From: Matt Gingell (mjg223@is7.nyu.edu)
Date: Mon Apr 10 2000 - 22:15:30 MDT
Billy Brown wrote:
> Matt Gingell wrote:
>> The basic problem that I see is that there is no real competition in
>> the i86 operating system market.
>
> True, but let's remember the whole situation. There are several non-i86
> platforms that could suddenly become viable if either Intel or Microsoft
> stumble badly enough, and several OSs (notable Linux) that run just fine on
> the i86. So, Microsoft's position is not equivalent to being, say, the only
> diamond mine in the world. There are lots of potential competitors out
> there, and MS's market share depends on its ability to fend them off
> successfully.
We'll see, I guess. Five years ago I don't think I ever would have
believed that Linux could become as mainstream as it's becoming now.
It's still a tiny slice of the market, and I don't believe it's a
practical option for the typical desktop user, but maybe in a few
years it'll get there. We'll have to see what happens with application
support.
I met a guy a while back who was involved with a start-up to make
Window's device drivers work on Linux - device driver availability and
quality being one the major impediments to alternative OS platforms.
The idea was to put a wrapper around the binary driver interface and
try to emulate the, typically pretty lightweight, services required by
low level interface logic. They got drowned in letters from Microsoft
lawyers claiming that the interface was Microsoft intellectual
property and building software based on it constituted a violation of
copyright and the Windows license agreement. Whether the claim would
stand up didn't matter - the threat was enough for the VC to back out
and so it never happened. That's the sort of thing that really underlies
my desire to see someone bigger step in, I think. It's just so
incredibly, unnecessarily nasty - there are legitimate reasons for
copyright law, but the legal excuse to quash something genuinely
innovative and useful isn't one of them. I suppose it's the desire to
see a parent figure step in and make them play nicely. The market is a
jungle - it's survival of the fittest and I freely acknowledge that -
but there comes a point when you have to ask what sort of world you
want to live in and whether there might be some cause to ask people to
treat each other in a more civilized way.
>> Now, obviously I think monopolization is a bad thing - so you can
>> insert all those standard arguments here.
>
> You are using 'monopoly' in a different sense than I do. Having a 90% market
> share (or even 100%) doesn't make you a monopoly. A monopoly is an
> organization that can charge its customers above-market prices for its
> products, or somehow coerce them into buying products they do not wish to
> purchase, without fear of loosing ground to competitors.
>
> By that standard MS as a whole doesn't qualify, because there are no
> barriers around its markets that prevent competitors from emerging.
Well, I've pointed out that I think application support is the major
issue. I ran OS/2 from the early 90's till I finally threw in the
towel a couple of years ago. I thought then, and still do now, that it
was a better, cleaner, more innovative platform than anything offered
by MS. But eventually I got sick of fighting with the Windows
emulation every time I wanted to run something, tech support people
who wanted to know if my OS/2 was the Win 3.1 version or the Win 95
version, and trying to parse badly mangled Word documents via Ami Pro.
Now, I don't really want to open the 'why OS/2 died' can of worms - it
was bungled by IBM from the beginning - but unless you've been in love
with an unsupported platform I don't think you can appreciate the
frustration.
>> What it really comes down
>> to, I think, is that control of the PC platform is too important to
>> trust to a single corporation: there's too much potential for abuse
>> and we loose the advantages of distributed economic optimization. This
>> is an abstract compliant, not one specific to Microsoft. Perhaps you
>> don't object to monopolization as strongly as I do, either because you
>> think it's unstable or because it's a lesser evil than government
>> interference.
>
> Yes, for both of those reasons. I thing people who favor anti-trust tend to
> forget that what they are advocating is a law that requires the government
> to seek out the most successful businesses in any given market and
> confiscate their property for the benefit of their competitors. That harms
> the customers, by preventing them from buying the products they wanted.
I really think confiscation of property should be the absolute last
resort of any resolution - as much as I feel intervention is
warranted, I think it should be as minimal and as undisruptive as
possible. I'm a bit of a crackpot, but I'm not so wacky I think people
should be prevented from buying Windows if they want it or it should
be forcibly placed into the public domain. The idea of forcing
Microsoft to license the sources at some 'fair' price - whatever the
hell that might be - appeals to me. That way other vendors could take
a shot at putting out their own version, so we could see competition
in the Windows implementation market. (Does that make me a Commie
Statist?)
