The Psychotic State
			      Number 15
			     May 1, 2000


	       Why Bad Software Is Good For Bill Gates


I haven't written one of these in a while.  It's not that I haven't
been crazed - quite the reverse.  In fact, I've been too nuts to
write.  The true psychotic can only rant when he's not fully sedated.

This is an essay in two parts.  The first half is a Psychotic musing
on how Microsoft got to the monopoly status it now enjoys.  As you
will see, my ideas are a bit off kilter from some of the other things
you may have heard.  This first part I've been meaning to write for a
while.

Next I will play the pundit, and voice an opinion on what could be
done about it, or, more accurately, what might happen if Microsoft
were broken up.

Before I begin, I thought I could state some of my assumptions that I
won't be explaning here.  I'd be happy to discuss them if there are
those who disagree, but for purposes of this essay you should regard
them as a given.

*) Microsoft is a monopoly.  (This is now much less debated, at least
by everyone not in Microsoft's employ, since it is a matter of fact
according to the Federal courts.)

*) Monopolies serve their own interests, not the public's.  (All
private enterprises serve their own interests, but at least if we
believe the arguements of free market economists, the presence of
other companies in their market forces them to serve the public's
interest in order to compete.)

*) Microsoft's software is entirely unremarkable in terms of the
features it offers.  (MS is in general a very _non_ innovative
company, despite what they usually proclaim.  I don't entirely doubt
their sincerity as they say it; I just don't think they can make any
sort of reasonable case to support their contention.)

*) Microsoft software is not particularly easy to install or use, nor
is it especially robust, secure, or bug free.

There are those who may feel that these views make me some sort of
anti-Microsoft zealot.  I will say in my defense that I feel this way
about a great deal of other software as well, and that I believe that
I could support all of these points with calmly reasoned arguments
that I just don't have room for.


Part I

I have always found it surprising that Windows 3.1 took over the
world.  It has always struck me that Windows 3.1 was extraordinarily
difficult to use, since it has a truly terrible interface.  Also,
since there was no real thought given to how different programs would
interact with each other, it was very difficult and costly to
maintain.  (These observations have been only somewhat - not entirely
- mitigated by the introduction of Windows 95 and NT.)

During the heyday of Win 3.1, I only rarely used it.  I was at the
time in an environment quite different from that of the typical
American business: I was surrounded by technical people.

There are quite a number of usual explanations as to why technical
people don't use Windows as much as the rest of the world, including
lack of support for the applications they want to use, community
convention, training, and others.  I'm sure that all of these play
their role in the state of affairs, but I believe there is another as
well.

Consider the following observation: If the installation, use, and
maintenance of Windows requires more technical expertise than other
systems, why is it that it is less prevalent where there are technical
people around?  To me, this is a deeper question than it seems.  It
seems to fly in the face of the usual sort of economic argument: if
the price of something is high, you'd expect to see poor people using
less of it than rich.

In intro economics classes they sometimes mention the concept of a
Giffen Good.  This is a product that has the unusual properties that
(a) it is inferior to other products, and (b) the more expensive it
is, the more people consume.  That's the reverse of what usually
happens - when the price of something is raised, it is supposed to
sell less.  Mr. Giffen discovered this type of good during the last
century during one of the Irish potato famines.  (There is, evidently,
still some scholarly discord over whether his observations were
correct.  Not to worry - what I'm saying is likely to be even more
controversial.)

Suppose you were a dirt poor worker during the famine.  Let's imagine
there are two kinds of food you can buy: potatoes and meat.  Meat is
considerably more expensive than potatoes, and your preference
function is this: you have to get 2000 calories a day, but within that
constraint you will eat as much meat as you can, the rest of the
calories you make up with potatoes.  Since it is a famine time, your
income is very low, and food consumes the vast majority of your
budget.  You can't afford to eat much meat, but there's a small amount
of money left over with which you can get just a bit.  Almost all the
calories you get come from eating potatoes.

Now consider what happens as the famine worsens and the price of
potatoes goes up.  You are now effectively poorer and have less room
in your budget for your one luxury - meat.  You have to make up the
calories by eating more potatoes, even though you like them less and
they are now more expensive than they were.

I believe that something akin to this operated with Windows.  Most
American businesses are quite poor - not in money, but in people with
technical training.  Early on, businesses used multiple operating
systems: MSDOS, Mac, CPM, and others.  No business was an island, and
it had to exchange documents with other organizations.  Therefore
there always had to be a machine around running, say, DOS and MS Word.
(In technical organizations this situation still obtains today - there
is always a box with NT on it to run the occasional Windows app.
Nowadays the engineer would have it dual boot with Linux for when s/he
needs to go back to real work!)

