NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/arch/11760

Internet Message Format | 1992-12-12 | 5.1 KB

Xref: sparky comp.arch:11760 comp.sys.dec:6533 comp.sys.sgi:18187 comp.sys.hp:14226
Newsgroups: comp.arch,comp.sys.dec,comp.sys.sgi,comp.sys.hp
Path: sparky!uunet!mcsun!dxcern!dscomsa!zeus02.desy.de!hallam
From: [e-mail] (Phill Hallam-Baker)
Subject: Re: Comparison of Alpha, MIPS and PA-RISC-II wanted
Message-ID: <[e-mail]>
Sender: [e-mail] (usenet)
Reply-To: [e-mail]
Organization: Deutsches Elektronen Synchrotron, Experiment ZEUS bei HERA
References: <[e-mail]> <[e-mail]> <[e-mail]> <0f=[e-mail]>
Date: Fri, 18 Dec 1992 15:01:49 GMT
Lines: 97

In article <0f=[e-mail]>, Sean McLinden
<[e-mail]> writes:
es: 48
|>
|>>I don't reagard it a "real" UNIX, then again I wouldn't buy a "real" UNIX,
|>>1970s software technology is not something I would want to buy today.
|>
|>[And here I was beginning to like you because you commented, correctly, that
|>BSD was still in demand. Oh well.]
|>
|>Statements like the above excerpt I would expect from Bill Gates but not from
|>a real programmer. To say that you wouldn't use Unix because it is 1970s
|>technology is a bit like saying you wouldn't speak French (or English or...)
|>because it is centuries old.

If I wanted to spend my time looking at centuries old technology I would be an
archeologist.

|> What we call Unix, today, is hardly anything
|>like what came out of AT&T Bell Labs in the early days and is even a long
|>way from early System V and BSD 4.2 of a few years (well, maybe more than
|>a few) ago.

Bugs they still need to fix:-

    1) Online manual still designed with aim of minimizing disk space rather than
providing information.
    2) No help facility.
    3) No standardized user firendly shell.
    4) Command qualifiers hard coded into applications making multilanguage
customization impossible.
    5) Requires expensive expert to have a chance of any security.
    6) File system limited to sequential file, forcing applications to create their
own file system on top of the UNIX one, thus preventing any standardization or
application independent optimization.

|> Like most modern languages, Unix has survived because it is
|>dynamic and accommodating and while it is hardly the perfect operating
|>system (though most people's complaints, including your references to 'ls'
|>have more to do with the command shells than the kernel architecture) it is
|>not the dinosaur that the NT zealots makes it out to be.

I disagree. It passed it's sell by date years ago.

The kernel is a major factor in determining how useful the system is. Only AT&T
had the capability to introduce a shell to replace the standard ones that could
have achieved a degree of acceptance.


|>>Apollo in my view were the only UNIX vendor to realize that they had to put
|>>work into the basic operating system. They had ACLs, shared libraries and
|>many
|>>other essential features five years ago.
|>
|>They are essential for some tasks and totally unnecessary for others. It
|>depends upon the environment in which you operate. Unix didn't, originally,
|>have ACLs because the people for whom it was implemented didn't need it.
|>They had a different concept of a programming environment than you do.

They had a limited problem and provided a limited solution. I find it hard to
see how you can consider UNIX to be a superior system to VMS while admitting
that it is inadequate for many tasks. It would appear to me that the strongest
statement to be made would be that UNIX was superior for some tasks (presumably
through simplicity).


|>>What I find disgusting about UNIX is that it has *never* grown any operating
|>>system extensions of its own, all the creative work is derrived from VMS,
|>>Multics and the operating systems it killed.
|>
|>So? If originality was important Microsoft would still be a niche vendor.
|>First of all, your statement was inaccurate and inflammatory. More to the
|>point, it is meaningless. I mean, who cares. How many people have ported
|>VMS to their boxes? How many open systems are being developed in Multics?
|>What is the point?

If UNIX was ever allowed to become the sole O/S it would halt O/S deelopment
completely. The only effort UNIX has ever made is in catching up.


|>Unix (whatever that is) is not just an operating system, it is has been the
|>seed of a marvelous technosocial experiment which is still being carried out.
|>It isn't fair to say that Unix isn't everything...it wasn't intended to be.
|>But I doubt that the world would have evolved on a production basis the
|>concept of horizontal integration of systems without something like Unix to
|>make it possible.

There were other choices. PICK for instance was a very highly regarded system
but was killed through over agressive attempts to control it. Imagine a world in
which EBSIDIC, not ASCII was the standard for charaters. We would see people
praising the merits of the EBSIDIC standard as if the deliberate convolutions of
the system were an advantage rather than a serious drawback.


--

Phill Hallam-Baker