Re: It's the economy, stupid

by "Ben Z. Tels" <optimusb(at)stack.nl>

 Date:  Sat, 8 Aug 1998 13:51:05 +0200
 To:  <hwg-software(at)hwg.org>
  todo: View Thread, Original
-----Original Message-----
From: Robin S. Socha <r.socha(at)control-risks.de>
To: hwg-software(at)hwg.org <hwg-software(at)hwg.org>
Date: vrijdag 7 augustus 1998 23:44
Subject: Re: It's the economy, stupid


>> I believe strongly that eventually MS will conquer the problems that
>> they built into NT and it will reach the point where it is a viable
>> enterprise scale server OS.

>That assumes that U*ix will not be developed any further.

Why? Doesn't mean anything like that. Development of any type, flavor or
derivative of Unix is completely independent from development of Windows NT.
Server or Workstation.

>This will not
>happen. It also assumes that MS will for the first time ever produce a
>program that is both bug-free and actively maintained.

This is also an assumption based on nothing. No software package the size of
a full OS is ever bug-free; not Windows, not MacOS and not even any Unix.
Even the best programmers make typos.
As for maintenance, Microsoft maintains all active OSes. I would assume most
UNIX-vendors do the same.

>This will not
>happen, either. If you want to sell a commercial product, you have to
>add new "features" which will inadvertantly break the program.

Or will NOT inadvertently break the program. Side-effects are never
guaranteed.

>Following
>the Cathedral Style of Programming, you will *never* be able to get rid
>of the bugs or even reliably detect them before release.

No style of programming will ever do that. If there were a way to reliably
detect bugs in software, don't you think it would be in wide use? The only
way to detect a bug is to stumble into it; this is the nature of bugs.

>Free U*ices
>follow the Bazaar Style of Programming and are "beta tested" by hundreds
>of thousands of users - who can (at least theoretically, because the
>sources are open) fix the bugs.


Fix the bugs, create new ones...... But as with what you call "cathedral
style programming", ONLY if they find those bugs.

>> There are pro's and con's to all OS's.
>
>This sentence does not have any meaning, does it?


It does. In any situation, one OS may have certain advantages over all
others. And also disadvantages that others do not have. For instance, you
would not wish to use a time-sharing OS in a situation where real-time
performance is required. The time-sharing OS might have wonderful features;
it might even correctly implement Shortest-Job-First scheduling (and THAT is
hard to do). But you STILL wouldn't want to use it as a real-time OS,
because it simply does not have the same (required) operating conditions.

>> NT has its strengths and its adherents just as UNIX has its own set
>> of strengths and adherents.
>
>That, however, is incorrect. NT claims to be a replacement for U*ix as
>a networking OS. Therefore, it *has* to have strengths where U*ix has
>weaknesses and vice versa.

No, that means it has to be able to facilitate network communication and
synchronization. It must correctly distribute processes over processors as
required and see to it that all processes have progress. It must facilitate
access to shared data for all users and must adequately protect data from
users not authorized to access that data (this requirement holds for data
both in primary and secondary store). It must facilitate communication
between processes as necessary and must synchronize the use of resources. It
should also avoid trap states. These are the services it must/should
deliver.

>What are these strenghts, Alan? I have used
>many OSes and I have used both commercial U*ices and NT to a larger
>extent (both as an admin in University and in my jobs). NT has no
>advantages over the U*ices I've worked with except that the food they
>server at their propaganda meetings is usually better than the crisps
>and coke one usually gets at a Linux hacking run.

The point about an OS is not that it is BETTER than any OS, the point is
that it gets the job done. Any OS or even generally program which
accomplished the task it was meant to accomplish, is a good, correct
program.

>*rotfl* REALITY CHECK 1-2-3!!![1] NT is competing with BSD and Linux -
>f-o-r-g-e-t it. Forever. NT is a third rate file and print server
>that cost infinitly more money than both above U*ices - and will
>inevitably lead to losses in productivity and security.


I would disagree, but it seems somewhat irrelevant to do so.

>Does the word "scalability" ring a bell? Unlike in real life, size
>doesn't matter in the Real Computer World(c). Give me hardware and
>I'll give you power - *my* OSes are scalable.


The idiom "size does not matter" is incorrect in this case. It is a
misconception.

>Open sources guarantee security and stability. They also gurantee
>progess.

I see no direct relationship between the two. Open sources guarantee that
anybody may see how something works. Security and stability follow from the
routines involved with these two entities functioning correctly. Making the
sources open does not guarantee their functioning correctly.

You seem to think that the fact that anybody may make changes means that
every change is for the better. It seems very likely to me that allowing
anybody to change anything introduces as many bugs as it removes.

>> Is one system inherently better than the other? (Emphasis on the
>> word "inherent" please!) Heck if I know.
>
>Get a clue, then. Install Linux and see for yourself. Read the advocacy
>groups and check the U*ix-users' expertise against that of the NT-users'
>- what a joke.


The correct answer to the question "Is one system inherently better than
another" is no. It depends on the operating goals for each system. One
cannot say that a OS intended to run a network is inherently better than an
OS intended to run only a single processor. It depends on what you want to
achieve.

As for reading the advocacy groups etc, I shall not do so. I see no reason
to, especially since they are offered in support of a ridiculous point.
Also, if for some strange reason I would ever try to determine which OS is
the best (ridiculous question, BTW) I would hardly consider ADVOCACY
NEWSGROUPS (which are intended for people who want to pick fights) a
reliable source of information.

Ben Z. Tels
optimusb(at)stack.nl
http://www.stack.nl/~optimusb/
UIN:2474460

"The Earth is the cradle of the mind, but one cannot stay in the cradle
forever."
                                        --Tsiolkovsky

HWG: hwg-software mailing list archives, maintained by Webmasters @ IWA