Bill Gates may be the most boring billionaire on the planet. However,
his testimony at the latest installment of Microsoft's antitrust case was
more interesting than anything served up by any other witness. It was
also more credible. He made several points that should infuriate Open Source
advocates, but that doesn't mean they weren't true, and shouldn't be
taken seriously. His testimony, and Microsoft's defense, suggests that there
is something fundamentally wrong with the whole notion of "structural
remedies" in high technology. Is there a better way to go about all this?
Dull, dull, dull: No matter how tech writers have tried to color
the testimony in the latest Microsoft trial, it has been mostly a
retread of all the litigation that has preceded it. We have seen the same
witnesses make the same claims they have made since 1998. In earlier cases the
National Association of Attorneys General usually posted trial transcripts
within hours of testimony. This time, they
have been running a week behind.
Last week, I suggested a lack of inspiration had caused the states
to put on an appallingly bad case, featuring bad research, an inept
strategy and poorly prepped witnesses. Instead of focusing on the impact of a
series of proposed penalties governing Microsoft behavior, they handed the
company a second chance to retry the original case, two years after the
facts were established, and the company's guilt was confirmed. The state's presentation was so hard to follow that Judge Colleen Kollar-Kotelly found herself agreeing to most of the objections made by Microsoft's lawyers. I suggested that it was hard to imagine a more poorly presented case.
This week, Microsoft's team nearly returned the favor. The company's witness
list was full of clients, "researchers" funded by Microsoft, ex-employees
with a bag full of stock options or clueless executives who needed a
favor from Redmond. The low point was hit early on when Jerry Sanders,
AMD's sometimes erratic chairman, stood up to testify. It was clear from
the outset that he was unfamiliar
with the facts of the case and that he took the witness stand only
because he needed support from Microsoft. (And
he got it!)
The only witness who stood out from this sorry crowd was William
Gates III himself. Unlike his testimony in the earlier antitrust trial where
he chose to be too busy to attend in person and almost perjured himself
on the witness stand, this new, PR-prepped and properly dressed Gates
gave some of the only credible testimony anyone had heard over the last two
weeks. While reporters for The Register dismissed it as nothing more
tedious advertisement for Microsoft," it is just possible that Gates saved
Several experienced Microsoft haters, including Mathew Ingram
of Toronto's Globe and Mail and
The Boston Globe's Hiawatha Bray came away fairly impressed. Compared to almost everyone else that testified, Gates
made some credible points. At various points in his oral and written testimony (or at least the testimony prepped for him by his lawyers
and PR people), Gates basically conceded that the Microsoft code base was bloated, complex and under-documented. It should come as no revelation to learn that
Windows is not so much an OS as it is a patchwork of operating platform
elements connected together to provide some level of backwards compatibility. He
suggested, quite correctly, that the only firm likely to be able to
clean up that code, and modularize it, was Microsoft itself. By that logic,
if the court wanted Microsoft to make changes in the code, it would
have to depend on Microsoft to make them. How could that possibly work?
Further, Gates and Microsoft maintain that they might not be able to
comply, unless they radically simplified the code, effectively
abandoning millions of legacy users and marooning thousands of small applications
developers in the process. Its an infuriating suggestion, and it's
Gates suggested that many of the distinctions being made between
applications and the operating platform are artificial. Code is code. There really
is no logic that mandates specific boundaries between applications and the
operating platform. As he said, "There is no clear dividing line
between where a particular [application] block ends and the operating system
begins." Developers of older, proprietary computing systems frequently combined
elements of what is now considered the operating system with
application elements. If Microsoft does the same thing on occasion, how does that
make it different than developers who have gone before it? And he asked
a fairly obvious question: Is the court going to successfully define
what constitutes the proper boundaries of an operating platform or
judge where APIs should be made available?
