1995-01-06 - Re: floating point crypto?

Header Data

From: tcmay@netcom.com (Timothy C. May)
To: rah@shipwright.com (Robert Hettinga)
Message Hash: c3a088ab14783c68012b324893f8e9d43bbe8a9d45e6112d3ec8b989d0730454
Message ID: <199501060251.SAA26936@netcom13.netcom.com>
Reply To: <v01510100ab3251fa70de@[]>
UTC Datetime: 1995-01-06 02:56:29 UTC
Raw Date: Thu, 5 Jan 95 18:56:29 PST

Raw message

From: tcmay@netcom.com (Timothy C. May)
Date: Thu, 5 Jan 95 18:56:29 PST
To: rah@shipwright.com (Robert Hettinga)
Subject: Re: floating point crypto?
In-Reply-To: <v01510100ab3251fa70de@[]>
Message-ID: <199501060251.SAA26936@netcom13.netcom.com>
MIME-Version: 1.0
Content-Type: text/plain

Robert H. has asked that we reply in e-mail to him, to avoid
"cluttering the list more than I already have...," but the logic of
this is faulty. The few lines of a response such as this one, or even
of several such responses, are as nothing compared to dozens or more
people sifting their own archives so they can each independently send
Robert what they find. Hence my public reply.

Robert Hettinga wrote:

> I casually mentioned somewhere else that I saw something on this list about
> floating-point math being used in crypto, contrary to popular belief, and
> somebody had the *timerity* to call me on it. ;-).
> I think it had to do with factoring, but maybe even in key-generation,
> though that doesn't sound right at all...

The thread was "Pentium bug and CRYPTO," and it hit on 1994-11-21 and
lasted a few days. Posts by Derek Atkins, Mike Duvos, and others
stated persuasively that no floating point operations are included in
PGP, that no FP coprocessor is needed or used for PGP, and that the
Pentium bug could not affect PGP.

(In another thread, which I have no intention of trying to dig up now,
though I recall either Norm Hardy or Hal Finney was one of those to
comment, it was noted that some clever uses of floating point hardware
can help with ostensibly integer-only computations. But PGP, as noted
above, does not do this, and I expect this trick is not common.)

> So, are there c-punk archives I could look in? I remember hearing something
> about that, too.
> However, if someone remembers off the top of their head, or if they have an
> actual copy of the posting, that would be great, too.
> Please send me whatever it is by e-mail. No point cluttering the list more
> than I already have...

(I will send Robert several of these article, so others don't have to.
Game theory and all that good stuff.)

> Of all the nerve....

Not to sound strident, but if folks would keep copies of articles and
spend some time organizing them in data bases or in other searchable
forms, this would help the list. In my opinion, having personal access
to past posts is several orders of magnitude more important than
having MIDI-MIME JPEG-II TeX players that can display "Cypherpunks R
Us" in the correct font and with the "R" reversed according to spec.

--Tim May

Timothy C. May         | Crypto Anarchy: encryption, digital money,  
tcmay@netcom.com       | anonymous networks, digital pseudonyms, zero
408-688-5409           | knowledge, reputations, information markets, 
W.A.S.T.E.: Aptos, CA  | black markets, collapse of governments.
Higher Power: 2^859433 | Public Key: PGP and MailSafe available.
Cypherpunks list: majordomo@toad.com with body message of only: 
subscribe cypherpunks. FAQ available at ftp.netcom.com in pub/tc/tcmay