Six issues

To: The World Wide Web Consortium and interested parties
Re: Proposed RAND policy change
Date: 8.October 2001 [8:45 Central time]

Summary:

This posting lists six major objections posted in response to the proposed
RAND policy change. Each one is explained briefly. (NOTE: this posting is
based on an article available on-line at
http://www.lisa.org/2001/rand/rand.html)

****************

Dear W3C,

I am writing this message in the hopes that I may have some positive impact
on the current debate over RAND licensing. In particular this response
concerns the potential impact that the creation of standards subject to RAND
licensing fees might have. It does not address the disclosure or revised RF
parts of the proposal, which are not the primary issues at stake.

I have read many of the postings on this subject here and found that many of
them show little or no understanding of the issues involved. Between threats
and irrelevant comments that show a lack of comprehension of the issues, it
is somewhat difficult to distill what the relevant criticisms of the policy
really are. This clearly makes the W3C's job harder. Nonetheless, after
going through many of the posts it seems that opposition to the RAND
provisions in the proposed policy is focused on six items:


1. For software development with no revenue model no licensing fees could be
reasonable. For free software RAND is a contradiction in terms. RAND would
force free software developers to either leave features out of their
software or even cripple existing features in it.

2. The RAND policy violates the W3C's mandate by letting the W3C create
standards that could stand as barriers to universal access to the Internet.

3. Only a few countries currently recognize software patents, meaning that
the legal framework behind the RAND proposal is not codified in national law
throughout most of the world. The RAND proposal would still require the
payment of licensing fees, even though the software might be sold or
distributed where the patents hold no validity.

4. Some smaller software developers feel that RAND would put them at the
mercy of larger corporations and not free to fully implement standards.

5. RAND would create a climate of fear or uncertainty in which developers
would not know if or when a company might choose to exercise patent rights
on a standard, making planning difficult and perhaps forcing developers to
dispense with features entirely.

6. If RAND goes into effect, companies might choose to not follow
patent-encumbered W3C standards, or might choose to create alternate de
facto standards outside the W3C. (This is commonly being referred to as a
"fork".)



Each of these criticisms deserves a brief, but more detailed, examination:

1. RAND discriminates against free software

The RAND proposal does seem to assume a revenue-producing business model
behind software development. In the absence of a revenue model there is no
apparent way in which free software developers could pay licensing fees.
Brooke also points to cases in which free software has had to abandon
central features (such as GIF image or LZW image compression support)
because of licensing fees being imposed on standards after such software had
used them for some time.

The W3C's original proposal does not mention open-source or free software
anywhere in the document, and this is an area that appears not to have been
considered in the Working Group to any real extent (at least as far as can
be determined in the proposal).

The W3C's public response to criticism has not addressed this point either.
This issue is a major one, involving not just the W3C, but questions of
patent rights in general and how non-revenue models can co-exist with
businesses that must be concerned about revenues.

This issue must be publicly debated before the proposed RAND policy takes
effect.



2. RAND erects barriers to Internet access

The W3C states that one of its three goals is "Universal Access: To make the
Web accessible to all by promoting technologies that take into account the
vast differences in culture, education, ability, material resources, and
physical limitations of users on all continents"

Opponents claim that charging for use of W3C standards violates this goal
(particularly the part about "material resources") by putting financial
barriers in place to those seeking to implement standards.

If all software used for Internet access were commercial (or at least had
some revenue basis to pay patent licensing fees) this issue would not
particularly matter, but a significant amount of the software used on the
Internet is either free or open source (the Apache web server being a prime
example). Given the vast amount of such software used for Internet-related
activities, particularly by lower-budget operations, RAND fees could end up
providing a major block to open and free access to the Internet.



3. Only a few countries recognize software patents

The W3C's response to this criticism has been that they have had
participation in the working group from people in all regions of the world
to this point, and they have posted descriptions of software patent law in
those countries that have such law. This does not really answer the
question, however, of how RAND licensing fees could be required in countries
that do not recognize software patents. For software developers outside the
U.S., Australia, China, Japan, Korea, Malaysia, New Zealand, Singapore, or
Thailand, this rule could cause problems since they would be subject to
patent laws of these countries if there is the possibility that any of their
software could be distributed or imported in these countries. This would, in
effect, give national law in these few countries an international reach.

