> If Microsoft has not created standards, then how can there be thousands
> of products conforming to them? How can there be a certification program
> in place for people?  How can such an industry have evolved around it?

$$$$

> How can I develop a Web Browser using this which is compatible with
virtually
> every website out there, including the MSN communities, in 2 hours?
> How can I write a program in Visual Basic that exploits a timer and
> multitasking in 30 minutes that runs on my own PC and a family of Pocket
PCs
> with different CPUs?

Just because you can use a product does not mean it is any sort of standard.
You've confusing "widely-used" with "standard". We've successfully used
Ventura Publisher to create documents for years. So have lots of other
people -- that doesn't make it any sort of standard.

> Granted, it took me a while to figure out the big picture well enough to
> write the Web Browser and the Pocket PC demo app.  But any system this
complex
> will take a while to learn.  And the documentation I used, in the MSDN
Library
> (which I view as the source of the Microsoft standards), was very clear to
> me. I am working at a higher level of abstraction when I do this
> (Visual Basic), but it's still a standard.
>
> As for your issue about "clear" and "unclear," isn't that a matter of
> background and perception?  What is clear to one person isn't
> necessarily clear to another.
> The background you need to have to understand some of these standards may
be
> extensive, but that does not detract from the standard.  It simply means
> that the standard is written for a specific audience.

You seem to miss the point, that the MSDN Library (which I agree is the best
source of information on Microsoft APIs) has huge omissions. My previous
example (that there is no description of the dynamic semantics of styles) is
just one of many omissions. A "standard" that does not contain enough
information to actually create something does not qualify as clear in my
view.

> Randy, your arguments simply don't make sense in the face of
> the industry. Just open your eyes and look around ... the Microsoft
> standards are there and working.
>
> You are welcome to try to contradict what I just said.

You forget that RRS has been using Microsoft operating systems nearly
exclusively for the entire time that there was such a thing. We were
building a compiler on 86-DOS *before* Microsoft bought it from Seattle
Computer. We've been doing Windows since 1987 (I think). By no means am I
arguing that Microsoft doesn't have a lot of users or clout. I'm even
willing to grant the "agreed-upon" part of the definition (based on all of
the customers).

But documentation that is so incomplete that the only way to figure out how
a program will act is by trial and error does not meet the definition of
"clear" by any stretch of the imagination. To Microsoft's credit, they make
it relatively easy to get all of the those OS versions to do the trial and
error (I've got 6 different Windows OSes installed here, and I'd have more
if I had more computers...)

Anyway, I don't think that there is any chance of convincing you. (I still
think there was something in the water at that conference...:-). So, I'm
going to try to drop the thread...

                        Randy.