*BSD News Article 31978


Return to BSD News archive

Xref: sserve comp.os.386bsd.development:2291 comp.os.linux.admin:9371
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!agate!howland.reston.ans.net!EU.net!Germany.EU.net!lemis!grog
From: grog@lemis.uucp (Greg Lehey)
Newsgroups: comp.os.386bsd.development,comp.os.linux.admin
Subject: Re: All UNIXs should merge!!!
Message-ID: <3121@adagio.lemis.uucp>
Date: 16 Jun 94 07:20:45 GMT
References: <2tj1bi$b6c@harbinger.cc.monash.edu.au>
Followup-To: comp.os.386bsd.development
Organization: LEMIS, W-6324 Feldatal, Germany
Lines: 94

In article <2tj1bi$b6c@harbinger.cc.monash.edu.au> billm@jacobi.maths.monash.edu.au (WE Metzenthen) writes:
>[ Article crossposted from comp.os.linux.admin ]
>[ Author was matt.kangas@analog.com ]
>[ Posted on Thu, 19 May 1994 18:48:31 GMT ]
>
>In article <1994May19.155350.9049@liberty.uc.wlu.edu> zbuckner@sage.wlu.edu (Zach Buckner) writes:
>>This may sound like a stupid question, but why can't all UN*Xs merge to
>>form one single operating system, which would run on MANY different computer
>>types?
>
>You're right... it is a stupid question. People have been asking this for
>at least the last decade, and it's still a stupid question. ;-)

The best reason for calling it a stupid question is because it has
been done, and the product is out there on the market. Where have you
people been?

One of the main reasons why people got so upset about the USL/UCB/BSDI
lawsuit is because USL's offering has borrowed liberally from BSD over
the last 15 years - look at the diagram on page 5 of Leffler&Co for an
overview. In System V.3.2, they folded in Xenix. In System V.4, they
folded in 4.3BSD and SunOs 4. Rumours are that the kitchen sink is in
there somewhere, too, but nobody can find it.

The result is that SVR4 is the sum of the three (four?) threads.
Without achieving much more than the BSD flavours now on the
market(*), the stock SVR4.2 kernel on a PC runs to 1.8 MB.

 (*) OK, OK, there are some things that SVR4.2 has, like shared
 libraries, shared memory, semaphores, STREAMS. I don't want to get in
 a religious argument about their merits, but with the exception of
 STREAMS it's reasonably easy to demonstrate that they don't take up
 anything near this much space.

>>I know that UNIX has been around for a while, but sooner or later..... There's
>>going to be a need for organization. A program, in my opinion should run
>>on any UNIX with compilation.

Good requirement. Unfortunately, SVR4 is one of the most *difficult*
UNIX platforms to port to. This isn't directly SVR4's fault: the
definition is well-stated and reasonably complete, but most
generally-available software has been developed on BSD-flavoured
systems, and it shows. I would expect (though I don't have any direct
proof) that it's easier to port most software to BSD than it is to
Linux, for the same reason.

Another point to be made here is that it isn't usually the kernel that
makes it difficult to port: it's the file system organization and the
header files. These are (very gradually) being fixed to be compliant
with POSIX.1 and ANSI C, two standards which *are* having a positive
effect. Interestingly, the Linux, BSD and FSF camps are closer to the
goal than most commercial implementations.

>>All the UNIX makers should merge to form a committee which SETS STANDARDS. If
>>a change needs to be made, it is mentioned at a committee meeting, and all
>>UNIXs change.
>
>Har har. And you think committees get anything accomplished? A committee
>would have killed Linux off long ago... it is only because it is the 
>product of anarchy that it has succeeded. (Think about that for a
while...)

Agreed. But I, for one, will be very interested to see how Linux
develops in the next couple of years. This anarchy could kill it, or
it could be a raging success, displacing commercial UNIX
implementations. I wish I knew which.

>Take a look into the history of UNIX, laddy - why BSD is different from
>System V at all (thank goodness!), what happened to Univel and the 
>OSF (Hint: Open Software Foundation is an oxymoron), etc.

The fact is that committees tend to define a uniform, rather than a
good implementation. Their speed of reaction can be best comprehended
by a geologist.  You should try SVR4 compared to Linux or BSD.  I
don't think you'll like it--I suspect that many people wouldn't switch
to it even if it were free. Compared to Linux it *does* have the
advantage of not changing all the time--even the bugs seem to stay in
there for ever. This is a result of the Big Business attitude of "the
source code is holy, you can't change that unless you're a high
priest", along with product maintenence strategies which are closely
tied in with (possibly incorrect) perception of future revenue
implications.

>If corporate committes decided everything, we'd all be using AIX. Yow!

Wrong. AIX was developed out of frustration with industry standards
committees. IBM has always preferred to do what it wants without
getting committee approval ahead of time.

Greg-- 
---------------------------------------------
Greg Lehey              | Tel:  +49-6637-1488              
LEMIS, Schellnhausen 2, | Fax:  +49-6637-1489
36325 Feldatal, Germany | Mail: grog@lemis.de