Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 15 Nov 1999 20:40:11 +0000 (GMT)
From:      Terry Lambert <tlambert@primenet.com>
To:        davids@webmaster.com (David Schwartz)
Cc:        keramida@ceid.upatras.gr, freebsd-chat@FreeBSD.ORG
Subject:   Re: Judge: "Gates Was Main Culprit"
Message-ID:  <199911152040.NAA22265@usr06.primenet.com>
In-Reply-To: <000101bf2d70$84b9a810$021d85d1@youwant.to> from "David Schwartz" at Nov 12, 99 04:46:27 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> > > 	Yes, and they crushed it by putting out a superior product.
> > > What is wrong with that?
> >
> > Nothing, but IE is not a superior product.  Superiority is never defined
> > as taking advantage of `internal' knowledge of the OS, in order to make
> > a program load faster.
> 
> Ahh, I see. Superior has nothing to do with user experience. Superior is to
> be judged by experts according to arcane principles. That's much better than
> letting the market decide.


Actually, from a user experience perspective, and from an "outernet
portal" perspective, integrating the browser into the desktop is a
win.

However, Microsoft engaged in predatory tactics by doing this, _and_
controlling the default destination (the "outernet portal"), _and_
bundling the developement costs into the OS pricing.  They did a
similar thing to Artisoft and Banyon, both of which had arguably
better peer-to-peer networking, including user level security that
exceeds what is available in Windows 2000's peer-to-peer code; both
companies were strangled by Microsoft integrating peer-to-peer
networking in the base OS; never mind that the networking was
intentionally not very good to ensure future sales of NT servers,
it was "good enough" to leverage to strangle the competition.

IE is not the best browser... though I will give them the GIF
decoding JNI in their Java/JavaScript, since this is much faster
than Netscape's "pure Java" implementation.

On the other hand, their clients consistantly do truly stupid
things, such as attempting to assign meaning to the contents of
RFC822 encoded data streams during download.  This means that
the application that fails because of poor interpretation code,
such as the OutLook mail client or IE browser, ends up not
completitng the transaction that its use of the protocol has
contracted it to complete.  E.g., I get the same 100 messages
downloaded multiple times from my POP3 server, merely because
the 101st message maked OutLook barf.

Worse, it's possible to create RFC822 non-conformant messages,
e.g. messages with unbalanced quotation marks.  At least now,
with the most recent versions, you have to go into a list of
destination users in your address book to trigger this bug.


As far as superior technology, Microsoft does what it has to
to meet the bar, without raising it in excess of the minimum
required amount.  USL has historically done this as well, and
it appears to be symptomatic of large marketing-driven (_not_
"market driven") companies.  USL went so far as to forbid their
employees from engaging in outside research and developement
unrelated to their job function, in order to ensure that the
free UNIX variants did not raise the bar too quickly.



> > For instance, half if not more of the IE
> > libraries (aka DLLs in the Windows world) are loaded because they are
> > part of the base OS of MS.  That makes IE `seem' to load faster than
> > other Web browsers, because a large part of it is already loaded.  This
> > is not superiority, it's plain good ol' cheating.
> 
> Cheating is a great way to make a product better. All the user cares about
> is the user experience.

The issue here is that the DLL's that are shared with Windows have
proprietary, undocumented interfaces.

As far as user experience, Netscape could load much faster by
lazy-binding their plug-in loading, among other things.  The
appearance of speed has much more effect than actual speed, as
any of us who have written DOS programs that use a screen memory
page flip in order to achieve an "instantaneous load" can attest.

One real problem that I haven't seen pointed out is that their
virtual memory manager favors Microsoft components.  For example,
VFAT32.VXD and other Microsoft components grab large chunks of
real memory, and then only give pages back one at a time, when
signalled by the OS that a low memory condition exists, based on
allocations by third party components, yet behave differently
when signalled by Microsoft components.

If the issue is truly "the user experience", then it would be
of most benefit to the user experience to _not_ engage in
practices that result in the user experiencing poor performance
from products that they did not buy from Microsoft _solely_
because they did not buy them from Microsoft.


> If integrating IE into Windows improves the user
> experience, then that's a legitimate reason to do it.
> 
> You may consider some of Microsoft's innovations cheating, but that's not
> your decision to make.
> 
> Go ahead, take memory management out of the OS. Take disk compression out.
> How far do you want to set us back?


With respect, Microsoft did not include memory protection until
OS/2 forced it to do so.  They also intentionally swapped a byte
order to make the OS/2 LAN Manager client and server incompatible
with Microsoft clients and servers (see the SAMBA source code for
details on this OS/2 variation for use of SAMBA with OS/2).

Microsoft has done more than merely engaging in predatory
practices, wielding monopolistic power in the marketplace, in my
opinion they have also violated the RICO anti-racketeering statutes.


> > And more examples
> > like this one can be found at closer inspection, like those rumours that
> > non-MS products are offered the great honour of a few extra wait-states
> > by the scheduler of the OS in question, which is another way of making
> > all the _others_ look like they're tooo slooow when compared to ma' MS's
> > finely intergrated products, etc. etc.
> 
> Look, it's Microsoft's operating system. If they didn't want to sell it at
> all, no one could force them to. If you like it, use it. If not, don't. But
> do you really want the government telling Microsoft how to design its
> products?

