From: Billy Brown (bbrown@transcient.com)
Date: Sat Apr 08 2000 - 22:00:42 MDT
Matt Gingell wrote:
> This is argument by anonymous authority. I don't find it persuasive.
Actually, its argument by summary - I don't really have time to try to think
of every argument you might possibly want to advance, especially when I'm
taking the side that almost all of the experts are on. If you think my
generalizations are wrong, give me a specific example of how.
> >There is no plausible mechanism by which Microsoft could exert such an
> >alleged advantage. The fact that they own the OS does not automatically
> >enable them to build superior programs, and it certainly doesn't let them
> >charge more or work more efficiently.
>
> This is simply false. Control and access to a proprietary kernel ABI
> and secret system interfaces confers advantage.
No, it doesn't. If Microsoft made all the most useful APIs in their OS
secret it would help their applications division, but at the expense of
crippling their OS's ability to compete. Remember, anything Microsoft can
implement, other OS vendors can also build. If those secret features are
actually important, a competitor can make a mint by building public
implementations of them.
In actual fact, Microsoft has never relied on secret APIs. Sometimes they
get lazy about documenting everything, but I have yet to see anything
especially important in an undocumented API function (and I've done a LOT of
Windows programming). What they have done is to write their apps to take
every possible advantage of their OS. Many other vendors only use OS APIs
that are common to many different platforms, which lets them port their code
easily but cripples their ability to exploit the unique advantages of each
OS. But that is a necessary result of writing cross-platform code - anyone
else who wants to write OS-specific code can get the same results as
Microsoft.
> The ability to
> distribute software via pre-install and to control what gets shipped
> with a PC via exclusionary contracts with hardware vendors confers
> advantage.
Microsoft does not own the hardware companies, and it does not have a magic
wand that controls the industry. If hardware vendors pre-install MS
software, they do so because they thing their customers want them to.
Besides, when has MS ever done this with a program that went on to become a
market leader? Computers don't come pre-loaded with Office, or SQL Server,
or anything else that MS actually makes money on. You just get free
utilities and cheap junk that no one would actually pay retail price for.
> The ability to leverage the revenue provided by the
> 'Microsoft Tax' and dump products at below cost confers advantage.
No, it isn't. A large company can not gain a competitive advantage against
smaller rivals by selling its product below cost. Why? Because the bigger it
is, the faster it looses money by dumping.
A company with multiple product lines also can not gain competitive
advantage by use revenue from sales of product A to subsidize below-cost
sales of product B. If they do, they cripple the ability of product A to
compete - their competitors can either undercut their price, or spend more
on R&D, or both.
Furthermore, neither tactic would actually do a dominant company any good in
the long run. Dumping may drive weak competitors out of the market
temporarily, but the instant you try to raise prices back to market levels
they'll be back - and if you try to raise them above market levels, you are
waving money in the face of every company that could possibly choose to
enter your market. In the mean time, the actual dumping activity has very
little potential to harm consumers - you might slow down innovation a
little, but you're giving everyone your product below cost, so the net
effect is unlikely to be negative.
All of this is standard economics today. If you don't believe me, ask any
expert in the field.
> The ability to deliberately break competing software via changes to the
> system confers advantage.
Maybe a small tactical advantage, as long as you don't do it very often.
Maybe.
The problem is that an OS that doesn't run a customer's software is
worthless. Most customers care a lot more about their apps than they do
about their OS. I would therefore say that the only way MS could get away
with this without getting clobbered in the OS market is if their competitors
their are so inept than that consumers consider them an even worse option.
Besides, once again, this can't possibly explain the actual events of the
past 15 years. Windows isn't perfect, but it maintains at least as high a
level of compatibility as Mac OS (or, for that matter, the few Unix vendors
that actually bother to add features to their products). MS goes to
considerable lengths to give everyone in the industry a chance to try out
new OS versions long before they are released, and it is pretty unusual to
encounter real problems with anything except antique software and obscure
utilities (oh, and games, but that is obviously due to the game
manufacturer's penchant for bypassing the OS to write funky,
hardware-dependant code).
> Putting aside for a moment the factual
> matter of whether such strategies were actually used, these are
> eminently plausible mechanisms.
By my count we've got two mechanisms that are considered discredited by
mainstream economics, one artful way of cutting your own throat, and one
strategy that might or might not give a dominant player some advantage
*after* if becomes dominant. None of this makes a compelling case for
massive government intervention in the market.