More general constraints seem reasonable too - like restrictions on
pre-announcement, requiring disclosure of platform changes to
competing application vendors to place them on a level field with MS.
Basic stuff like that doesn't seem overly onerous to me.
> Personally, I think this is a stasist way of looking at things. No one
> 'controls' a free market. Every purchase is voluntary. The only way a
> company can succeed in the long run is by giving customers what they want -
> and the 'long run' in the PC business is only a couple of years, so the idea
> of waiting for market forces to work things out is not a big problem for me.
We'd be going round in circles if I went into that again. I appreciate
the point but I don't share your perspective.
>> There will never, without government intervention, be an
>> implementation of Windows offered by anyone other than Microsoft. The
>> API is too big and the target is moving too fast. Consider the
>> failings of Win-OS/2, and the Wine project.
>
> Isn't this an argument against intervention? The two justifications for
> anti-trust law are to encourage innovation and to keep prices low. If your
> statement is true, that knocks out the first justification. The fact that
> low prices are a primary complaint of MS's competitors seems to knock out
> the second as well. That leaves only a general distrust of companies with
> large market share - and if you let the government confiscate private
> property on mere suspicion of possible future wrongdoing, no one will ever
> be safe.
This is very convoluted. I'm not sure I follow it.
>> So, then, what is are appropriate remedies? I'd like to see, as I've
>> suggested before, a prohibition against government itself purchasing
>> software based on non-open standards.
>
> Well, the government can certainly make any rules it likes for its internal
> purchasing. It does have one side effect you've missed, which is that it in
> effect forbids the government from ever buying from a dominant vendor. In
> the computer industry the general pattern is for dominant players in any
> given market to define their proprietary standards, which then become de
> facto industry standards if they are any good (ODBC is a good example of
> this). The minor players then band together to create an 'open' standards to
> compete - they make it open because they have to in order to pool enough
> market share to make their efforts matter.
>
> Now, requiring the government to always side with the underdogs probably
> won't really hurt anything, but will tend to lead to higher costs and more
> frequent vendor turnover than making a informed decisions on a case-by-case
> basis. Of course, the government isn't very good at informed decision
> making, so they may not be loosing much.
It doesn't prevent purchasing from the dominant vendor - they can buy
an Ada compiler because Ada is a public standard, whether the
particular vendor they're buying from is an underdog is neither here
nor there. If there's sufficient market pressure major players will
implement the standard - witness the Posix subsystem in NT. (Not, I
imagine, that anyone has ever used it...)
>> I'd like to see a market where I don't buy an 'operating system,' I
>> buy a memory manager, a tasking subsystem, a file system, a desktop
>> environment, etc, from independent vendors - and they all work
>> together because they all conform to open standards; in the same way I
>> can purchase a nut and a bolt from two different manufactures because
>> they come in standard sizes.
>
> Everyone wants software to be modular and reusable, but there isn't much
> agreement on how to achieve this goal. Is there some particular action you
> think the government could take that would yield this result?
I don't know, really. I'm just pontificating about the way I'd like
the world to be. In the Linux world, you get a X-Server, a kernel, a C
library, a desktop environment from different people. It's nice - you
cross-pollinate - and I don't think that the average user in the
Windows world understands the number of different, casually related,
pieces that make up the 'Operating System.' I have a very traditional
view of what an operating system is - basically a glorified device
driver who's sole reason for existing is to place a thin abstraction
between the hardware and user-mode software. The idea that a web
browser can be considered part of 'the operating system' is insane -
it's crazy talk. A command line shell isn't part of the OS - it's
just a piece of user software that makes system calls. In that view,
only a very small part of what Microsoft call Windows is an operating
system - the rest is environment and tools. But all that debate is
metaphysical, really, and of course these definitions evolve over
time. If you can call IE part of the OS though, then why not Office or
whatever the hell else they want to cram in there?
-matt
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:27:57 MST