mitra@cdp.UUCP (03/14/89)
As I understood it DES was artificially reduced in complexity by the NSA so that they could crack it and no-one else - this was supposed to be the reason that they objected to it's implimintation in software cos the complexity might then be increased. If this is true cant we go to a larger key or larger whatever to increase the complexity beyond commonly available machines (and ideally the NSA!) again.
caulkins@cdp.UUCP (03/14/89)
I attended a meeting with crypto guru Marty Hellman many years ago in which he argued strongly for a larger key size for DES. The government people present (including those from NSA) argued against this on the grounds that the key size was large enough. Hellman made the same accusation: that they wanted a more easily broken DES. They denied this. As far as I know the accusation was never proved, although it is clear that a larger key would make it lots more secure. As I recall Hellman was arguing for 100 bits. All this revolved around the design of the then-unreleased DES implementations in silicon. As far as I know NSA or anyone else has never objected to making stronger DES implementations in software. Of course, any such would be non-standard and thus little used. Dave C
new@udel.EDU (Darren New) (03/15/89)
Rather than inventing another and greater cryptography system with longer keys, hoping to avoid whatever NSA did to DES, why not just go back to the original Lucifer specification? If NSA actually did simplify DES for their own purposes, this would make great sense. In addition, there is probably already much study completed about the security of Lucifer. - Darren New