[cryptography] NSA's position in the dominance stakes

James A. Donald jamesd at echeque.com
Fri Nov 19 22:42:01 EST 2010

James A. Donald:
 >> I know they are patent free, and you know they are patent
 >> free, but Certicom does not admit they are patent free and
 >> it is not likely a judge or jury would be able to tell the
 >> difference.
 >> The history of patents is that is harder to enforce a
 >> valid patent in a application where it actually applies,
 >> than to enforce a totally and outrageously bogus patent,
 >> because the judge and the jury is likely to find a valid
 >> patent's application to a valid case entirely
 >> incomprehensible.
 >> Let us suppose you use Tate pairing key agreement, and
 >> Certicom's lawyer claims you are using MQV

On 2010-11-20 9:35 AM, Jon Callas wrote:
 >   Forgive me, but that is insulting to both judges and
 >   juries. In that particular case, it is easy to defend
 >   because the question is "are you using MQV" and the
 >   answer is no.

But the defendant is always going say he is not using MVQ,
and the plaintif is always going to say the defendent *is*
using MVQ.    How are the judge and jury going to figure out
which one is telling the truth?  The judge and the jury do
not know MVQ from a dog that bit them, reading the
plaintiff's patent will not enlighten them in the slightest,
and they cannot read the defendant's code.

In order to understand the plaintiff's patent or the
defendant's code, you have to know programming, and a bit of
public key cryptography.  Few judges know programming, none
know cryptography, and any juror who knew programming would
be thrown off the jury for being dangerously smart.

Thus there is simply no way that these patents can be
enforced honestly and competently.  Indeed, the same is true
for every patent in every field, for analogous reasons.

 > > Conversely, suppose you really are using MQV - but you
 > > have changed the variable names, and are using the Boost
 > > metaprogramming template library to create and utilize
 > > cryptographic types and operations.  Now imagine
 > > Certicom's lawyer trying to explain boost metaprogramming
 > > to a judge and jury!  Obviously he would not try.  He
 > > would instead claim that Certicom owns ECC no matter how
 > > implemented, and slip the judge a suitcase full of small
 > > unmarked bills.  What else could he do?

On 2010-11-20 9:35 AM, Jon Callas wrote:
 > So you're saying that patent suits are won by bribing the
 > judge?

If you examine most patent cases, it is usually perfectly
obvious that the fix was in before the case started.  As to
whether this was done by bribing the judge, bribing the
politician who controls the judge's career, pulling strings on
the old boy network, threatening the judge, or whatever, I
have no way of knowing.

I do know that threatening the judge with physical harm is
becoming a more popular and increasingly common move in
corporate lawsuits, though this move was unknown and
unthinkable not very long ago.  As to whether this is true in
patent cases, again I have no way of knowing.

More information about the cryptography mailing list