Subject: Re: How accurate is Metcalfe's law? (Was: Ximian software)
From: "Gerald P. Dwyer, Jr." <>
Date: Sun, 06 Jan 2002 11:10:30 -0500

At 04:31 PM 1/6/02 +0100, Bernard Lang wrote:
>On Sun, Jan 06, 2002 at 08:51:39AM -0500, Gerald P. Dwyer, Jr. wrote:
> > At 06:02 PM 1/5/02 -0800, Josh wrote:
> > > >  Bill Gates has become very very wealthy by
> > > > selling relatively shoddy software. This
> > > > happened because his revenue was based on license
> > > > rather than on quality developer time, and that
> > > > happened because society granted him a property
> > > > right in something which is not obviously property.
> > >
> > >I have to disagree here. From being a Microsoft fan since the 1980's, as a
> > >kid who came to know DOS quite well, I watched Microsoft get to where they
> > >are, and I honestly believe it's their marketing / mindshare that got them
> > >there. I believe the licensing issues were not a factor at all, nor any
> > >other IP issues. It was all marketing, and some backstabbing. The fact is,
> > >Microsoft knows their customers quite well (doesn't always listen to 
> them).
> > >They know them a lot better than IBM did at that time.
> >
> > It is fun to bash Microsoft. I also think that it misses some things that
> > MS has done right technically.
> >
> > The PC world is standardized on software, not hardware. That's the big
> > difference between MS and Sun. MS software will run on almost anything.
>is this a joke, or can you explain.
>   There were several attempts to port MS on other platforms than the
>PC, and they were all abandonned at some point.  But Linux and BSD
>   Or do you mean all PCs ... and then, they got much help from
>hardware manufaturers who could not sell their components without
>making sure it was MS compliant.  And Linux or BSD did not have that

To clarify, I am not comparing MS DOS and Linux or BSD or any other free 
software. I am comparing MS DOS to Sun UNIX, HP's, IBM's and a lot of other 
solutions that tied the hardware and the software. MS did not sell hardware 
tied to their software. MS adapted to the hardware that PC sellers wanted 
to sell. PC sellers wanted to sell hardware that ran a cheap OS with a lot 
of cheap software available.

MS did not do this because they wanted to be benevolent. The hardware 
sellers and manufacturers did not do this to be benevolent either.

Linux and BSD have a shot precisely because they run on a lot of hardware 
available from different vendors.

> > This is hard to accomplish and I think that MS does a pretty decent job of
> > it. OS/2 was better software and there were installation problems there 
> too.
> >
> > IBM and MS are largely responsible for the standardization on software.
>IBM did it ...  MS just benefitted ... if they could have prevented
>you from looking at DOS code, they would have ...
>   One major reason Apple lost is that they owned both software and
>hardware, and kept it proprietary.
>    IBM did not believe, and is why they won ... they did not try to be
>very proprietary.  At the time, harware was more costly than software,
>and the competition on hardware permitted by IBM allowed both design
>improvements and loower prices.
>   That, and the fact that the name IBM (and even IBM compatible) was a
>marketing advantage.  MS just benefitted from the situation.

Arguably, IBM lost. They should have bought DOS. Gates worked out a great 
deal for MS. I agree that IBM's name was crucial.
It is not clear that MS would have kept people from looking at the OS. 
There was a lot of Assembly language programming back then. Utilities, all 
kinds of stuff. Spreadsheets and word processors were in 8086 Assembler.

>   And the lesson should be in favor of free software.  It was in a
>sens free hardware: anyone could modify and reproduce ... for a
>licence fee that was small enough compared to the cost of reproducing.
>Of course, in the case of software, the licence fee is seldom small
>enough since producing software can be done without monetary
>investment, and the marginal cost is close to zero.
>    This comparison gives me an idea regarding software patenting.
>Since the economy of software is without marginal cost, and only in
>initial investment, if there are software patents or similar
>protection schemes, licencing should not be on a royalty base, but
>only in up front payment, so as not to distort market structure.

An up-front payment would be more "efficient" in economists' terms if you 
could get software produced that people want to use. Some people, who have 
more faith in the government than I do, have suggested that the whole 
patent and copyright system should be replaced by government rewards.

Free software may be a more "efficient" solution than patents and 
proprietary, copyrighted code. I think so.

> > MS used to take care of programmers. There was a continual battle between
> > MS and Borland. Now, there hasn't been an upgrade to the C/C++ compiler in
> > several years. Nobody wants to write anything that'll be widely 
> distributed
> > on Windows unless it's a niche product. As Karsten Self said, they've 
> eaten
> > their seed corn.
>Is that so ?  I am sure people still produce for windows because it is
>the larger market in the short term, which is all a start-up company
>can consider.  Yes, it is suicide in the long term if they are
>successful ...  but it is better to die later than right away.
>   If the seed-eating theory were any good, people like Adobe should
>have more stuff running on linux.

I think that Adobe will. I may be wrong. Mathematica runs on Linux. I am 
pretty sure that SAS runs on Linux. We'll see.