[linux-elitists] Re: [IRR] No-fee license vs. RAND (was: Greg Aharonian...)

Eugen Leitl eugen@leitl.org
Thu Dec 16 03:31:21 PST 2004


----- Forwarded message from "David L. Black" <david.l.black@verizon.net> -----

From: "David L. Black" <david.l.black@verizon.net>
Date: Wed, 15 Dec 2004 22:42:24 -0500
To: <somelist>
Subject: Re: [IRR] No-fee license vs. RAND (was: Greg Aharonian...)

>Only if you all remember that my name is "Alan" not "Alex."  Unless you're
>talking about some other Alex...

My thinko, mea culpa ... please don't blame Keith ... sorry.

>Anyway, here's the basic pro-free position:

[ ..  definitions snipped ...]

>In specific, when the W3C was debating which policy to adopt I worked for
>a small company that made a server.  One of the competitive advantages of
>our product was that it supported the full suite of Web services protocols
>(SOAP, WSDL, UDDI, etc).  In the discussion it became clear there were as
>many as 20 patents that might be relevant to each of these standards.
>Now, let's assume they're all licensed under RAND terms.  Let's further
>say that the RAND price is an insanely low $100.(*).
>
> (*) I say insane because customary license pricing is at least an order of
> magnitude more than that

That's not true.  In the media space, many technologies have patent
pools that charge a *lot* less.  Here's an MPEG example where the
maximum per-copy fee for codecs is $0.20, and small-scale
distributors (less than 100,000 copies/year) pay nothing:

http://www.mpegla.com/avc/AVC_TermsSummary.pdf

>A little basic math shows that we would have had to pay $600 per copy of
>our software just in RAND licensing fees, and that's not including any of
>the administrative overhead costs.  $600 was more than twice the
>single-copy price of the product.  So we would have had to at least
>triple the cost of our product if W3C had gone with RAND.

I can't argue with Alan here, because I agree that a multi-hundred-
dollar per-copy patent license obligation makes open source software
impossible.  Unfortunately, the W3C royalty free patent policy is
incapable of completely solving this problem because the policy only
applies to those who choose to be bound by it - to take one extreme,
no "patent farmer" would ever join the W3C.

>Now, from a political point of view, I also believe that it's not the job
>of big standards-setting bodies like W3C to help companies enrich
>themselves off intellectual property.

The flip side of that position is that it's also not the job of such
bodies to confiscate patents.  The way that the W3C patent policy works
is that if someone participates in a standards effort, and holds a
patent necessary (the precise word is "essential") to the standard,
then anyone who implements the standard is entitled to a free license
to the patent.  That effectively confiscates the patent, as showing
that one has a contractual right to a license is usually a winning
move in defending patent litigation.  There are exclusion procedures,
to exempt a patent from the licensing obligation, but they're backwards -
they're opt-out instead of opt-in, and cause problems analogous to
opt-out mailing lists in that a patent holder who isn't paying careful
attention to all the procedural details can easily get shafted.

An unfortunate result is that if <X> is involved in a W3C standards
effort, and the group does something that requires one of <X>'s patents,
and the W3C participant doesn't know about the patent, slurp ... it's
gone once the appropriate time period runs out, and the participant's
lack of awareness of the patent doesn't make a difference.  In contrast,
IETF policies only apply to patents that the participant knows about,
or should know about (to disallow intentionally withholding patent info
from someone who would otherwise know it).

>I think that a good standards
>process requires up-front and open disclosure of possibly conflicting
>intellectual property.

Here I strongly agree, and this is the bedrock principle of IETF's
intellectual property rights policy, although the "up-front"
aspect could use some more teeth in practice.  This seems like a
good juncture to lay out my views ...

-- What is the problem?

I'm going to start by disagreeing with Alan about the problem.
The issue in the IETF was not patent licensing terms, per se, but
rather what rules the IETF as a standards organization should
put in place for intellectual property and what disclosure and
licensing obligations should be imposed on IETF participants
across the board.  As noted above, the IETF intellectual property
rights policy is fundamentally disclosure-based, so the main issue
is/was:

What licensing obligations should be required in a disclosure?

