Re: Re:Driver Programming Fundamentals/Philosophy, was: Re: Calling NdisRequest() from ProtoclBi

Microsoft could have required that all drivers run in ring 1 instead of ring
0, but that paradigm would not have worked for some of the other processors
they were supporting. The CPU design is limited, the OS design is limited,
and the driver model is limited - with some stacks having more limitations
than other stacks. We got Windows and until Steve Jobs takes over
Microsoft, legacy support will continue. Good, bad, or just adequate, it is
what it is and we have little choice. Ask Compuware - Vireo - Numega about
VtoolsD, DriverStudio, and SoftIce. They were there for a while, but have
all died. KMDF, UMDF, and all the various old and new ‘miniports’ are
Microsoft’s attempts to design a better system for one objective or another.
Just as open source has limitations, so does closed source. At least
Compuware’s frameworks included source so you could find the problems, but
if you pay that much for it why should you have to find the problems?

Maybe there is a lot of truth to the saying: “Life sucks, and then you
die”. Being a device driver writer could be having a ‘life’ of some sorts.
There is the fact that because it is so hard to do this, it does provide job
security to some degree.

“cristalink” wrote in message news:xxxxx@ntdev…
> “Martin O’Brien” wrote in message
> news:xxxxx@ntdev…
> pact.
>> Microsoft contents that most system crashes are 3rd party drivers, and
>> all
>> evidence is this is true.
> I am glad my TV does not crash every time they load a bad show into it.
> That’s what I call a good design.
> —