On the proposed patent policy of the W3C

I quote from the Overview:

  | our goal is to affirm the Web community's longstanding
  | preference for Recommendations that can be implemented
  | on a royalty-free (RF) basis.

If that is truly the goal, then the proposed policy should
require technology mandated in Recommendations to be
available royalty-free. Until now, that has been the de facto
policy of the W3C. It would be extremely unfortunate if
this were to change; it is preposterous for the document
proposing such a change to claim to be "affirming" the
very policy it liquidates.


Consider some of the implications of non-RF technology
entering the W3C's standards. For the moment, I'm going
to be pessimistic and see what's the worst that could
plausibly happen.

1. Reduction of choice.

There are many web browsers. Internet Explorer has much the
largest market share, but there are plenty more. Almost all
of them are distributed for free. What will happen to this
situation if essential web standards become unimplementable
without royalties? Answer: it will be impossible to create
free software that implements those standards. In view of
the huge importance free software has had in making the WWW
the success it has been, this seems clearly regrettable.

What, then, will remain? Internet Explorer and a couple of
other browsers owned by small companies. What next?

2. Single-vendor lockin.

The obvious next step is for Microsoft to buy out those small
companies. This is clearly to their advantage commercially.
The consequence? It is impossible to get a standards-compliant
web browser from any other source.

What then?

3. The end of the W3C.

At this point the W3C becomes irrelevant. There is no more need
for universal open standards for the web when the only browser
is Internet Explorer. This in turn will speed up the flight from
"openness", of course. (Perhaps it won't be long before it's
impossible to *create* valid web pages without software from
a single vendor.)

Now, there's another possible consequence which would
doubtless greatly please at least one author of the Patent
Policy Framework.

4. Disaster for free operating systems.

So, Microsoft is the only source of compliant web browsers.
Will they provide them to users of Linux or FreeBSD or any
of the other minority operating systems out there? (I'll consider
MacOS in a moment.) I don't see any reason why they should.
They'd have nothing to gain.

Well, now. Who's going to deploy Linux in their company
when doing so makes the WWW inaccessible?

... And what about MacOS, or any other non-free minority
OS out there? Remember that Microsoft have the only
web browser in the world at this point; other companies
making web browsers have been bought out. Apple could
make their own web browser; Microsoft probably wouldn't
try to buy them. So maybe they're OK. It might involve quite
a major investment, though. Does Apple have the spare
resources to do it? Who knows?


OK, end of dystopian fantasy. Let's suppose that doesn't
happen; perhaps it might turn out that the only W3C
Recommendations that aren't implementable RFily are
sufficiently inessential that free browsers continue to
be usable. (Note: free browsers are crucial in avoiding
the nightmare above, because they cannot be bought out.)
What then?

There are still a number of problems.

The W3C would no longer be able to distribute reference
implementations of its standards. (I appreciate that many
W3C Recommendations don't come with reference implementations,
but it would be sad for it to be impossible for legal reasons.)

The proposed requirements on W3C Members have no teeth,
for two reasons.

  - Members can opt out. There is nothing whatever in
    the proposed policy that says that a WG cannot, or
    even should not, approve a Recommendation that is
    covered by patents not licensed on RAND terms.

  - A member can simply claim not to have been aware
    of relevant patents. Consider the following situation:
    EvilCorp Exploitation Inc. gets a representative onto
    the WG discussing the next revision of XHTML, and
    gives that representative strict instructions to push
    for such-and-such a new feature to be mandated.
    This feature is not implementable without using
    techniques patented by EvilCorp, but the representative
    is not told this. (He may think it very likely, of course,
    but he isn't required to make "extraordinary effort" to
    find out whether it's true.)

The definition of "RAND" terms is highly unclear. Specifically,
it permits "reasonable and non-discriminatory" royalties or fees
to be charged. What does that mean? The definition does not
say. For instance, would a level of royalties such as to make it
impossible for most people in Third World countries to purchase
compliant browsers be acceptable? It's anyone's guess.


The one consequence that can be predicted with confidence
here is that free implementations of WWW standards will
become rarer. It seems to me that the success of the WWW
(as of the internet generally) has been largely driven by the
availability of free, standard, implementations. I've considered
above some of the unfortunate consequences if users prefer
"standard" to "free" when the two become disconnected; but
if instead they prefer "free" then the result will be the
marginalization of the W3C and its standards. That would
be a shame, too.


I have one final comment. I first heard about this proposal today;
the very same day as the Last Call period closes. (In fact, in some
time zones it has already closed.) Looking at the archive of
messages sent to www-parentpolicy-comments, it would appear
that hundreds of other people did too. It does not seem that the
proposal was well publicized. I suggest that instead of "the
Patent Policy Working Group invites both public and W3C Member
comment" it should perhaps have said "will grudgingly accept".

Gareth McCaughan

Received on Sunday, 30 September 2001 18:45:32 UTC