No, of course not.  I just want the government to "bitch-slap"
them when they are found in violation of the Sherman Anti-trust
Acts and the RICO statutes, just as I was happy for the government
to "bitch-slap" AT&T for owning everything from the copper wire
ducting machines to the production of the chairs used by telephone
operators.


> A company should use every resource at its command to provide the best
> products possible to its customers and the most competition. Technological
> leverage is the bedrock of pretty much every high tech company. They all ask
> "how can we take the technology, expertise, and knowledge we have and use it
> to make new and better products". If that is a crime, then innovation is
> over.

I would change this to read "...evey _legal_ resource...".


> > > Harmful to consumers? Please -- show me any evidence of monopoly harm
> > > to consumers. (Do you know what monopoly harm is? Or am I wasting my
> > > breath?)  Show me reduced output. Show me higher prices. Show me
> > > reduced quality.

The Monopoly Harm includes:

o	OS purchases subsidized IE developement, such that users
	had to pay for IE, even if they installed Netscape.

o	The "Anti-DR-DOS" code test and warning dialog in
	Windows 3.x.

o	Anticompetitive non-uniform pricing across PC vendors
	allowed Microsoft to effectively control the PC pricing
	model for vendors.

o	Products that are only as good as they have to be to
	stifle the commpetition, while at the same time leaving
	room for retroactive bar-raising.

o	Etc. (I could go on for many pages, without even resorting
	the the "findings of fact").


> > Compare the price of running Netscape on *BSD, with that of running IE
> > on Windows.  Both browsers are free, but with IE you find yourself in a
> > lack of choises.  You absolutely _must_ use Windows to have IE running,
> > even if you would prefer to run *BSD as your primary desktop OS.
> 
> Neither Netscape nor FreeBSD are commercial products. I'm talking about
> comparing Windows to other commercial operating systems.

BSDI is a commercial OS.  DEC UNIX is a commercial OS.  OpenVMS
is a commercial OS.  OS/360 is a commercial OS.

Where is IE for these systems?


> Do you think you could start a company, build an proprietary operating
> system from the ground up, and sell it for prices that compete with
> Microsoft? Of course not. Microsoft can only do so because its enormous
> volume gives it tremendous economies of scale. That makes it nearly
> impossibe to compete with Microsoft.

DR-DOS was sold by Novell for $6 a copy.  Microsoft was selling
DOS for ~$50 a copy, the so-called "Microsoft tax" on PC hardware,
whether or not it ran a Microsoft OS.  This was also found to be
actionable under the Sherman Antitrust Acts, and Microsoft signed
a consent decree that held them accountable.  One of the causes of
action of the current case is their failure to comply with the
previous consent decree.

Microsoft controls the availability of their applications for
other platforms.  A lawsuit by Apple, settled by an investment
of cash and porting of Office to the Macintosh platform, proved
this.


> 	This is how competition is _supposed_ to work.

No, it's not.


> > Monopoly harm begins when you start to get your choises limited, and the
> > choise of one's operating system is IMHO a very fundamental one.
> 
> But you have the choice of numerous operating systems. Most people choose
> Windows simply because they find it superior for the tasks they need to
> solve. Build a better operating system, and nothing Bill Gates can do will
> stop you from ultimately triumphing in the market. But Microsoft is powerful
> and is a fiendish competitor. This is a successful market.

Actually, the Microsoft stranglehold on the application space,
including tools designed to cause people to use non-portable
coding practices and closed technologies (e.g. DCOM) when open
standards exist (e.g. CORBA), is enough to cause such a company
to fail.

BeOS comes close to being useful, though its total lack of the
idea of credentials other than the logged in user/password tupple
unfortunately makes it unsuitable as a server OS.

Where is "Office 2000 for BeOS"?

> > > And in any event, killing the competition is what companies are
> > > supposed to do. Our antritrust laws exist to _ensure_ the most
> > > vigorous competition possible.
> >
> > Forgive me if I am wrong, but companies are not supposed to "kill"
> > competition.
> 
> 	Yes, they are. They are supposed to compete so effectively that their
> competitors have to provide better products at lower prices or go out of
> business. That's a strong, competitive economy.

Actually, "killing the competition" is the tactic of a coward
who chooses to fight on ground that has nothing to do with
eventual customer benefit.

Microsoft competes in the domain of "copy plus one", minimally
raising the technological bar, while at the same time preventing
others from being able to gain sufficient market share that they
can raise the bar.  Every Microsoft product that has followed
this strategy (e.g. WordPerfect -> Word, Lotus 1-2-3 -> Excel,
etc.) has succeeded, and any that have not (MS Money -> Quicken)
have failed.

When they can not control market share through normal channels,
they "give the product away with the OS", code for "add to the
OS cost and amortize develeopement costs for otherwise poorer
products".  Witness IE, Outlook Express, etc..


> There is nothing a company can do to stop consumers from buying a better
> product at a lower price.

I disagree.  And so does the Judge.


> The hard part is producing a better product for a lower price.

Microsoft did not become the largest company in the world,
money-wise, by having thin margins.


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199911152040.NAA22265>