> Maybe for you rich California engineers. :-) :-) :-) > I live in a place where plumbers make more than Software Engineers. :-) We all do! > Plus, IMHO, support means more than buying one device and one motherboard/PCI card. :-) For host controllers, there should just tiny to no differences, the specs are quite complete so two OHCI controllers by two different vendors should look the same (except for HW bugs, but that is a different story), two UHCI controllers should look the same and two EHCI controllers should look the same. As for USB devices... you are mixing things. First you need all the "basic" stuff (bus manager and busses) in place. USB devices are later implemented as class drivers atop the USB core. True you need some sample driver to test the core, but a zip drive or a flash reader should be enough for testing control, interrupt and bulk endpoints. A simplified(uncomplete) HID driver will also do for testing. For isochronous endpoints some USB speakers will do (thought i suggest to start with the "easy" stuff, build a prototype that handles interrupt control and bulk, and later think how isochronous can be added and redo the core.) Also note that implementing all class drivers is beyond the scope of a single person (as it is testing all the oddities and annoyances of buggy USB hw)... too much work > >> b) getting support, etc > >> c) getting documentation > > > >That's also easy. EHCI is no longer classified as confidential information > >by > >Intel. Go to: http://www.intel.com/technology/usb/download/ehci-r096.pdf > >For the USB 2.0 spec itself: http://www.usb.org/developers/docs.html > > That is cool. I wonder how many other chipset makers supply that info. Compaq's OHCI spec has always been public. Intel's UHCI is public as well. manuel,