>> We all know WinXxxx's don't work under Linux. (or OS/2, etc.)
>> We all also know that WinXxxx's sap the main CPU, thus are Evil.
>> But WinXxxx's are a popular market option, and to some extent a
>> barrier to Linux (or OS/2, etc) adoption.
> Frankly, these Winmodems and Winprinters are a scam by the hardware
> manufacturers to save a few bucks on their designs by passing some
> functions off to the CPU. I don't see why you would want to subsidize
> poor hardware design and encourage that kind of OS dependent thinking.
I agree, but they're probably more of a statement that a lot of people
are buying big CPU "because it's the newest and neatest", and then
barely scratch the surface of what it can do. In that case the WinXxxx
is tapping underutilized resource.
Notice that I have no plans to EVER buy such a beast, myself. I run
every CPU I've ever had into the ground in short order, and want more.
Quote:> Are the many hours needed to reverse engineer those toys worth the
> savings? Is your time worth that little?
I wasn't asking the question from a reverse-engineering standpoint,
more from questioning whether there were any disciplined or documented
interfaces there. I guess the WinModem software has proprietary metal
at the bottom and the Windows comm.drv interface at the top. Then it
sounds like the WinPrinters have proprietary metal at the bottom, and
GDI at the top.
Given the proprietary nature of top and bottom interfaces, I'm
surprised these things work as well as they do. Graphics drivers
seem to be a real bear. Actually, I guess the comm.drv and GDI
interfaces are published, but I would think normally only from
the top, not coming up from the metal below.
(NOT speaking for IBM)