Yes,a DOS I/O program does run in Win2k, but as you said it goes thru' a number of internal Win OS layers, which thoroughly messes up the timing

I once tried to use the parallel port to generate a range of frequencies with millisecond resolution, which ran agreeably fine even on Win '98. But in Win ME (which as far as hardware access is concerned is the same case I think)I got signals that were totally,as u put it, wierd :D

what really confused me was that the frequency generated wasn't less than but greater than what I had programmed... i.e. it was toggling the parallel port O/P MORE often than it was programmed to! I also noticed that if no programs were running the frequency meter still registered a count on the parallel port.

Does that mean Windows is always messing around with the paralel port even when it has no reason to?

I also noticed that even in Windows '95 the C BIOSCOM function for serial communication using RS-232 does not work.

kundi