Nokia is one of the companies which submitted a number of withdrawal
requests for previous disclosures. In no case (that I'm aware of) our
intention has been to sneak out of a licensing commitment. Instead, we
submitted withdrawal requests with the intention to keep the IETF patent
database a useful tool---to do our share of database cleanup, so to speak.
For example, we removed disclosures where
-the patent went away (e.g. an abandoned application with no intention to
re-file the case)
-the scope of protection changed in such a way that the previous disclosure
became irrelevant, or
-an I-D went away and, in our estimation, the protected technology has not
been picked up in any other IETF document we are aware of. (If it were, we
would submit another disclosure for the same patent, but against a different
draft. This has happened once in case of Nokia).
We believe that these actions have been of advantage to the transparency of
the IETF patent system, and transparency is important. When writing
"transparency", I mean transparency to the technical IETF contributor, who
typically has neither interest, nor the qualification, to accurately
interpret the legalese of patent disclosures. (All too often guys just
state "there's a patent on this draft", because they found something in the
tracker---and in some WG, in practice, that can kill a draft.)
We also think that in an organization like the IETF, where language and
practice suggests the disclosure of (unstable) patent applications against
(unstable) I-Ds, there is a need for a cleanup mechanism of some sort. This
is in contrast to organizations where one needs to declare only once at
least one of the documents is reasonably stable.
I personally believe that the impact of a removal of a disclosure to a
licensing promise is rather negligible. The paper-trail of a disclosure can
quite easily be reconstructed during litigation, if a need arises. The
IETF's patent database should focus on the practicalities required for IETF
My suggestion would be to either continue the current practice, or implement
something along the following lines:
-an "invisible" flag, under control of the discloser
-an "expert" mode in the database, which provides the whole paper-trail,
-a "standard" mode which lists only the most recent update of a disclosure
(or the information that the request has been flagged "invisible" by the
On 8/14/08 12:25 AM, "Simon Josefsson" <simon(_at_)josefsson(_dot_)org> wrote:
Harald Tveit Alvestrand <harald(_at_)alvestrand(_dot_)no> writes:
Simon Josefsson skrev:
Brian E Carpenter <brian(_dot_)e(_dot_)carpenter(_at_)gmail(_dot_)com>
I wasn't even aware, during my tenure as chair, that the 'remove' button
existed. The only removals I recall, which may or may not be in the
numbers Simon quoted, were completely bogus and nonsensical disclosures
clearly filed by someone who was just fiddling around on the Web.
Some of the disclosures that are now removed were certainly not bogus.
For example, the patent license given in #833 was important input to a
lengthy discussion relatively recently.
definitely agree on that one "for the record".
OTOH, to give a counterexample, I don't think there's any value to the
community to having both #941 and #942 on file - they're duplicates.
Removing one out of two duplicates doesn't remove any patent-disclosure
related information, so I don't think it is a good counter-example.
If removals should be permitted, the reasons for accepting a removal
request should be well established. I can think of at least two reasons
that are valid:
* Exact duplicates
Beyond this I'm less sure we can get away the liability concern.
False positives for spam could be a issue, so I'm not even sure the
second one is OK.
Ietf mailing list
Ietf mailing list