Newsgroups: comp.arch
Path: utzoo!henry
From: henry@utzoo.uucp (Henry Spencer)
Subject: Re: Content Addressible Memories
Message-ID: <1988Dec9.175121.3321@utzoo.uucp>
Organization: U of Toronto Zoology
References: <12371@srcsip.UUCP> <367@enint.Wichita.NCR.COM>
Date: Fri, 9 Dec 88 17:51:21 GMT

In article <367@enint.Wichita.NCR.COM> lpelleti@enint.Wichita.NCR.COM (Larry Pelletier) writes:
>> It seems that content addressible memories, although present an
>> awful lot in the literature over the years, have never really taken off.
>> Any comments on why?
>
>There was never a "popular" programming paradigm that could really take 
>advantage of content addressible memories...

Also, and more significantly, content-addressible memories are inflexible
and are significantly more expensive to make than ordinary memory (they
are more complex and there is much less demand to bring production volumes
up).  It has generally been more cost-effective to punt this problem to
software.  Schemes like hashing, done cleverly, are often almost as good.
And they don't need expensive special hardware.

(If you want a parallel, consider Multics vs. Unix.  Despite its vastly
greater complexity, Multics is arguably a better system.  Unfortunately,
it ran only on specialized and expensive hardware.  Unix doesn't do some
things nearly as nicely, but it runs on anything, which is why it is
orders of magnitude more popular.)
-- 
SunOSish, adj:  requiring      |     Henry Spencer at U of Toronto Zoology
32-bit bug numbers.            | uunet!attcan!utzoo!henry henry@zoo.toronto.edu