LWN: Comments on "An introduction to Elliptic Curve Cryptography"

Elliptic Curve Cryptography patents

giraffedata — 2006-03-11T23:09:57+00:00

I too have only a vague understanding of this, but years ago, I know the US stood out with its "first to invent" rule, which meant a patent was good only to the first person to invent something, so if you had private notes proving you had invented it earlier than the patentee, you not only had a right to use it yourself, but the patent was invalid and you could file your own patent and exclude the original patentee from the invention!

But I also heard the US was thinking of changing to the more practical "first to file" rule, where whoever files a patent first owns the invention. I don't know if that happened.

Incidentally, one point missed from the original posting is that it contemplates developing an original implementation "from the math." In that case, there's no independent invention going on anyway -- using that math for that purpose is the invention, and you got it from someone else. It's not only a patent infringement, but a deliberate one.

An introduction to Elliptic Curve Cryptography

LetterRip — 2006-03-11T08:33:20+00:00

[QUOTE]You're infringing even if you had made the discovery earlier, unless:
- you've patented the discovery, or
- you've made the information into prior art (public)
Just inventing it earlier is not enough. [/QUOTE]

Actually I believe that is incorrect (at least for the US).

I'm not a lawyer, but as I recall as long as you documented your prior invention - then your invention has primacy. Things like lab notes and other formal documentation procedures can be used to establish this.

LetterRip

An introduction to Elliptic Curve Cryptography

oak — 2006-03-09T19:13:52+00:00

> With patents if you independently discover something that has
> already been patented, you are still infringing.

You're infringing even if you had made the discovery earlier, unless:
- you've patented the discovery, or
- you've made the information into prior art (public)
Just inventing it earlier is not enough.

Patents are about a possibility for the rich to buy (time-limited)
monopolies.

An introduction to Elliptic Curve Cryptography

jzbiciak — 2006-03-09T18:39:40+00:00

There is one benefit.... you avoid triple damages for willful infringement. Cold comfort, I know, but hey.

An introduction to Elliptic Curve Cryptography

Nelson — 2006-03-09T15:28:49+00:00

The math isn't that deep, you need to have some math background though. On the upside, we know tons and tons about elliptical curves where some of the number theory about primes is relatively light (there haven't been many advances in that area) it gives you a degree of confidence in the EC methods. Also, the speed difference is huge, it's much much faster. You can play with the EC code in OpenSSL, it's not too difficult and there are tests in the test suite. I'm not sure where it stands exactly with NSS.

An introduction to Elliptic Curve Cryptography

jamesh — 2006-03-09T05:46:46+00:00

With patents if you independently discover something that has already been patented, you are still infringing.

You might be able to claim that the idea was obvious, but you'd still be risking a lawsuit.

An introduction to Elliptic Curve Cryptography

kirkengaard — 2006-03-09T04:59:54+00:00

I feel like I'm usually more insightful than this, but here it is anyways:

Would I be right in saying, then, that the patent landscape is irrelevant if one is implementing cleanroom from the math? Or is that copyright-mindset, so you're alright cleanroom from the math as long as you don't wind up substantially similar to an existing patented implemetation? And that's where the trouble lies, in determining exactly what implementations are foreclosed?