Re: The Burden of Proof

On Tue, 2001-10-02 at 18:15, Daniel J. Weitzner wrote:
> Hello Adam,
> 
> You've take a lot of time to provide you're views on our patent policy
> proposal, so I hope you can spare a bit more to reply to my questions.

I appreciate the opportunity.

> ----- Original Message -----
> From: "Adam Warner" <lists@consulting.net.nz>
> To: <www-patentpolicy-comment@w3.org>
> Sent: Tuesday, October 02, 2001 12:49 AM
> Subject: The Burden of Proof
> 
> > Hi all,
> >
> > Having read through:
> >
> > http://www.w3.org/2001/10/patent-response
> >
> > "The draft policy does attempt to answer this question: In a world where
> > patents exist and may be used to constrain conformance to standards, how
> > should W3C best proceed in order to accomplish its mission? Even if a
> > patent holder claims that a patent is relevant to a W3C Recommendation
> > and that party offers a license, this does not mean that W3C (or anyone
> > else) shares the belief that the claim is valid, or that an implementer
> > has infringed upon it."
> >
> > Thanks for asking and answering the wrong question.
> 
> Perhaps there were other questions we should have answered as well, but W3C
> recent experience with patents established that we needed to answer this
> one.

Yes, of course :-)

> > In this respect what I originally wrote still stands:
> >
> > "What isn't clear is that the appropriate response is for the W3C to
> > condone RAND licensing terms and to actively promote non-free licenses."
> >
> > What I meant by that is obvious: If the W3C agrees to a Recommendation
> > that includes RAND fees/restrictions and the official W3C logo/body is
> > used to promote that Recommendation it can't help but promote those
> > non-free licensing terms as well (for incorporation into our future
> > WWW).
> >
> > What we need from the W3C are some well reasoned examples for why RAND
> > licensing terms are necessary for the future functioning of the W3C and
> > for the BENEFIT OF THE WWW (that's the overriding criterion of everyone
> > here after all). So we can all nod our heads and go "Yeah, that makes
> > sense."
> 
> I believe that we've already provided two sets of answers to this question
> in the policy.

Daniel as the Director of the W3C's Technology and Society division I
understand you are more than qualified to address issues of Internet
public policy. Please point out any areas of my ignorance so we can
overcome those issues quickly.

As you are also a teacher of Internet public policy I hope you will be
able to provide some hypothetical situations that we can discuss--where
RAND licensing is necessary--in more than abstract terms.

> First is a set of general observations about how the patent
> environemnt around Web standards has changed including a number of
> situations in which W3C working groups have already run into patents not
> available on Royalty-free terms

OK. This is concrete. We have hit a philosophical divide. Those who are
looking to maintain free Internet standards will want the W3C to stop at
this point and make one of, say, three decisions: work around the
patent; lobby hard to get the respective company to provide the
technology on RF terms (especially if the patent is held by a member
organisation); or disband for the time being and if relevant (i.e. it is
widely concluded the patent is specious) call upon the Internet
community to challenge the patent.

> and convergence between the Web and the
> telecom & consumer electronic industries. As you know, those industries and
> their standards bodies work in a patent-intensive environment. (see section
> 2.1 Larger Role of Patents on the Web Landscape [1] for more.)

(Following URI [1]
http://www.w3.org/TR/patent-policy/#sec-patent-role-web)

"1.Convergence"

Why does convergence have to go the way of RAND licensing? OK it's
pretty much guaranteed it will happen if the W3C adopts RAND licensing
as a viable, sanctioned option.

But what if the W3C is the standards body that upholds the RF rights of
its members and the wider community? As these other technologies
converge towards free Internet standards perhaps they will have to adapt
to a free Internet rather than the other way around.

And if the W3C is not willing to draw this line in the sand, what will
stop another organisation with an invigorated moral mandate drawing that
line in the sand themselves and others moving to support them?

The free Internet is too wonderful an ideal to give up so easily. The
only reason we are having a conversation now is because of free Internet
standards enabling me to be an independent publisher at extremely low
cost. As media industries converge towards the WWW they will desire to
increase the cost of independent publishing. W3C sanctioned RAND
licenses will help to achieve that objective.


"2.Rise in patent issuance."

Which can be dealt with by the grant of RF licenses by W3C members.

Is the W3C concerned that they will lose their member base if they do
not allow RAND licensing? (NB: Is there somewhere I can view individual
member contributions to the W3C?)


"3.Experience of Internet-related standards bodies: A number of
standards bodies including W3C, IETF, the WAP Forum, and others, have
encountered potential barriers to acceptance of standards because of
licensing requirements perceived as onerous."

That's how many people view reasonable and non-discriminatory fees and
audit requirements, especially those who are accustomed to the free
standards and software of the Internet.


