May 26, 1997
An Intel-igent proposal for the Wintel
By Arthur Tisi
A recent project that I worked on brought out
everything that is wrong with the Intel-Windows
platform. It showed me that the openness of the
Wintel platform is a myth and leads me to
propose what some might think a far-fetched
but I think a reasonable solution: that Intel and
Microsoft develop a "nonopen" version of their
platform that would provide seamless
interoperability between applications and its
processors. It might be something like Windows
CE, except not for mobile computing.
A few weeks back, I needed to prepare an extensive presentation
incorporating multimedia, screen-capture and audio in a short time. I
selected Windows NT and PowerPoint, thinking that it was the most
robust and simple approach. Many of my friends in the Apple
contingent suggested that if I really wanted it to work, I should have
used a Macintosh. But I felt comfortable with the Intel-Microsoft
approach, which I had adopted many years ago. So off I went.
The project necessitated putting some CD-ROMs on the computer and
capturing graphics and video. We decided that the most
straightforward approach was to use Microsoft Camcorder. But it didn't
work on Windows NT. So, with the help of Microsoft (which, in all of
this, was extremely responsive), I ran the application on Windows 95,
losing the benefit of a dual-processor 200MHz Pentium Pro that
worked amazingly well under Windows NT but, under Windows 95,
lost half of its speed.
There were other complications. Microsoft Camcorder uses
nonstandard AVI files, although I'm told there's a patch you can
download. And stock AVI files are a special order from CompUSA.
The current notion of plug-and-play is nothing more than a notion.
There's a lot more work to do. The only way for it to work is for Intel to
put it on a chip. It would be cheap; it would be fast; it would work. No, it
would not be the latest and greatest. No, it would not be rich with
options and functionality. Intel might be able to achieve what Apple
has been doing for years, but with the advantage of processor ubiquity
and more horsepower.
Who would object? It is fair to say that both corporate and consumer
customers "just want it to work." Also, both business users and
consumers don't care about computer software vs. computer
hardware; they are solely interested (and rightly so) in a robust and
reliable computing platform. This idea may seem farfetched, but
conceptually, it's just an extension of the NetPC--a desktop you don't
have to mess with.
And Intel has already developed a model for this in its
videoconferencing, which, although rather rudimentary, does provide
the sorts of things to let grandmother and grandfather see their
grandkids across the country. If we extended this metaphor, we might
have something. Third-party software companies could take part, but
they'd have to adopt standards published by Intel.
Intel has achieved critical mass in terms of the reliability of its three
vectors of performance: office productivity, floating-point and
multimedia. What better way to combine these vectors than in truly
efficient, user-friendly and robust applications tied to hardware?
Yes, this screams monopoly. Yes, it begs the question of whether or
not the Justice Department might be involved at some point. But if the
goal is ubiquity and ease of use, then this solution is as good as any.
Or perhaps I'm simply re-creating what already exists. Perhaps, for all
the talk, Apple truly is the way to go when you're in a pinch.
There is no off position on the genius switch. ...David Letterman
<> email@example.com <>