This is somewhat different from RAND vs. royalty-free terms,
because standards benefit from broad participation of those who
could/would use the standard, and hence the effect of proposed
rules and obligations on participation need to be considered -
if the rules and obligations are too draconian, they will drive
people away to other forums.  It's also the case that the more
that is required in a disclosure, the harder they are to get,
the longer they take, and the bigger the disincentive they
pose to participation.

Lest you believe this is hypothetical, W3C's patent policy
would be a real obstacle if I wanted to participate
in W3C.  I should also point out that I chair or co-chair two
IETF WGs, and hence get to deal with patent issues whether I
like it or not ...

-- Why RAND disclosures?

It's very important to understand who the target of patent policies
are.  Patent farmers, patent shells (company collapses into a
shell containing patents and lawyers), and the like won't join a
standards organization and aren't affected.  Disclosure obligations
instead fall on participants who have some interest in seeing
the standard completed, and some incentive to cooperate.  The
requirement for RAND disclosures sets a minimum bar; an IETF
WG is free to ask for more, evaluate other aspects of the situation,
etc.  Different WGs will reach different conclusions in different
circumstances.

The opposition to RAND appears to assume the worst - that a
RAND disclosure will always result in per-copy license fees -
and hence is wont to declare RAND disclosures incompatible with
open source software.  Needless to say, there are plenty of
counterexamples to that conclusion because RAND disclosures
don't always lead to per-copy license fees, and in particular
a RAND disclosure allows royalty-free licensing if the patent
holder so chooses.  There are other options, such as fixed fees
or fees based on volume or revenue.

Keep in mind that disclosures are always prepared by lawyers,
who prefer to say as little as possible to keep as many options
open as possible.  In requiring RAND disclosures as a minimum,
the IETF is attempting to strike a balance between patent concerns
on one side versus breadth of participation and promptness of
disclosure on the other.  An IETF WG is always free to go back
to ask for more, which is a good point to pick up on Alan's
email, because he used a rather bad example of asking for more ...

-- The Sender-ID IP snafu

>I won't rehash what happened to P3P, but I will
>say "tolja so" about Sender-ID, which the IETF recently had to throw out
>in part due to IP snafus.

That's a peculiar "tolja so", because the most visible IP snafu
involved a patent for which a royalty-free patent license was
offered.  The issue was that the offered license did not permit
sublicensing, but the offered license was close enough to the
requirements of the W3C royalty free patent policy that it probably
could have passed muster with a little work - needless to say, the
W3C royalty free patent policy doesn't require sublicensing.

Had the IETF adopted the W3C royalty-free patent policy, at least this
IP snafu would still have happened, because the open source community
through its counsel chose to pick a fight over sub-licensing.  IMHO,
the fact that the patent was held by Microsoft (viewed as an enemy by
significant portions of the open source community) had a great deal
to do with this - I find it hard to believe that the same IP snafu
would have happened if IBM had been the patent holder.

-- Summary

RAND disclosure requirements are better than royalty-free because:
- They encourage broader participation in standards efforts
- They increase flexibility for both patent holders and WGs,
	recognizing that not all situations are the same.
- They allow a WG that believes royalty-free requirements are
	necessary to its work to demand them, or work around
	patents based on a patent holder's lack of cooperation.

Also, standards group intellectual property policies are incapable
of solving the complete patent problem because obligations cannot
be imposed on patent holders who are outside the standards process.
The royalty-free requirements approach has the peculiar side effect
of making things worse for precisely those patent holders who are
trying to cooperate with the standards efforts.

-- Final Thought

IMHO, a significant part of what is going on here is that there is
a strong belief in a portion of the open source community that
software patents are evil and should be fought in every way possible.
People are entitled to that opinion, but also need to realize that
software patents are a fact of life as the US patent system currently
stands.  I have no problem with opposition to software patents and
actions directed to governments to change the patent system, but I
have severe problems with those who fight software patents by proxy
through standards organization patent policies in a way that
pays no heed to the potential collateral damage to the standards
organization.

Thanks,
--David
_______________________________________________
...
----- End forwarded message -----
-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a>
______________________________________________________________
ICBM: 48.07078, 11.61144            http://www.leitl.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE
http://moleculardevices.org         http://nanomachines.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
Url : http://allium.zgp.org/pipermail/linux-elitists/attachments/20041216/972d12b4/attachment.pgp 


More information about the linux-elitists mailing list