The alternative might be to open up the entire code base. Gates
suggests, quite plausibly, that the result might be hard to interpret, and the
most interested parties would come from the Open Source community. He
suggested that the result would likely be relatively little in the way of
commercial software and lots of Open Source freeware of varying quality. He argued
that Linux platform developers would almost certainly
emulate Windows APIs. Gates claimed that Microsoft would then be competing
against freeware. He suggested that this would quickly kill the Windows
operating platform as a commercially viable product for the company. This
suggestion has infuriated many Microsoft haters, but it is quite possibly true.
Gates claimed that whatever flaws there were in Microsoft's past
practices, it had been much more supportive of open platforms than its
competitors. He pointed out that, unlike Apple, Microsoft did license its code to
OEM equipment makers. He pointed out that he had urged Apple to do so as
far back as 1985. He suggested that Apple's unwillingness to do so led to
it being marginalized, rather than anything Microsoft did. He went on to
argue that Sun Microsystems could have positioned either the Sun OS or
Solaris as an alternate OS platform for Intel hardware, but chose not to do so.
He suggested that both companies had tried to set up closed
hardware/software/application environments in ways that Microsoft had never considered. He suggested that today Sun is facing a threat similar to the one Apple had faced, now that lower-priced servers running on Windows 2000 or Linux are
coming onto the market. He suggested that the blame for this situation hardly
lays with Microsoft. This is probably true as well.
To be sure, Gates made a number of fairly absurd claims as
well. He claimed that Microsoft "doesn't track development costs." If that was true, no wonder his COO resigned. Gates also claimed that the
proposed settlement would allow developers to make use of the Windows
trademark even if there was little or no Microsoft products on a PC or other
system, and that they, rather than Microsoft, could judge for
themselves if a product was Windows-compatible.
One claim woven throughout Gates' testimony could be called the
monopolist defense. Throughout his testimony, Gates returned again and again to an
argument made by monopolists and
their defenders since the days of John Pierpont Morgan and
John D. Rockefeller. Gates suggested that most of his company's actions were
aimed at providing consistency and rationality to the software
marketplace. In their time, both Morgan and Rockefeller claimed to be "rationalizing"
the oil, steel and banking markets. Rockefeller never denied
being a monopolist -- he simply stated that whatever he was doing, that the
price of kerosene was a lot lower and had more predictable quality than it
did before he put his competitors out of business.
In and out of court, Gates and his allies make similar claims. They
have suggested that software is a "natural " monopoly. Gates' testimony is
full of references to the need to make sure all code elements work together.
Outside of court Microsoft defenders point to the many incompatible
flavors of Unix and the meager financial statements of most commercial Linux
companies as evidence of what happens without such rationalization. And. like
Rockefeller, Gates claims that the cost of a fully loaded desktop PC costs less,
does more and works better than would be the case if Microsoft had not used
its dominance to dictate standards and rationalize the market.
If that is all that Microsoft's monopoly has done, where is the
harm? And what should the penalty be?
Microsoft and the Justice Department have recently suggested that
the harm isn't necessarily all that great. This is one reason that their
suggested penalty is relatively minor.
Outside the United States, regulators have often believed that such
"rationalization" is a good thing. In the past, European and Asian governments used to
turn a blind eye to monopoly behavior, and permitted the formation of
cartels that "rationalized" markets in products and services in everything from
chemicals to banking. Monopolists and cartels don't lay people off and
they don't go bankrupt. Only recently have these governments noticed
the cost of monopoly is stagnation, as they realized that their
"rational" economies were falling behind that of the United States.
Outside the United States, government agencies are re-examining other "natural"
monopolies in the telecoms and electric power markets. In Europe, especially, cartels and would-be monopolists have been subject to dawn raids, and heavy
fines, typically 10%
of the companies' gross receipts. In Microsoft's case, such an
approach would have cost the company $2.5 billion dollars for every
infraction of antitrust laws. That wouldn't put Microsoft out of business, but it
would certainly get its attention.
And it would have been a lot easier to manage and adjudicate than
any of the proposed "structural remedies" being discussed in and out of
court over the last couple of years.