"4.Popularity of business method patents: Beginning with the State
Street decision in the United States and continuing through high-profile
litigation between Amazon.com and Barnesandnoble.com, business method
patents have become increasingly significant factor in the ecommerce
marketplace."

I'm a New Zealander. How much sympathy do you think I have for United
States business method patents? But regardless, what has this to do with
RAND licensing? If Amazon.com (to quote your example) wants to work with
the W3C to promote a Recommendation then I'd sure hope they'd generously
provide an RF license to (say) their single-click patent. And I'd also
hope the W3C would never dream of incorporating it on more than RF
terms.


The W3C's arguments are weak. They are not sufficient to displace the
history of free licensing that has made the Internet what it is today.

We also see no analysis of how RAND licencing may affect different
sectors of the Internet community, especially those in the free software
community.

What's strange about the Patent Policy FAQ? It contains three sections
on implications: "Member Implications," "Working Group Implications" and
"Legal Implications." Here's a guess for why this policy change was off
people's radar (apart from the terrorism of last month): You completely
forgot to study "Internet Community Implications." I bet, for example,
that if the FAQ had mentioned this could affect the free software
community someone would have spotted it before me (please point out
where it does if I've missed it).

Secondly, where is the rigorous economic analysis of the benefits to the
community you serve of RAND-style licensing?

And by the way that economic analysis should include investigation of
how the dynamics of standard setting will change within the W3C because
of some members' preference for RAND-style licensing. I'm not the one
who wrote this:

'From a licensor perspective, everyone would prefer the comfortable
choice of a "RAND commitment" (which need not identify particular
licensing fees or terms).'

It's your PP FAQ:
http://www.w3.org/2001/08/16-PP-FAQ#[4-4]

That economic analysis should be made public to allow independent
comment.


> This first set if answers points more to trends than concrete situations, so
> the policy offers the second set of answers by creating a process which
> requires that just the question pose be answered when we create an activity.
> Section 5.1 of the policy requires that:
> 
> "Activity proposals and/or draft charters proposing licensing modes must
> state clear reasons for selection of the licensing mode proposed. Members
> commenting on these proposals during the Advisory Committee review process
> should state reasons for their views on the appropriate licensing mode given
> the specifics of the Activity, relevant market, along with any other
> factors." [2]
 
> With this, I believe we have assumed the burden of answering the question of
> the wisdom of RAND or RF for each activity that we start.

"Trust us, we know what we're doing." The answers above are more
concrete than this.

What "relevant market[s]" are we talking about? If the W3C envisages
these markets may not provide for free WWW participation you better let
us know now.


> The question will
> be asked an answered again by the entire development community when they go
> to implement a specification created in either a RAND or RF group (I expect
> that many will refuse to implement specs they believe to be chartered in the
> wrong mode) and each time a Working Group published a public working draft
> (every 3 months) for public comment. W3C certainly may make the wrong
> decision in some cases, but I'm confident that we'll heard about it and hope
> that we'll learn lessons about how to make these choices.

Well a power struggle is going on here because I'd rather the W3C not
have the ability to make the "wrong" (RAND track) choice. But that's
because of my preference for free standards.

Perhaps you could point out the constituency that wants RAND licensing.
Let them put their hands up and be counted (a few of them already have,
in this comment archive and by their unwillingness to provide the W3C
with licenses on RF-terms).

The W3C is trying to serve two masters. I can understand how difficult
that is.


I have just spent some time trying to come up with some really elaborate
examples of how a higher-level service can become a lower-level one over
time (since this would refute your theoretical underpinning for RAND:
"Recommendations addressing higher-level services may be appropriate for
licensing on reasonable and non-discriminatory (RAND) terms.")

But then I see Ron Arts has just given a beautifully succinct example:

http://lists.w3.org/Archives/Public/www-patentpolicy-comment/2001Oct/0217.html

"Another problem is the low/high level separation. If you look at
the history you will notice that what currently is high-level,
can in a few years be seen as low-level (for example: HTTP used to
be high level but is now included in the linux kernel). Accepting
a high-level standard under RAND may sound reasonable, but will
result in a few years in a low-level standard under RAND."

We could imagine that media services might become a part of kernel
operations in the future as well. What is your comment on this
phenomenon?

I can't remember the author/post but someone also brought up the very
important point that high level services may also need to be processed
at a lower level (for example by converting it into another format). So
even if encoding ability was free and decoding ability was intended for,
say, consumer electronics devices (where the W3C also believes
convergence in this area may justify RAND licensing) it would impact
upon the server infrastructure as well.

I look forward to reading your comments on this matter and I also invite
others to respond to our respective comments.

Regards,
Adam

Received on Tuesday, 2 October 2001 07:40:00 UTC