
beef and chicken.  The soup is served to everyone and everyone agrees
the soup is the best they have ever had thanks to the flavor of the
magic stones. 

For many CASE products it is much the same, I think.  The real value
is not in the product, but in the ingredients provided by the customer
themselves in their customization, self-examinations and preparation
for use of the new system.  The CASE product is merely the catalyst.
This isn't to put down the hard work of the developers of CASE
products--I am managing development of one myself.  But I am concerned
by the knowing glances I get whenever I demonstrate one of these  magic
stones.   

There are some big challenges ahead for our industry.  We need to develop a
large CASE market with a taste for stone soup, or we are going to have to
learn to make some hard choices and see how individuals respond to
pre-prepared onion soup, chicken noodle, chunky beef and creamy mushroom
soup.  Soups where you just add water, not water where you just add your
favorite soup ingredients.   The challenge is particularly stong for those
of us working on Integrated Project Support Environments (IPSEs)
We may have to build systems with some of the decisions already
made at the factory--and we may succeed or fail based upon the correctness
of our predictions. 

So here's to better soup and may we not starve while waiting for
someone else to bake the bread.

--Scott McGregor