Portable TCL-based crypto toolkit

[henry strickland <strick@osc.versant.com>]

# One thing that frustrates me is the difficulty of easily providing
# implementations of cryptographic algorithms that would be useful on a
# wide range of machines.  A lot of these algorithms are really simple,
# almost trivial.  Yet to write programs to implement them takes pages and
# pages of code, and making them portable so that people on PC's, Mac's, and
# Unix machines can use them is almost impossible.

My experience has been much better.  I do have a TCL-based crytpo
tookit running, currently on SunOS, although some of the work (RSAREF
wrappers) I did on macintosh.  I think most of the pieces in this
list port to MAC or DOS, using ANSI_C+POSIX emulation:

	tcl7.0 (John Ousterhout's "Tool Control Language") sprite.berkeley.edu /pub/tcl
	gmp (gnu miltiple precision) prep.ai.mit.edu /pub/gnu
	gdbm (gnu database manager) /pub/gnu
	alo-des (by Antti Louko (alo@kampi.hut.fi)) kampi.hut.fi
	md2, 4, 5 (reference implementation) ftp.uu.net /inet/rfc/rfc{1319,1320,1321}
	tclRawTCP (TCP socket, listen, connect for TCL) harbor.ecn.purdue.edu
	RSAREF 1.1 (beta?) <rsaref-administrator@rsa.com>, 
			includes limited license to practice RSA & DH

Along with the tcl7.0 for MAC (for THINK6.*, altho in about 2 hrs
I got it running under THINK5.* as well) comes ANSI and enough-of-POSIX
libraries, sufficient to do most of what you need.

I assume ANSI and enough-of-POSIX are available for PC as well?
(I've never done programming on PC, so I can't speak from experience.)

And I assume assume we can find TCP (Berkeley Sockets functions) for MAC and PC.


This toolkit is sufficient to do most anything we've talked about.  I
want to supplement this with more stuff -- IDEA, UDP, cme's trans,
tripleDES, etc. -- but it already contains at least one implementation
of what you need to prototype almost anything we've talked about.

TCL is the trick.  Using this toolkit, I implemented Knapsack in about
2 hours (because it was my first one), El Gamel in half an hour, and a
DH-exchanged- DES-encrypted TCL-shell session over TCP in 2 hours.
Most anything becomes a one-evening job, except DC-nets, because it has
so many componenets....

I'm trying to shape this into a release.  This will have to be a
strictly-US-citizen-in-the-USA highly-controlled release, like
RSAREF and RIPEM are.  Sorry...


					strick

					"stricks write code"


p.s. perhaps someone could mail me the ftp path to the ITAR again...  thx