On 04-28-97, CHRIS HOLTEN said to BOB DAVIS:
CH>NT defaults to polling, I would imagine that any printer and drivers,
CH>postscript, bidirectional, whatever, that work in NT could not be trying
CH>to use direct interupt printing as NT doesn't allow direct hardware access
CH>of the printer port. Not sure about ECP ports, but those are a superset of
CH>the original PC parallel port architecture in that not only do they have a
CH>hardware interupt, but they also use DMA. I don't know how NT handles ECP
CH>ports if it even does treat them differently from the standard IBM PC
CH>architecture paralell port. --- Maximus/NT 3.01b1
I've had problems with my HP5L if the sound card is set to IRQ7 with
bidirectional support enabled in the printer driver.  That doesn't 
ecessarily
imply that bidirectionality forces use of the interrupt but it appears to do
so in this case.
My empirical research indicates that the driver forces the use of an 
nterrupt
if the hardware and driver are so configured.  The original assertation by
Mark Wright that Postscript (and, later, PCL) are what causes the use of an
interrupt is flawed...
---
*Durango b200 #NR* DurangoMail for Windows NT/95
 * KMail 3.10o Knight Moves
--- QScan/PCB v1.17b / 01-0406
---------------