Comment on Microsoft Settlement

From: "Charles Karney" <ckarney@sarnoff.com>
Reply-To: charles@karney.com
To: microsoft.atr@usdoj.gov
Subject: Microsoft Settlement
Date: Sat, 5 Jan 2002 15:51:02 -0500

From: Charles Karney
      702 Prospect Ave.
      Princeton, NJ 08540-4037
      E-mail: charles@karney.com

To:   Renata B. Hesse
      Antitrust Division
      U.S. Department of Justice
      601 D Street NW
      Suite 1200
      Washington, DC  20530-0001 
      E-mail: microsoft.atr@usdoj.gov

Date: January 5, 2002

Re:   Proposed Settlement in United States vs Microsoft

I have used computers professionally for the past 30 years.
Currently I am employed by Sarnoff Corporation, Princeton, NJ.
In this position, I have, for the past two years, coordinated
the research computing needs of a start-up company, Locus
Discovery, Inc.  Locus Discovery uses novel computational
techniques to design small-molecule drugs, and at the core of
this method is software running on a 2000-processor computer
cluster running the Linux operating system.

Linux has provided a incredible opportunities to deploy massive
computing resources in an extremely cost-effective manner.  This
has resulted in great opportunities for America's small
companies to realize innovative technologies.

In this context, I find that the proposed settlement in the
United Status vs Microsoft to be too narrowly drawn and to allow
Microsoft too many opportunities to interpret the settlement to
its advantage and to the detriment of the consumer.

The proposed settlement governs the market in desktop PCs
running the Windows operating system.  In most companies, this
computing environment coexists with other platforms: Windows and
non-Windows servers, desktop systems running other operating
systems, handheld computing devices, etc.  With the settlement
of this case, the U.S. has an opportunity to foster as wide a
range of choice as possible in all these areas of computing and
to prevent Microsoft from extending its illegally maintained
monopoly.  This will allow U.S. companies choose the best tools
for a particular job enhancing the overall competitiveness of
the U.S. economy.

I believe that there is a straightforward way for the U.S. to
ensure this while allowing Microsoft the continued "freedom to
innovate":

Microsoft should be required to publish technical specifications
for all its network protocols, all its data formats, and all its
application programming interfaces.  The specifications should
be sufficiently detailed to allow others to offer competing
implementations and Microsoft should explicitly allow such
implementations, providing, if necessary, royalty-free licenses
to permit this.

Let me provide two examples:

(1) Microsoft should publish the specifications of the file
formats using by its "Microsoft Office" suite.  This will enable
other office suites to read and write files in a compatible way.
These office suites will likely be available on non-Windows
platforms, and this will mean that consumer will not be forced
to purchase a Windows platform merely because of the need to
read documents produced by Microsoft Office.  A consumer may
still choose to use Microsoft Office (because he perceives that
it provides to the best way to produce his documents) and he may
choose to use Windows because of his perception of its benefits.
However, he will now have a choice.  This will be at no cost to
Microsoft's flexibility to create good software.  On the
contrary, it will offer an incentive to Microsoft to improve the
implementation of its office suite since it will now need to
compete against comparable compatible products.

(2) Microsoft created a protocol called "Server Message Block"
to permit files and printers to be shared between Windows
systems.  There is a free implementation of this protocol called
Samba which allows the sharing to take place between Windows and
non-Windows platform.  This offers a clear benefit to consumers.
Unfortunately, the Samba implementation is hampered by the need
to "reverse engineer" the details of the protocol.  Microsoft
should remove this impediment by publishing the protocol and
specifications of additional protocols, e.g., for user
authentication.

Some of the restrictions in the proposed settlement are overly
restrictive and should be removed.  Two such examples are:

(1) The proposed settlement limits the provision of information
to companies which Microsoft considers to be bona fide
businesses.  This restriction would exclude the "Open Source"
community which is responsible for Samba and Linux.  Microsoft
should make the information available to all, e.g., by
publishing it on a publicly available web site, and the needed
royalty-free licenses should permit implementations by anyone.

(2) The proposed settlement also allows Microsoft to avoid
disclosure of information which would "compromise security".
Unfortunately, this provision is open to abuse by Microsoft.
Good security protocols can be (and are) published in full
without compromising their security.  This provision creates a
perverse incentive to Microsoft to craft poor security protocols
which rely on "security through obscurity", a rightly derided
mechanism for computer security.

By requiring the publication this information, the U.S. would be
stimulating innovation in the entire computer industry.  The
situation would be similar to two other periods where the
establishment of computing standards lead to explosive growth:
the creation of the PC market by the publication of the hardware
standards for PCs, and the creation of the world wide web by the
publication of standards for HTML and the underlying networking
protocols.

I urge the U.S. to reject the proposed settlement and to create
one which will have a clear benefit to the consumer.