Subject: Re: Embrace and extend [was: Returns to service professionals]
From: "Stephen J. Turnbull" <>
Date: Wed, 7 Jul 1999 13:52:15 +0900 (JST)

>>>>> "Crispin" == Crispin Cowan <> writes:

    Crispin> "Stephen J. Turnbull" wrote:

    Crispin> You're ignoring the explicit
    Crispin> "embrace and extend" strategy that MS uses: take a
    Crispin> standard protocol, use it, and "extend" it with
    Crispin> proprietary features.

    >> No, I am not ignoring this.  I just don't think it matters
    >> much; otherwise other companies would use the strategy
    >> successfully, too.

    Crispin> But this strategy can only be successfully employed if
    Crispin> you are a market leader with a substantial majority
    Crispin> position.

In other words, it's marginal; you have to have the preexisting
dominance.  That's what I said.

    >> I simply do not believe they can get away with it indefinitely
    >> on that basis alone.

    Crispin> Certainly not on that basis alone.  It only works in
    Crispin> combination with other factors.  It is a strategy for
    Crispin> maintaining market dominance; it blows chunks for gaining
    Crispin> share in a fragmented market.

So what's your point?  All I can see you doing is echoing my points,
and _asserting_ that that proves my position is incorrect.

    >> This strategy is a Red Queen's race.  You must maintain
    >> technological leadership[1] or the fact that your product emits
    >> broken protocol will break _you_.

    Crispin> I think this is incorrect: you must maintain a dominant
    Crispin> market share, by whatever means, or else emitting broken
    Crispin> protocls will break you.  Microsoft has shown that
    Crispin> technological leadership is irrelevant :-)

Sigh.  Is shouting in footnotes allowed?

MSFT may not be at the bleeding or cutting edges, but clearly MSFT
technology is sufficient to the jobs that the market wants done,
whether they build, beg, bluster, buy, or borrow the technology.  In
the end others copy Microsoft (for the last few years).  Microsoft
leads the industry as a figurehead leads a ship (thanx, Robert
Townsend, _Up the Organization_).

If Microsoft fails in either of those factors, that market share will
slip, and fast.

If you want to post preferred definitions of "technology" or
"leadership," please do and I'll use those, and come up with other
terms for the concepts relevant to my argument.

    >> However, far more important than "improved" protocols, as far
    >> as I can see, is the ever-increasing integration of MS
    >> products, leveraging overwhelming dominance in one market into
    >> dominance in others.

    Crispin> That's clearly what all the shouting is about in the DOJ
    Crispin> case.  However, I agree with other posters here that the
    Crispin> file format wars may yet prove to be even more important
    Crispin> than the integration battle.

And will similarly be over fairly shortly (in human years, if not in
Web years).  Can't prove that, but I don't see any particular argument
for it giving MSFT a protracted advantage, either.

Even if consumers don't know or care about broken protocols, Word,
Excel, Outlook Express, and the like _don't matter_.  It's app-to-app
communication that is going to matter, and there open standards are
going to win hands down.  Eric Raymond points out (at every chance)
that 90% of the programming jobs are for software that will never be
sold in the market.  Sure, if MSFT does a bang-up job of making sure
that OLE always works as advertised (raaaaight), then you can live in
Visual Blather and depend on the library routines for data
interchange.  But if not, you want an open protocol with multiple
vendors, preferably with open source, so that you can fix the
inevitable upstream bugs.

Those application programmers are not going to standardize on Word
HTML, no matter how attractive the presentation in Word.  Well, some
of them will; the ones who sell form over substance anyway.  But with
current browser technology it's hard to see how massive improvements
in presentation technology can be made, so interapp interchange is
where the action will be.  And there you don't want to depend on
someone else's proprietary library routines if you can avoid it.  (You 
may use them, of course, for efficiency in development or use; but you 
want to have alternatives available so you can work around.)

If you can do 95% of whatever in conformant HTML/XML/SGML/CORBA/etc,
with some extensions that allow you to do the rest, then proprietary
formats are going to lose market share in those applications, unless
they can keep up with what the apps need.

    Crispin> The writing is on the wall for the win32 API.  The file
    Crispin> format wars are just warming up.

Don't bother buying asbestos clothing, and don't expect them to melt
ice.  IMHO, of course.

University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
What are those two straight lines for?  "Free software rules."