[Cryptography] 40 years of "Diffie-Hellman"

[Jeff Burdges]

On Mon, 2016-06-27 at 16:10 -0700, Jon Callas wrote:
> > On Jun 26, 2016, at 10:11 AM, Ray Dillinger <bear at sonic.net> wrote:
> > 
> > Maybe, but it's arguable whether it counts if he didn't publish.
> > The world at large didn't benefit from his invention of it, but
> > we all benefit from Diffie and Hellman's (re)invention of it. I
> > say credit where credit's due means give the credit to the guys
> > who did something that gives the world a benefit. It's not like
> > they stole it from him.
> 
> Hear, hear.
> 
> Invention credit and scholarly credit goes to to those who publish
>  or at least publicize. 

I dislike that phrasing as it might endorse plagiarism if taken too
literally.

It's simply that if you willfully fail to contribute to human knowledge
then we should probably not name a discovery after you.  This should
apply even if you were compelled by preexisting obligations, like a
contract or security clearance, although obviously one could imagine
exceptions. 

Imagine some GCHQ person had anonymously published Diffie-Hellman key
exchange first, so we would not be calling it Diffie-Hellman.  If that
person was Williamson, and we know that fact, then we should call it
Williamson key exchange.  If that person was not Williamson, then we
should probably call it modular key exchange, even if the anonymous leak
outed Williamson as the discoverer. 

We nevertheless always have an obligation to acknowledge other
discoverers if we know them.  Wikipedia handles this correctly with
Williamson being acknowledged in the third paragraph : 
https://en.wikipedia.org/wiki/Diffie%E2%80%93Hellman_key_exchange 

In the case of Diffie-Hellman, at least Diffie's stated motivations
represents a major early contribution to cyperpunk (political)
philosophy.  Williamson cannot lay claim to that, so the name and lion's
share of the credit unquestionably belong to Diffie and Hellman. 

Jeff


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20160628/2d8b5c7a/attachment.sig>