Web Services and MS strategy

Clay Shirky clay@shirky.com
Mon, 29 Oct 2001 07:48:28 -0500 (EST)


I made this separate mail, because this is a different Holy War than
worse-is-better. 

> Any thoughts? I wonder, if this is the case, why it is. Why are
> Microsoft/IBM pushing this so hard? To sell more stuff? To provide an
> alternative to Java? To promote a charade of populism while in fact only
> they have developers skilled enough to specify/infer the context needed to
> build a successful Web Service? Other?

IBM has a pretty simple strategy right now -- bet on openess, since
that drives adoption, because even if their software business takes a
hit from stuff thats cheap or free but good, and they sell enough
hardware and enough services to come out ahead.

MS is more complicated. Half a decade or so after announcing Windows
Everywhere, Windows is still mainly on the desktop. They don't have a
majority market share, much less a monopoly, in any other device class
-- not phones, not set top boxes, not PDAs, not game consoles, not
even servers, which was supposed to be a cakewalk. They may eventually
get to majority share on at least some kinds of devices, but it won't
be quick, it won't be cheap, and even then it may look more like 60%
than 90%.

Even worse, whole new device classes keep sprouting like dandelions --
Palm, Tivo, and the Blackberry all came out of nowhere, and despite the
downturn, things like embedded Linux and 802.11b keep lowering the bar
for creating new kinds of devices.

So what they want is to triangulate -- their ideal strategy in a world
where they only have one aging monopoly is to find a way to sell
software onto all devices, with an MS OS or not, while not hampering
their drive to own those devices eventually as well. So they'd like to
let those devices use MS software _as clients_ without creating a
world where two clients can interoperate without an intermediary
server.

Both their insistence on calling SOAP-compliant devices Endpoints,
rather than anything like nodes or transceivers, coupled with their
constant emphasis on being able to maintain _huge_ server farms in all
the initial .NET MyServices (nee Hailstorm) literature, suggests that,
as with "C#: The Benefits of Java Without the Threat of WORA" and
"Shared Source: Developer Energy Without Developer Freedom", this
strategy is "Hailstorm: The Ubiquity of Interop Without Giving Up
Control of the Transactions."

Now they're smart enough to know that they can't get all of the
benefits of openess with none of the downside, so I'm pretty sure that
part of the lack of definition around their Web Services strategy is
that they are still figuring out how to triangulate -- the risk of too
much openess is visible out their left-hand window in things like the
IBM-compatible PC and the Hayes-compatible modem, and the risk of too
little is out the right window in companies like DEC.

And getting just the Goldilocks amount of third-party participation is
a big part of that triangulation.

-clay