Just say No to RAND! (and RF)

I've been browsing some of the comments about this recent
working draft, and find a lot of things to agree to in comments
arguing against permitting W3C Recommendations (or in
effect, "standards") to entail global licensing taxes.  I'll hope
that this isn't a case where public comment will be ignored.

The "Patent Policy" TR should correct its usage of "Royalty
Free" (RF).  Every RF licence I've seen has involved hefty
up-front fees, as the _alternative_ to "reasonable" royalties
based on unit sales.  Excluding 4(e)5 makes this TR's usage
be confusingly different from common practice, and calls into
question the intent to truly exclude such up-front fees.  At
the very least, "Zero Cost" licensing needs to be a "mode".

I'm also pleased to see the number of comments that seem
to be coming in at the last moment.  It seems that this TR
was not particularly well known until recently.


I - NO ADDITIONAL STRUCTURAL BIASES

As a world-wide organization claiming to represent all
sectors of the web community, W3C should not be
granting additional structural preferences to the first-world
corporations which are the primary beneficiaries of
today's malfunctioning patent systems.  (The problems
of the US PTO are mentioned in the TR, although that
mention appears to treat that bug as a feature.)

Any kind of mandatory fee-based licencing is by
definition discriminatory; it can't be "RAND".  In
addition, adoption of such fees by W3C would in
effect be creating a world-wide technology-based
taxing authority, furthering first-world profit agendas
at the expense of the rest of the world.


II - OPEN SYSTEMS REQUIRE NO FEES

The genesis of the Web was in "Open Systems" processes
where the only real barriers to entry were ability to master
technical complexity and to deliver real value in the form
of a service (or product).  Encouraging RAND, as opposed to
zero cost licensing, creates new kinds of barriers.  More, it
turns the development paradigm into "Closed Systems", in
the sense that it excludes the Open Source development
community which has driven much of the web innovation.

(That's particularly on the server side; but it's starting to get
corresponding traction on client side systems.  I know
that some corporations don't really want competitors, and it
sure looks like RAND could eliminate some great ones.)

Clearly W3C needs a patent policy, to prevent "submarine"
patents with surprise costs.  But I'd argue that the overall
community would be far better served by having that policy
dictate "zero cost licences".  Vendors can and should use
non-W3C processes to extract revenue from patents; and
W3C should continue to provide the level playing field
which it was founded in order to create.


III - HOW TO DEAL WITH PATENTS

When W3C members declare relevant claims, they should
have at least two options:  granting zero cost licences, so
that the REC would have a safe IP foundation; and
specifying licensing terms.  If the WG continutes, it
would design around any for-profit licences; and
otherwise it would dissolve. (The process might move
outside of W3C into some vendor-oriented context,
though W3C should not grant permission to use any
of its work results in that new context.)

In no circumstance should a member's not declaring
they have a patent give them the ability to "submarine"
licensing taxes of any kind out of a resulting REC.
Such taxes are unfair to other WG participants, as
well as to others who contribute to such a REC.

Giving up options to impose such world-wide taxes should
be costs assumed by members in return for the ability to
control and influence development of W3C standards to
use their technologies.  They would maintain the ability
to use non-W3C mechanisms to extract such taxes.


IV - ADDITIONAL SAFEGUARDS NEEDED

Unfortunately the "good faith" standard is toothless,
and hence useless.  I understand that corporations
will demand such loop-holes, but can't support the
ability of any organization to torpedo standards in
such a manner.  Tax-inducing pattents applying to
standards are significant structural defects.  Stronger
efforts are needed to prevent them from appearing.

There also need to be procedural safeguards against
WGs which might be tempted to create standards that
are incomplete, and which can only be implemented
in usable form with the use of for-payment patent claims.
For example, if it needed six claims from some patent,
but only three were deemed "essential" in some REC,
the community would be severely disserved.

The W3C has unfortunate precedent in producing
REC documents that are incomplete.  I'll use DOM
as one example:  both DOM L1 and L2 can't be
implemented, or even tested, without additional
"back door" functionality which isn't even enumerated
in the specifications.  When such functionality needs
to use patented technology, it looks like the current
patent policy TR can't cover the threat.

- David Brownell

Received on Sunday, 30 September 2001 14:32:56 UTC