Given the patchwork of laws and the varying legal attitudes regarding
software patents at the very least the W3C owes an explanation of what it
sees the legal basis for the RAND policy to be. The W3C also needs to
explain how licensing might work in international situations.



4. RAND puts smaller developers at the mercy of larger companies

RAND stipulates that the charges be reasonable, but when potential licensors
range from tiny start-ups run from someone's garage to large companies, how
is reasonability to be determined? (The free software issue really
represents the extreme end of this question.)

The RAND agreement would allow per-use or per-copy charges as licensing
requirements, or could require a flat fee for use. However, if a flat fee
were charged it would have to remain the same for every entity seeking to
license the patent rights, so companies would be unlikely to actually make
this an option, or would be forced to make it very expensive (but reasonable
for large companies). The use of per-copy or per-use licensing would force
the implementation of accounting systems to track these numbers in areas
where such tracking has traditionally not been done.

This issue needs to be addressed before any RAND policy takes effect because
of the potential costs and difficulties involved in adapting the policy to
smaller companies.



5. RAND would create a climate of fear

Opponents of RAND claim that RAND, by allowing costs to be attached to
standards, would make it difficult and uncertain for companies to plan
standards implementation since they would not know if or when fees would be
levied, and they would be open to allegations of violation of RAND
agreements. The so-called "back-door RAND" problem could allow companies to
retroactively impose fees on further use of standards that were initially
released as royalty free.

In addition opponents claim that even though the W3C stipulates that
adherence to agreements be made on a "good faith" basis and that members not
be required to do extensive searches of patent portfolios, it would be
difficult to prove that any lack of compliance was unintentional, so
companies would be forced to expend considerable resources to ensure
compliance. This would give larger patent-holding companies considerable
leverage in their dealings with smaller companies, since the threat of
examining whether smaller companies were complying with RAND licensing
arrangements would pose the specter of large legal fees that smaller
companies might not be able to bear, keeping them out of certain arenas
entirely.

The W3C, on the other hand, states that RAND is needed precisely to
eliminate a climate of fear and uncertainty. They hold that the present
situation allows patent holders to hold the standards process and
implementation to standards hostage to patents not disclosed early on in the
standards-creation process. By making RAND a rule for all companies in the
W3C, the member companies would be forced, as a condition of participation
in W3C, to make any patents they hold that are used in standards available
at a reasonable cost, thus eliminating the need to worry about intellectual
property rights and the possibility of one company putting the entire
process in jeopardy.

It seems that both sides of the argument have valid points on this issue,
and it is hard to see how the two sides can be easily reconciled on this
issue.

Even if the "back door RAND" issue is just a chimera, then the basis for
prevention of abuse needs to be explicitly stated before any RAND policy
goes into effect. While there may be a legal framework to prevent abuse of
the policy the present documents do not make this clear. This issue must be
dealt with.


6. Developers might not use patent-encumbered standards

Various individuals in their response to the proposed rule have openly
advocated ignoring any patent-encumbered W3C standards, and some have gone
so far as to suggest the creation of an alternate standards body dedicated
to the RF model, should W3C implement the RAND policy. Other, more moderate,
writers have acknowledged that starting an alternative to W3C is inherently
unlikely, but have worried about the potential that developers would simply
ignore RAND standards and "go it alone", leading to the creation of
alternate and competing implementation of features on the Internet (like the
ongoing feature war between Microsoft and Netscape browsers in which each
side adds incompatible features and then lobbies to get its own
implementation made standard).

At present there seems to be no way to ascertain whether or not this would
actually be a problem inasmuch as there are no RAND standards yet and no one
has experience with the actual licensing fees and arrangements. The concern
expressed by opponents to RAND seems to be substantial, and needs to be
addressed by the W3C before any decision is made.



Given the serious nature of these objections and the fact that the present
policy document does not show that the Working Group has considered these
issues, the W3C needs to explicitly address them before any policy goes into
effect. (It is possible that these issues were considered in the drafting of
the policy recommendation, but they are not apparent at all and a lot of the
fury over the subject derives from this fact.)


I hope that the W3C will give serious consideration to these issues before
any actions are taken. Please do not tar all the serious issues with the
same brush reserved for postings by people with no understanding of the
issues.

Regards,

Arle Lommel

LISA (Localisation Industry Standards Association)
-and-
BYU Translation Research Group

Received on Monday, 8 October 2001 10:26:54 UTC