> Microsoft dominates the i86 operating system market. Microsoft
> dominates the Windows office suite market. Microsoft dominates the
> Windows development tool market. I don't know the break down on web
> browsers, last time I looked it was somewhere around 50% - though I
> think it's clear at this point to most observers that Netscape is on
> the way out. I don't know what the situation for databases and
> server products either - mail servers, etc. - but they're a significant
> player in those areas too. I don't think you can seriously believe
> that they've accomplished this success based solely on the technical
> merits of their products.
No, I think they did it my doing a superior job of giving customers what
they actually wanted to buy, at a price they were willing to pay. Marketing,
strategic alliances, and backroom deals with other vendors also had a role
to play, albeit a smaller one. But these are all legitimate business tools,
and the overall result has been immensely beneficial to MS's customers.
The people who object to all this mostly seem to think that the public has
somehow been flimflammed into buying an inferior product for a period of
more than a decade now. That is self-delusion talking. If people don't buy
the OS you consider to be superior, the rational conclusion is that most
people aren't looking for the same thing in an OS that you want.
> The Posix standard is an example here - because OS vendors had a
> mandate to work within a publicly available and publicly influenced
> standard, today an application written for one *nix is (reasonably)
> portable to platforms offered by a variety of different
> implementors. I can choose AIX or Solaris or Linux or whatever because
> I like it, not because it's the only place I can run Word. As a result
> there's competition in the Unix market, and I think we've all
> benefited.
If the benefits were really all that great, you wouldn't need a government
mandate - people would be buying it on their own.
There are good reasons why *nix OSs never caught on with consumers, and why
their market share is so precarious now. Unix-style OSs are optimized for
high speed, small memory footprint, low storage requirements, and maximum
programming power. In the 60s and 70s this was a great combination, which is
why they were so popular back then.
However, this isn't the 60s. These days hardware capacity is dirt cheap,
and manpower is very expensive. That means customers want an OS optimized
for ease of use and fast coding. They want gobs and gobs of functionality in
the OS, even if they rarely use it, because they've got disk space to burn
and they don't want to have to code anything themselves. And most important,
they want a vendor that will actually ask them what they want and try to
give it to them, which is something most *nix vendors aren't particularly
interested in doing.
Now, this is not to say that Unix is dead, or that Linux will never go
anywhere. To the contrary, I think there will still be people using Unix
when the singularity arrives. But complaining that customers are buying
something else is missing the point - the purpose of an industry is to give
customers what *they* want, not what you think they *ought* to want.
> Programming languages are another example. Because C is a public
> standard, I can choose compliers based on quality and appropriateness
> to a particular task. The same can't be said for Visual Basic, where
> I'm lucky if my code is even portable across versions.
There is an inevitable trade-off between backwards compatibility and
innovation. If you value stability, write C code on Unix. If you value rapid
improvement, use VB or another 4GL on Windows (or wait for MS's new
Intentional language, but that's another story).
In any case, I see absolutely no value in a government-mandated choice of
programming languages. You get maximum benefit for everyone if vendors are
free to offer any language they want to, and customers are free to buy
whichever offering they prefer. If the government wants to pick a single
standard for internal use, fine, but trying to dictate what the public has
to use is a very bad idea.
> Now note that I'm talking about a governments being restricted to
> purchasing implementations of a public standard - not that they make
> up some random junk and require that everyone observe it. I don't want
> some bureaucrat throwing darts at a spec anymore than you do. Think
> about the mandate that DoD projects use the Ada language - this is a
> reasonable model for the sort of thing I'd like to see.
Fine. What does that have to do with anti-trust law?
> Personally, I'd like to see innovation stifled. I got out of Windows
> programming because I was sick of working in an industry where
> interfaces are deprecated before they're documented, and
> the environment is obsolete before it ever works properly. I don't
> want more 'technologies.' I want less code that actually works. It
> should take forever to add anything to the standard - I don't know if
> you've done much Windows development, but it's like trying to build a
> skyscraper on quicksand. It doesn't matter how square the corners are,
> it's still going to sink. I have horror stories if this claim is too
> abstract.
I do web apps and n-tier development on Windows, so I'm right in the middle
of the mess. And, you know, I wish they would go faster. There are several
major revolutions in application development coming down the road, and I
don't want to have to wait another 2-5 years for all the pieces to come
together.
But we don't have to agree. I'm perfectly willing to let people like you go
one using whatever OS and programming technology you prefer. I just want you
to extend the same courtesy to people like me. Just let everyone take the
approach they prefer, and let the market decide which one works best.
> Because with 'features' you get incompatibility, and with
So, let's see, you're against features, you'd like to see innovation
stifled, and you think the government should make everyone use the same OS.
Which agency did you say you worked for?
Billy Brown
bbrown@transcient.com
http://www.transcient.com
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:27:55 MST