Re: Fwd: Re: [Cryptography] Encryption opinion

Nonsense.  There are many examples of successful capability system.  Even some that vanished were terminated for business reasons having nothing to do with the capability underpinnings.  The stated explanation, that it's too hard to manage policy, is even more ridiculous.  Capability systems naturally have dynamic policies.  It's policies embodied in ACLs and roles that get more or less frozen.  As for understandability, nothing stops you from handing out capabilities based on ACLs or roles.  Set up the ACLs (roles).  Hand out the capabilities.  Forget the ACLs (roles) ever existed.  Problem solved.

Even the statement about the Connection Machine is wrong.  The CM-2 was far superior to anything in its niche.  I know because I was chairing the Gordon Bell Prize judges' committee at that time.  What happened is that management wasn't happy with being in a niche and bet the company on the CM-5.  I interviewed with them at the time and asked if they could "survive the 5."  We know the answer.  A less ambitious company would have been happy in its <$1B/year niche.  A larger company would have survived the failure of the CM-5.  (By the way, I believe a simple feature could have saved the CM-5 by allowing it to switch from SIMD to MIMD in a few hops in its internal network.)

________________________
Alan Karp
Principal Scientist
Enterprise Services, Office of the CTO
Hewlett-Packard Company
1501 Page Mill Road
Palo Alto, CA 94304
1 (650) 386-4568
http://www.hpl.hp.com/personal/Alan_Karp