Ann: NOISE.SYS v0.4.1 (should be) available...

Latest version of NOISE.SYS has been uploaded to ftp.funet.fi,
and a few other sites, as "noise04.zip".

It is a random-noise device driver for DOS, which samples fast
timings between keystrokes, disk access, clock-drift, and even mouse
movement or audio card and hashes with SHA-2 algorithm to generate
some good-quality randomness.

Note changes since previous version: this one defines two devices akin
to random.c patch for Linux, /dev/random and /dev/urandom.  The latter
will output as many bytes as are requested, while the first will only
output as many bits as are estimated to be in the pool.

Source included. 386 req'd.

Take care,

--Rob

(Still waiting on comments or suggestions about earlier versions...)

On Wed, 17 Jan 1996, Mutatis Mutantdis wrote:

> It is a random-noise device driver for DOS, which samples fast
> timings between keystrokes, disk access, clock-drift, and even mouse
> movement or audio card and hashes with SHA-2 algorithm to generate
> some good-quality randomness.

Schneier mentioned last year in one of his conference reports that SHA
was being revised, yet I couldn't find it in Applied Crypto 2 (I admit
that I don't yet own the new one, and I haven't taken a good enough look 
while browsing it in the bookstores), anyone have any pointers to the new 
spec?

Please correct me if I am wrong.
TIA

I don't know if the revision is official or proposed. I first heard
about it in a post to alt.security (I saved the message somewhere)
which contained ref's in the federal register.  I've seen other
implementations that make the same fix.

The difference that when the expansion function is performed, it rolls
the dword 1 bit left before putting it in the W[] array.

--Rob

s1018954@aix2.uottawa.ca writes:

[..]
>Schneier mentioned last year in one of his conference reports that SHA
>was being revised, yet I couldn't find it in Applied Crypto 2 (I admit
>that I don't yet own the new one, and I haven't taken a good enough look 
>while browsing it in the bookstores), anyone have any pointers to the new 
>spec?

>Please correct me if I am wrong.
>TIA

-----BEGIN PGP SIGNED MESSAGE-----

Rob writes:
> I don't know if the revision is official or proposed. I first heard
> about it in a post to alt.security (I saved the message somewhere)
> which contained ref's in the federal register.  I've seen other
> implementations that make the same fix.
> 
> The difference that when the expansion function is performed, it rolls
> the dword 1 bit left before putting it in the W[] array.

Any particular reason someone called this SHA-2 ?  It sounds a whole lot like
the revision of the original SHA, called SHA-1, that came out quite a while
ago. (FIPS 180-1)  This is rather old hat unless they're making a _second_
revision to the standard, in which case I expect there would have been much
more noise made about it.

Futplex <futplex@pseudonym.com>

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQEVAwUBMQoL0inaAKQPVHDZAQGzmQf9FLDvD9TmpMfgDDac0xHsJX8RspJ/tIfS
yMU6eoVSclD1hdQzMxkSc1ffPxvrLvCzILeFZVzZ/4duAp2wn1q4GPnQRvjXh98V
GXVhHusiyB4RFWOsUewXt7r4aYtPeIZI51WEnRMXanCjcVU2ChukiruLAEQqC1JS
nInfVMNjNkb1IHrltnwznnfqY91xBRzrABI1s8dRFXU/jUAI+jGr3ThfMipowvwh
egbBkrhQJjlS3J9f2XL0rte0NDO5WxL5MrdR/N54ODI9ktrhWXWrAeK/NbA4tm6I
uLrHq8FiI6HhqbrO7cEMMU2cuODv3Yu/0Z/MyD03C/uO1D0m1m1VRg==
=zI2p
-----END PGP SIGNATURE-----

wlkngowl@unix.asb.com (Mutatis Mutantdis) wrote:


>>Schneier mentioned last year in one of his conference reports that SHA
>>was being revised, yet I couldn't find it in Applied Crypto 2 (I admit
>I don't know if the revision is official or proposed. I first heard
>about it in a post to alt.security (I saved the message somewhere)

Official--it's called FIPS PUB 180-1.
Michael J. Markowitz, VP R&D     mjmarkowitz@attmail.com
Information Security Corp.       847-405-0500
Deerfield, IL  60015             Fax: 847-405-0506