Browser Integration: Methods and Motives
By Jean-Louis Gassée
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Keeping our past newsletters on line has many advantages.
One of these is that they give readers a sense of the
history of our little company. Sometimes, sharp-eyed readers
kindly give us an opportunity to shed the light of logic, or
apology, on apparent contradictions. For instance, I wrote a
column titled "The War of Two Browsers"
[http://www.be.com/aboutbe/benewsletter/Issue38.html] on
August 28th, 1996. In it, I wrote "In the meantime, we
happen to agree with Microsoft: the browser ought to be
integrated in the system and presented to application
programmers as an API, a class library. After all, several
years ago, we didn't call our 'user shell' a Finder or a
File Manager, we called it a browser."
How does this square with our subsequent criticism of
Microsoft's integration of the Explorer Web browser in
Windows in subsequent columns: [http://www.be.com/aboutbe/
benewsletter/volume_III/Issue5.html], on February 3rd, 1999,
and [http://www.be.com/aboutbe/benewsletter/volume_III/
Issue6.html], on February 10th, 1999? Are we contradicting
ourselves? Did I even remember our August 1996 position when
I wrote my February 1999 columns? No and yes.
Let's start with two premises. First, the "border" issue,
and second, what you can no longer do once you've become a
monopoly. The "border" issue invoked here refers to the
difficulty of setting up a clean, well-lighted fence between
what we imprecisely call the operating system and the
applications. Today, it appears extremely difficult, some
say impossible, to write a usable English sentence as a test
of which is which. If the code meets condition A, it belongs
to the OS. We have trouble describing what "condition A"
would be. I don't know if the trouble lies more with the
speed of technological evolution making definitions
obsolete, or with the very protean nature of software. I
guess that it's more the latter than the former.
In any event, how does one write laws or regulations
governing borders that are imprecise at any given moment --
and that move very fast once you unfreeze the clock. This
would seem to support Microsoft's right to put whatever they
please in Windows -- a spreadsheet, a word processor
(there's a decent one already, WinPad), or a Web browser
whose capabilities would help many other system modules or
applications -- if we stick to a common- sense but imprecise
definition for these two.
Turning to the second issue: things that change once you
become a monopoly. In a market where no player dominates,
common sense and, it seems, the law both say that
exclusionary incentives can be offered. What I have in mind
is offering a retailer an incentive to distribute your brand
of yogurt and only your brand of yogurt. For instance, a
supplementary discount above the normal wholesale price.
Retailers who offer several brands of yogurt pay wholesale,
but the retailer who offers brand X exclusively gets a
supplementary rebate.
In a truly competitive market, this is legit. Other
retailers exist who might bet customers will prefer a choice
of yogurt. Smaller margins on brand X might be compensated
by more volume, or more store traffic and the opportunity to
sell fidgets, bidgets and ridgets, not just an assortment of
yogurt. Once you become a monopolist, things change.
Retailers effectively offer only one brand, brand X. In
addition to the wholesale price, the retailer gets a
substantial supplementary discount based on not offering
and/or advertising any emerging new brand. If he does, he
loses a sizeable amount of money. Sales from the emerging
brand Y cannot make up for the loss, and his competitors who
didn't budge pay a lower price for the monopoly product. The
market is locked.
This is but one example of what happens when you become a
monopoly: behavior that's acceptable when the market is
truly competitive must be rejected once a monopoly situation
sets in. With this in mind, let's consider the manner and
the motives for Microsoft's integration of the Explorer
browser into Windows. They could have done it "fair and
square," one module (or a small number of modules) somewhere
in the system, a sort of Web driver (I realize the metaphor
is crude, but the point is valid nonetheless). Today, you
can update or replace a driver and, if the driver works as
intended or better, the user benefits, applications get
better performance, the system is improved.
Even if a browser is more complicated that a dial-up module,
it features a finite list of pipes for information to come
in and go out of. One way of integrating the browser makes
it easy to disconnect and reconnect the pipes, another way
makes it deliberately difficult. Did Microsoft make the
disconnection and reconnection deliberately difficult and,
if they did, why? As a monopolist, are they free to make it
arbitrarily difficult to remove Explorer? Did they have an
anti-competitive motive for doing so? Did they leverage
their monopoly? Did they fear that another Web browser might
prevent them from gaining control of the HTML dialect mostly
spoken on the Web?
In our case, we found we could offer a browser and make it
easily removable. In other words our browser is integrated
in the sense that it's "in the box" when you install the
BeOS, it's in the system. But if, in the future, you want to
install a competitive browser, it's a simple drag and drop
move. Integration doesn't have to be anti-competitive.
|
