NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/comp/sys/dec/4957

Internet Message Format | 1992-09-12 | 3.1 KB

Xref: sparky comp.sys.dec:4957 comp.sys.hp:10344 comp.unix.questions:10998 alt.sys.sun:3130 comp.sys.next.advocacy:2191 comp.os.os2.advocacy:5313 comp.os.ms-windows.advocacy:2173
Path: sparky!uunet!haven.umd.edu!darwin.sura.net!spool.mu.edu!olivea!sgigate!odin!twilight!zola!zuni!anchor!olson
From: [e-mail] (Dave Olson)
Newsgroups: comp.sys.dec,comp.sys.hp,comp.unix.questions,alt.sys.sun,comp.sys.next.advocacy,comp.os.os2.advocacy,comp.os.ms-windows.advocacy
Subject: Re: net.views -- What is an "Open System"?
Message-ID: <[e-mail]>
Date: 12 Sep 92 22:04:33 GMT
References: <[e-mail]> <[e-mail]>     <[e-mail]> <[e-mail]> <[e-mail]>
Sender: [e-mail] (Net News)
Organization: Silicon Graphics, Inc.  Mountain View, CA
Lines: 43

In <[e-mail]> [e-mail] (Geoff Collyer) writes:
| I am deeply and
| negatively impressed that Sun or SGI can call their systems "open", wth
| straight faces, when they don't offer anything comparable to the
| Digital PDP-11/70 Processor Handbook, which described the instruction
| set, addressing and memory management unit, memory system, floating
| point processor, I/O controllers, console, kernel memory map including
| device addresses, and device register bit layouts, in sufficient detail
| to implement an operating system when read in conjunction with the
| Peripherals Handbook, and without "open system" hype, and all in only
| 276 pages.

You'll never find anybody in engineering at SGI (at the firstline
management level or below) claiming we have open systems at the
hardware/OS level, at least, none that I know of.  Perhaps we have some
pseudo-engineers who would claim this (or 'real' engineers who might
claim it in a more limited sense, as applied to particular standards),
but not anybody who ever deals with the hardware.

For all of the systems that SGI builds that I am aware of, there does
not *exist* adequate documentation to write an OS.  More recent SGI
systems are getting closer, but I (personally, not speaking officially,
etc. etc.) don't believe any of them are there yet.   They are close
enough now that with a number of good logic analyzers, etc., you could
probably do it, but it wouldn't be easy.  Those who have done it to
date have all had reasonably good contacts with SGI to supply the
missing info as far as I have been able to determine.

I would imagine this is true of many/most of the the other system
vendors whose marketing folks claim that they have open systems,
although I have no recent first hand experience with others.

Let's face it, "Open Systems" as used these days is a marketing, not an
engineering concept, so you will likely never come up with a good,
solid definition.

Some other term should probably be invented and used for talking about
systems that are well documented at an architectural and hardware
implementation level.
--
Let no one tell me that silence gives consent,  |   Dave Olson
because whoever is silent dissents.             |   Silicon Graphics, Inc.
    Maria Isabel Barreno                        |   [e-mail]