break microsoft!!!

re: the recent message that windows 95 has weak password encryption
due to a 32 bit random seed sent to RC4--

this seems like another EXCELLENT opportunity for some major 
cypherpunks press coverage and feather-in-the-cap accolades.

remember, the last break of the Netscape RC4 algorithm (or was it
MD5?) got front page NYT coverage and reverberated throughout the
entire media.

a "cypherpunk" accomplishing the same thing for Microsoft would 
potentially get *major* notice.

I imagine a short .exe file that when run on the proper computer
prints out passwords as it cracks them. this would get the
attention of a LOT of people.


I have the MS C++ compiler, and if someone could discuss the difficulty
of writing this proposed "exe" file, the rough time required to break
the keys, and other considerations (physical access to server computer
required?)  and maybe point to code pieces on the
net (RC4 etc.) to pull it off, I might start the effort myself.

Be careful not to sound too gleeful, lest you play into the evil nasty 
hacker stereotype. Keep the focus on the fact that real encryption is 
both possible and highly desired; the bad guys are lazy programmers and 
the US Government.

I have sent a pointer to the sci.crypt article to the win95netbugs list, 
which currently has eight Microsoft employees and nine major computer 
magazines on it. I might mention it to Microsoft's "technical people" 
when they drop by next week to address our networking concerns.

The answer, for anyone desiring one, is to turn off Win95's "multiple 
user profiles" features, turn off "encrypted password caching," and 
advertise the fact that Win95 is a totally insecure single-user OS, and 
will continue to be so as long as it uses the 1970's-vintage FAT file 
system. If real security is not available, the goal should be to 
eliminate the false sense of security that encourages people to leave 
sensitive files out in the open.

-rich