What happened as Windows came out was the price - in terms of
technical expertise - of running the Microsoft product was raised.
The tech person in the office _had_ to support Windows - too many
other organizations were using it to ignore it completely - but s/he
could no longer afford the time to support the other systems as well.
The expense of running Windows squeezed out the others.  Windows was a
Giffen Good.

As usual for someone in a Psychotic State, I can't support my theories
with more than anecdotal data, but I do think that you can really
understand why Windows is as dominant as it is - and not better than
it is even after all these years - by thinking about this explanation.


Part II

The latest develpment in the Microsoft antitrust trial is that the
plaintiffs, the federal and many state governments, are going to ask
the judge to order the breakup of Microsoft.  According to the latest
plan - the details of which I must admit not having read - Microsoft
would be cleaved into two companies, one controlling the operating
system, the other the office suite.  (For purposes of this essay I'll
call the two resulting companies WinCo and OffiSoft.)

One of the reasons I hate the punditocracy in the American media is
that they usually offer predictions of the future rather than analyses
of the past.  Usually these people proffer endless expostulations on
policy, without even knowing the relevant facts.  Usually these
discussions are little more than matters of uninformed opinion.

I'm Psychotic - I figure it's time for me to do the same.

For a long time I thought that breaking up Microsoft wouldn't work.
In the last few weeks I have come to change my mind.  Here's why.

I'll imagine that what happens is that WinCo and OffiSoft are two
separate companies.  They are allowed to compete, but not to
collaborate or use a common code base.  WinCo can still create a new
word processor, and, if it wants, OffiSoft can try to create a new OS.
WinCo will be prevented from the most extreme examples of bundling
that were perpetrated by Microsoft, but they can write whatever other
software they want to.

For a long time I reasoned as follows: in this situation, WinCo wins.
I do not believe that Microsoft controls the office suite because
their tools are better.  Rather, their control of the OS allowed them
to crush the competition.  So, for example, despite consumers clearly
expressed preference for Word Perfect way back when, MS could just
keep plugging away until their changes to the OS and their bundling of
Word with the installation prevented Word Perfect from keeping up.  We
saw it with Lotus 1-2-3 and Excel, and then later, with Netscape
Navigator and Internet Explorer.

Here's what I recently realized: The CEO of OffiSoft is no fool - s/he
is not going to buy the Microsoft version of history: Microsoft
dominates all these categories because they happened to create the
best products in _all_ of them, through excellence and innovation.
No, s/he knows what the future will bring.  The OS is not an
inherently valuable monopoly to have by itself, as WinCo does, so
they'll be looking to go after the office productivity tools sooner or
later.

What does OffiSoft do?  They immediately act to make sure that Windows
is not the only game in town.  They port all their tools to Palm OS,
to Linux, to Solaris, etc, and they make sure they run as well on Mac
OS as they do on Windows.  (An aside: those who argue that the MS
monopoly serves the public interest will please kindly explain why it
is good that you have to buy a Win CE device in order to open a
spreadsheet.  You can't run Word or Excel on a Palm - and I daresay
you never will, at least not until Microsoft feels, as they do now
with MacOS, that Palm OS is not a real threat to their core monopoly.
But I digress, and I promised I wouldn't debate my four assumptions.)
OffiSoft makes sure that Internet Explorer runs the latest version of
Java.  OffiSoft has the widest used productivity tools, and they can't
afford to stop watching their back.  WinCo will happily do to Excel
what Microsoft did to 1-2-3.

Just imagine how well the consumer is served when you have OffiSoft
being really hungry!

WinCo has a different set of options.  Knowing that the OS is nothing
without application support, I think they have to go after OffiSoft.
They aren't fools either, and they can see how easy it is today for
Microsoft to push Apple around, because Apple needs Microsoft to keep
supporting the Mac version of Word, etc.  With Windows only one of
many possible OSes on which the tools run, WinCo can never be too
sure.  Maybe WinCo goes out any buys Corel or something, and tries to
push their tools.  (If they have real chutzpah, they'll turn around
and claim that the Corel tools are "part of the operating system".  As
I said, I hope that the settlement precludes them from bundling
again.)

WinCo will also, doubtless, continue its assault on Sun, trying to
make NT into an enterprise OS.  They will continue their "Windows
Everywhere" campaign - going after TV set top boxes, and yes, palm top
devices.  In this new world, WinCo no longer has the option of
hermetically sealing their software, of making the user buy everything
WinCo in order to get the simple spreadsheet and word processor s/he
really wants.

With no further argument, I offer my Psychotic version of the future.
I hope I'm right.  It would be a real improvement.