W3C Royalty-Free Patent Policy Last Call Working Draft Comment

Dear PPWG,

Thank you for publishing the 14 November 2002 Last Call Working Draft. I
found the Addendum that sets out the rejected RAND exception process
enlightening and am very pleased it did not come to pass.

Since the Free Software Foundation (FSF) published their position on the
Last Call Working Draft on 25 November 2002 most responses in this forum
have strongly rejected the inclusion of limitation 3 in section 3, which
states: ``3. may be limited to implementations of the Recommendation,
and to what is required by the Recommendation;''

The FSF informs us that such a restriction will be incompatible with the
GNU GPL and render any GPL software that implements such a licence
limitation illegal to distribute.

Licence grants that are limited to a published specification appear
quite common. For example I'm thinking of Adobe's Patent Clarification
Notice: http://partners.adobe.com/asn/developer/legalnotices.html

``Accordingly, the following patents are licensed on a royalty-free,
non-exclusive basis for the term of each patent and for the sole purpose
of developing software that produces, consumes, and interprets PDF files
that are compliant with the Specification:''

We are even told why Adobe has done this: ``Adobe desires to promote the
use of PDF for information interchange among diverse products and
applications.''

This is precisely the type of goal that organisations making
contributions to the W3C recommendation process have in mind. Yet here
we see a company freely licensing their patents to implement a
specification because market share and being able to leverage network
externalities is critical to the future success of their product.

The W3C may not even need to apply any pressure to patent holders to get
them to agree to the Royalty Free Licensing Requirements. Patent holders
may ``reluctantly'' agree because the windfall of W3C approval will
hopefully lead to mass adoption of their ideas. Eventually extensions to
the specification will become widely used, because anything less is
nothing short of technological stagnation.

Once such a point is reached the patent holders may be in a powerful
position to demand whatever unreasonable and discriminatory licensing
requirements they can dream up. I strongly suggest you clarify that the
implementation of any extension (third party or otherwise) to a W3C
recommendation is still covered by the licence grant.

If it is not then what you may have done is deleted the Core/Extension
process and replaced it with an informal one that could force future
non-proprietary technology stagnation. We may be worse off than if you
had kept out of the whole process. So long as companies have an
incentive to get developers and users to adopt their technologies they
will continue to consider granting royalty-free licenses to implement
their specifications. If the W3C assists them in obtaining widespread
adoption then the incentive to provide future royalty-free licenses is
diminished. As extensions inevitably develop that also use the same
patented algorithms then future implementations may have to become
proprietary.

This seems to be a tricky dynamic problem that I will give further
thought to. The W3C should ask itself whether it has a role in
cultivating a climate conducive to future free standards. Allowing the
same patent problems to reemerge every time one develops an extension to
a recommendation doesn't appear helpful.

We revisit the point that whatever influence the W3C has should be used
to encourage standards implementation and adoption by all its
constituency. This includes proprietary and non-proprietary software
developers. If companies promoting a technology want their ideas to be
adopted by both groups (any W3C recommendation has this degree of
universality) then both groups must be able to compete as extensions to
a recommendation naturally develop over time. Anything less is a
disservice to its constituency.

Regards,
Adam Warner

Received on Tuesday, 7 January 2003 01:52:39 UTC