From: proff@suburbia.net
Date: Sun, 12 Jan 1997 01:11:32 -0800 (PST)
To: DataETRsch@aol.com
Subject: Re: IMDMP 8192 PKCS and IMDMP Summary
Message-ID: <19970112091106.18965.qmail@suburbia.net>
MIME-Version: 1.0
Content-Type: text/plain


Ok. I admit it. It is beneath me, but how could I resist such temptation?


> Hi,
> 
[..]
> cryptosystem method. The future versions will also support multi-party key
> integration features, four platform independent random number generator
			^^^^^^^^^^^^^^^^^^^^^^^^^(1)

> algorithms, as well as 1024 (8192 bit) keys. Note that the current version of
                         ^^^^^^^^^^^^^^^(2)
> IMDMP only allows 256 byte (1024) bit keys. The current shareware version and
                    ^^^^^^^^^^^^^^^^^^^(3)
> the future shareware version of UDCM will still only allow 5 byte (40 bit)
                                                             ^^^^^^^^^^^^^^(4)
> keys so as to comply with ITAR unless a key recovery infrastructure is
                ^^^^^^^^^^^^^^^(5)
[..]
> partially detailed summary of the IMDMP is now included in VENDOR.DOC file of
> the UDCM V2.0 software package / archive. UDCM V2.0 was modified to restrict
> keys to 50 bits so as to comply with the latest ITAR details. The extensively
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^(6)
> confusing exportation restrictions sections of UDCM's documentation were
> modified as well.
> 
> methods are not too comparable to PKCS methods. What I was referring to when
> I said IMDMP is more advanced than RSA, etc. is the actual encryption
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^(7)
> procedure itself, not the way keys are secured. Again, irrashional claims
                                                         ^^^^^^^^^^^(8)
[..]
> AND-ing algorithm without trying it first. I find it extremely hard to
> believe that the celebrated creator(s) of Blowfish, IDEA, etc. had to go
> through all of this ritualistic screening complexity too. (Please do correct
> me if I am wrong.)
^^^^^^^^^^^^^^^^^^^^(9)

> 
> (For the record: DataET Research's promotional agent has been fired.)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^(10)
> Thank you very much for your time.
> 
> Jeremy K. Yu-Ramos
> President
> DataET Research
> Data Engineering Technologies
> 

Eww.

(1)     Not possible, you need real-world entropy. Sadly, there is
	yet there is no platform independent way to efficiently
	collect it.

(2)     For symmetric ciphers, keylengths over 128 bits do not gain
	you anything - unless the cipher is poorly written and does
	not use all of its keyspace or leaks bits of the keyspace
	into the ciphertext. 2^128 is such an immense number that
	short of a fundamental change in complexity theory, exhaustive
	keysearching a 2^128 keyspace will never be a possible
	attack scenario, regardless of increases in computation
	speed.

(3)	"256 byte (1024) bit keys". Hmm. 256 x 8 = 2048

(4)	"5 byte (40 bit) keys". 5 x 8 = 40 and

(5)	Says (4) is to comply with ITAR. But...

(6)	"50 bits so as to comply with the latest ITAR details."

(7)	Find the oxymoron.

(8)	You're wrong.

(9)	That "irrashional" spelling, at it again.

(10)	Then why is he still writing?

-J