home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.next.misc
- Path: sparky!uunet!psinntp!cubetech.com!imladris!andrew
- From: andrew@cubetech.com
- Subject: Re: 100 Mips Intel NeXT.
- Message-ID: <1992Dec17.163753.6662@cubetech.com>
- Organization: Cube Technologies, Inc.
- References: <1992Dec13.204914.20442@fcom.cc.utah.edu> <1992Dec16.002043.9961@pencom.com> <1go1j6INN2dq@gap.caltech.edu>
- Date: Thu, 17 Dec 1992 16:37:53 GMT
- Lines: 28
-
- In article <1go1j6INN2dq@gap.caltech.edu> madler@cco.caltech.edu (Mark Adler) writes:
- >
- >>> But which chip they choose is very important and not an easy decision!
- >
- >Is it all that important? Especially considering the ease with which
- >applications are being ported to NeXTstep 486? And considering their
- >insanely-great production line that can be reprogrammed in no time?
- >There's no problem with coming out with a different processor every
- >year. Your distributed executables might start to get a little large,
- >but there would be some utility to keep only the part for your machine
- >so it wouldn't clog up your hard disk.
-
- On a technical level, it isn't important. All of the major RISC
- architectures (R4000, PA, Alpha, SPARC) are good, and NeXT can make
- NeXTSTEP scream on any of them. Also, users and even programmers
- shouldn't care what processor is in the box (except for byte-order in
- the case of programmers).
-
- BUT, for some reason users do care. That's why choosing a processor
- should be a marketing decision.
-
-
- andrew
- --
- andrew@cubetech.com | "C++'s runtime sucks shit! Virtual functions
- Andrew Loewenstern | can go straight to hell!" - Ron Pomeroy
- Cube Technologies, Inc. | ForYourEyesOnly public key:
- 0000000701B61D1ADF0DFC9C16185CEA055200000007EB4A9FEB1922065D471A89E905B5
-