[cryptography] NSA's position in the dominance stakes

Jon Callas jon at callas.org
Thu Nov 18 23:39:16 EST 2010


> I notice that not only has the concept of ECC been patented, but so has data compression, and graphical hyperlinks.  Why should we worry about one patent more than the other?
> 
> Does the fact that graphical hyperlinks are patented cause you to refrain from using them?

I want to take a slight issue with the above.

The concept of ECC is in fact *not* patented. I know that Marsh posted a thing from a Certicom person claiming this, but it is not true. Certicom's patents are on *optimizations* of ECC, as well as a few other things like MQV. There are Certicom-free ECC libraries. Certicom has never been clear as to what all they claim, which is of course to their advantage. It's also frustrating as hell, for all the obvious reasons.

Strictly speaking, data compression per se and links per se aren't patented. The full discussion is long and complicated and really a digression. I just feel like I should point that out. Sure, patent holders have said those things, especially when not under oath. But as Gershwin said, it ain't necessarily so.

You should *never* believe what a patent holder says their patent covers. They can and will put the best possible lipstick on their pig, along with a feather boa, and a tiara. You have to look past the lipstick and rhinestones to see what the actual pig is.

I was involved in a patent dustup that went sorta like this:

Them: "Take a look at this patent. Lots of people have licensed it, maybe you want to, too." (Any smart person starts this way.)

Us: "Hmmm. Nope, no, I don't think we do."

Them: "Are you sure? We'd like to have a meeting on this. We think you'll find licensing them cheap, and effective. Lots of other people have." (They named the companies who had licensed, so as to make a good impression.)

Us: "Sure, we'll take the meeting. But before you come, be sure to look at this patent and this one too. We think you might be interested in licensing it from us. One possible outcome is we like each others' patents." (This is suggesting a cross-license, which is no good for them, as they get no money and have to sign a contract.)

When they came to that meeting, I can assure you that our pig was wearing lipstick, mascara, a sexy one-piece bathing suit, and stylish-but-not-slutty pumps. We came out of that meeting agreeing to have a follow-up meeting, that has not happened yet, and that was years ago.

The reason it went that way was that it was two companies that actually make things. Companies whose assets are largely IP tend to be aggressive with their patents, as that's how they make money. Also, they are less likely to have to end up cross-licensing. Companies that make things do a much more subtle and refined dance.

	Jon




More information about the cryptography mailing list