> At 10:51 26.01.98 +0000, you wrote:
> >This leads to another question: how would one generate reasonably
> >accurate programming pulse times? EPROMs need pulse resolution
> >around a millisecond. From DOS, I could read the timer hardware,
> >but on Win95 I'd expect a message about an I/O access without
> >permission. As everything gets more virtual, such access becomes
> >less feasible. And I can't see OS writers providing facilities
> >for such fine-scale timing.
>
> I don't know if this is what you asked for, but you can do this from a DOS
> window in win95. Just clear the interrupt enable flag, and you have full
> control over your computer...

Would that it were that easy.  If you clear the interrupt-enable flag, your
software will THINK that it has full control over the computer.  And in
many cases Windows will refrain from doing anything to alter that impress-
ion.  Regardless of the interrupt flag, however, I/O operations WILL be
trapped and virtualized; the virtual I/O handler will usually just let the
I/O happen once it figures out that's what it should do, but there is no
guarantee that it will do so in any specific amount of time.

I would suggest that under Windows the only safe way to program anything
other than a 16C84/16F84 (which are very forgiving on their timing req'ts)
is to use an external PIC to generate the timings.  Even with a 16x84 an
external PIC wouldn't hurt.