Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!pasteur!ames!elroy!orion.cf.uci.edu!uci-ics!venera.isi.edu!raveling From: raveling@vaxb.isi.edu (Paul Raveling) Newsgroups: comp.os.misc Subject: Re: When you have to go as fast as possible Message-ID: <7699@venera.isi.edu> Date: 6 Mar 89 18:19:25 GMT References: <4434@freja.diku.dk> <3200001@hpsemc.HP.COM> <598@redsox.UUCP> Sender: news@venera.isi.edu Reply-To: raveling@isi.edu (Paul Raveling) Organization: USC-Information Sciences Institute Lines: 40 In article <598@redsox.UUCP> campbell@redsox.UUCP (Larry Campbell) writes: >In article <3200001@hpsemc.HP.COM> gph@hpsemc.HP.COM (Paul Houtz) writes: >} Careful, Karl. MS-DOS may be sold by one vendor, but it runs on more, >}different machines than Unix. MS-DOS run more programs than unix, and >}the number of MS-DOS users dwarfs the number of Unix users. > >Sorry folks, I just couldn't let this one go by. "MS-DOS ... runs on more, >different machines than Unix." That's utterly ridiculous. It runs ONLY on >Intel 8086-family chips. Period. MS-DOS will *never* run on anything else >(sigh, I know about the various software emulations of the 8086, big deal...) It's true that "IBM" PC's have reached lots more users than UNIX systems. The key reasons are: 1. Low cost 2. A well-defined hardware standard [This enabled low cost because of competition!] 3. An adequate file system standard 4. An adequate operating system standard. This isn't to say that I'm enthralled by any of the standards: The 8086-family chips, the limited UNIX-like file system, or the OS interface (originally a CP/M clone, then improved to UNIX-like). IBM and Microsoft could have and should have done better on all of these in the early evolution of PC hardware and software. I'd rate the most significant factor among these as being a hardware standard. In other markets, companies such as DEC and IBM have used the same strategy successfully in lines such as VAXes, PDP-11's, System/360's, and all the assorted child-of-360 mainframes. While I agree that portable operating systems are a good idea, it still seems that hardware engineering has been more adept at portability than UNIX-based software engineering. ---------------- Paul Raveling Raveling@isi.edu