[idn] Thoughts on patents

[RJ Atkinson <rja@inet.org>]

[new topic]

	An observation I'll make is that I will vigorously
oppose the adoption of ANY method for IDN that is patented
by anyone -- unless the patent filer/holder formally and
in writing agrees in advance to a no-cost, no-hassle licence 
for anyone implementing the IETF standard.

	Examples of past such agreements with IETF are documented,
for example, in RFC-1822, RFC-1988, and RFC-2339.

	The practice of designing around patent filings is 
quite common in the IETF.  The IETF routinely will re-design 
a protocol to not use a patent if a non-patented approach exists 
-- even if that non-patented approach is not optimal.  The only 
time I can recall where the IETF let a standard rely on a 
patented technology is with regard to the RSA patent on 
asymmetric cryptography (said patent is now expired and void;
that patent had less than 10 years left at the time it was included 
in a standard; said patent was never valid outside the US 
in any event).

	I have no serious doubts that a non-patented approach 
to IDNs can be found, so I suspect that anyone planning to get rich
by licencing a patent on IDNs might be surprised at the end
of the day.  Someone that has patent filings they want the IETF
to consider for IDNs ought to consider making a unilateral
grant of the form in the above 3 example RFCs.  Such a grant
would likely increase the chances that their approach would be 
adopted by the IDN WG and the IETF as a whole.  Obviously
there are no guarantees that making such a grant would necessarily
result in their technology being standardised since it might
not be obviously better than another un-encumbered technology.

	Finally, I will note that if the goal of a patent filing
is to prevent some other party from patenting one's technology,
there are alternative legal strategies that are equally effective.
In the US, one can make a Statutory Disclosure Filing with the US
Patent & Trademark Office placing an invented technology into the 
public domain.  In most non-US jurisdictions, any technology can't 
be patented once it is openly published (this is why the RSA patent
was not valid outside the USA).   So claims that a patent is
needed to protect oneself against another firm patenting one's
own invention are at best misleading, IMHO.

	The above is my personal view, not necessarily that of 
my employer nor of any organisation that I might be associated with.

	Let me discourage folks from a wide-ranging philosophical
debate about patents or nit-picking with my outline above.  We
should focus on finding a good (IMHO, non-patented) approach to
solving the IDN issues.

Yours,

Ran
rja@inet.org