NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/arch/11794

Text File | 1992-12-21 | 6.2 KB | 136 lines

Newsgroups: comp.arch
Path: sparky!uunet!spool.mu.edu!darwin.sura.net!Sirius.dfn.de!chx400!dxcern!dscomsa!zeus02.desy.de!hallam
From: [e-mail] (Phill Hallam-Baker)
Subject: Re: <None> (Should be Open Systems, bloody NEWS system...)
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]> <[e-mail]>
Date: Sat, 19 Dec 1992 20:31:23 GMT
Lines: 123

In article <[e-mail]>, [e-mail] (David Fox) writes:

|>>|>> So far, the most "open" system I can think of is a 386 running DOS:
|>>|>> - systems available from all sorts of people
|>>|>> - CPU's available from (at least) AMD, Intel, IBM, and Cyrix
|>
|>I don't agree.  For a system to be open requires it not to be
|>proprietary.  Of course, 386s running DOS are a lot more 'open' than
|>Macintoshes, for example.  But Unix (or at least _some_ unixes) foster
|>a much more open environment because source code to many things is
|>usually provided.  For example, your choices for a C compiler are
|>plentiful, but that depends on how much $$ you are willing/able to
|>shell out to the various software writers to try those compilers out.
|>With Unix, one can get gcc for free, including source.

In the old days you used to get C free with the O/S. In fact it was the biggest
selling point about ten years ago when a compiler was something few people could
afford.

GCC also works on VAX under VMS of course and under Alpha/OSF1. Despite the
complaints about the "obscure" link format of Alpha-VMS it is close enough to
the VAX on to make a GCC port at some point inevitable. Probably once people
have got over the fun of poping up Jpegs in seconds and decide to do some
serious work.

|>IBM was very open when they published the source code to the
|>PC BIOS.  However, that openness was short lived, and the PC started
|>becoming more and more proprietary.
|>
|>In my opinion, license agreements, non-disclosure agreements, and
|>trade secret information just doesn't make for 'openness.'

Yep now there is a closed system. Or rather it was intended to be. IBM did not
expect nor intend the clone market to exist. It is noticable that their
"openness" lasted just as long as it took to beat the rival ACT-Sirius machine
which was an infinitely better machine. The ACT was ahead of it's time offering
graphics (in a sensible place) and high capacity disks. IBM confined it's
inovation to the keyboard layout.


|>>I would guess that W/NTs single user functionality should outstrip that of
|>UNIX
|>>fairly easily. I bet that in '94 or whenever Microsoft will be comming out
|>with
|>
|>I don't see that.  Maybe by 1998, DOS will come with a 'dd' command.
|>>being able to offer additional functionality.
|>

zow07> man dd
Warning: couldn't stat file /zow/userb/hallam/man!
WARNING: terminal cannot clear to end of line
WARNING: terminal cannot clear screen
WARNING: terminal cannot home cursor
WARNING: terminal cannot move cursor to lower left of screen
WARNING: terminal cannot scroll backwards
-  (press RETURN)

Name
     dd - copy and convert data


Note the warnings caused by sysop hackery since my last login. Let us hope that
MSdos never has a dd command. A CONVERT command perhaps.

     conv=ascii          Convert EBCDIC to ASCII.
     conv=ebcdic         Convert ASCII to EBCDIC.
     conv=ibm            Slightly different map of ASCII to EBCDIC (see RES-
                         TRICTIONS).

Let us also hope that a future CONVERT command would consider *THESE*
conversions an irrelevance!


I think that a convert command of this kind would be something that a UNIX user
would condsider unnecessary. files should be files should be files was the UNIX
manifesto line. Now MSDOS is being crucified for the lack of a block size
conversion munger!

A little though could elimintate the need for this command entirely. The
apparent purpose of the command is to match device capabilities. This would
argue for attributes to be applied to the device:-

COPY /dev/magtape0(block_size=900,end_of_block=0)/* -
    /dev/magtape1(blocksize=512)/fred

Would seem to be a better choice.

Or better still make the tape drives intelligent enough to figure it all out for
themselves. That would appear to be the UNIX philosophy which I am not arguing
with. What I do take issue with is the frequently repeated line that a grad
student project thrown together by two hackers plus assorted hangers on should
be regarded as the culmination of O/S technology.  There are other approaches,
frequently these are superior approaches.

What I find objectionable is the idea that it is desirable computing should be
surrounded by a mistique and that initiation to the mysteries be as painfull as
possible. I wish that I did not have to spend as much time as I do answering
questions from people wanting to do a very simple task which is made complex by
poor O/S and application design.

It is not just a matter of lamebrains. If people waste computer resources
because the applications interface is baddly desingned then everyone suffers.
Like the stupidity of ftp which assumes ascii conversion mode as default. Just
think how many times *you* have forgotten the switch, how much faster would the
networks be if people did not waste resources in this manner? *You* may of
course be a genius (or someone without more important things to think about) o
never makes a mistake. But that dosen't stop the network being slower because of
*other peoples* mistakes.

Back to the question of "Open" systems. An open system to me is a system that an
uninformed but intelligent user can simply be given a password for an acccount
and a primner and told to get on with their stuff. If they have to open the
primer then the system ain't much good. I would prefer to spend more of my time
explaining to people how they can use the machine to solve difficult and
challenging problems and none at all answering problems caused my a mode of
thinking in which simple and straightforward tasks are reduced to mystic
cabalistic rites whose mysteries are to be known but to the few.


--

Phill Hallam-Baker