[M4IF Discuss] IEEE Spectrum Article

[Kris Huber]

There is a short article in the "Speakout" section of the IEEE Spectrum
magazine.  As an engineer who has been involved in MPEG to a limited extent,
and now seeing M4IF encourage the licensing terms to get established and
clarified, I thought it interesting to take a step back and look at this
standardization/licensing situation, and this is the first article I've
noticed on the subject.
S. J. Frank, "Can you patent an industry standard?", IEEE Spectrum, Mar.
2002, pp. 14-15.
The author is with Testa, Hurwitz & Thibeault LLP.  A link to the article is
http://www.spectrum.ieee.org/spectrum/mar02/departments/speak.html but it is
only available in the members-only portion of the site.  Many university
libaries subscribe to this magazine, however, as well as all members of The
Institute of Electrical and Electronics Engineers.
Best regards,
Kris Huber
P.S.  Here is a brief summary for those who may not be able to access it or
who prefer an abbreviated version:
"The standards-setting process is 'open', but that does not mean 'free'.
Increasingly, standards writers are patenting their contributions and
charging adopters."
For years contributions encumbered by patents were rejected by standards
bodies as a matter of policy.  The rules have loosened in recent years to
avoid foregoing the best technology.  Conditions under which standards
bodies today accept patented contributions are typically that they "disclose
relevant patents or patent applications before their technology is
considered, and must agree to grant licenses on a nondiscriminatory basis
and on reasonable terms."
"Companies that deviate from these guidelines may face the wrath of the U.S.
Federal Trade Commission (FTC), which has treated violations as unfair
competition."  An example is discussed about Dell Computer, who was a member
of VESA who created the VESA local bus for accelerated graphics access ('486
days).  After the design was widely adopted, Dell "contacted various VESA
members and told them that by using the VL-bus standard, they were
infringing a Dell patent."  To settle the resulting FTC complaint Dell
agreed not to enforce that patent, and "also agreed to be prohibited from
enforcing any of its patent rights intentionally withheld from any
standards-setting organization in which it participated in the future.
Various reporting requirements assist the FTC in ensuring Dell's
compliance."
In addition to FTC proceedings, private lawsuits are possible.  "Enforcing a
patent witheld from a standards body against adopters of the standard can
qualify as patent misuse or, depending upon the circumstances, constitute an
antitrust violation."
The author then discusses the difficulty in defining what it means to
license on terms that are "reasonable" and "nondiscriminatory."  One
difficulty of those wishing to adopt a standard is how to know whether they
are being treated fairly.  A seemingly fixed figure like a percentage can be
a "slippery" amount.  A difficulty is defining "percentage of what"?  Patent
holders may want the percentage to apply to a whole product while an adopter
may view it as a percentage of the value they perceive to be added to a
product by adoption of the standard.  Also, large-quantity discounts in the
IP terms can give advantage to larger firms.  Such volume discounts are
usually upheld as legitimate as long as they are offered to everyone.  An
initial one-time fee for the license is also often upheld, even though it
can sometimes be much higher than the administrative costs that justify its
stated existence would require.  The size of the issue fee may also vary
depending upon the identity of the adopter, as "the licensor is not obliged
to divulge the criteria it applies."
"Another license term patent owners seek is a grantback, meaning the right
to use technology created by licensees."  This can disadvantage the adopter
compared to the licensor.
"An economist might have little sympathy for companies that feel they are
paying too much (or giving too much away) in order to implement an industry
standard.  After all, if enough of the industry feels that way, the standard
will simply wither and die.  But this view assumes perfect information," and
it is that information that is most lacking according to the author.
Standards-setting processes are "transparent" (i.e., "see-through", allowing
one to understand) when it comes to technical criteria and mostly
transparent on the fact that relevant proprietary rights exist, "but the
process is essentially 'opaque' to licensing of those rights." 
The author's conclusion is that this results in industry players having to
"make decisions in total strategic darkness."  A company may feel forced to
adopt, whatever the terms.  Because of confidentiality, the licensor "can
keep the blinders on everyone by refusing to discuss how its license
provisions vary among adopters.  In the end, the realities of standards
adoption may undermine the economic logic of the market.  Ignorance is
bliss--if you own the patent."
The author believes the solution is for standards bodies to deal with
license terms as thoroughly as they deal with technical merit of
contributions.  "Participants holding (or intending to develop) proprietary
rights in the standard should be required, at the earliest possible stage,
to submit a detailed licensing proposal.  The proposal would essentially be
a form license, circulated to standards participants and accessible to the
public at large.  Should rights-holding participants envision terms that may
vary across the industry, let them say so and explain the basis.  Then let
industry judge the standard by financial as well as technical criteria,
before the snowball effect" of widespread adoption occurs and forces
acceptance of the standard at almost any cost.