Free drivers aren't enough

I bought it specificly to be used with my PPC machine. I ended up using the Ural-linux ported freebsd driver.

(of course I like to do lots of research before I buy hardware so I expect this issue)

The rt2x00 driver is great and all that, but it's definately not ready yet for kernel inclusion. It's being used as the forefront of developing the Devicescape 802.11 protocol stack.

When the devicescape stack gets into the kernel, so will the ralink driver.

As for why the Broadcom driver is in the kernel before the Ralink?
Because rt2x00 developers made mistakes with their development model. They went through several 'beta' re-writes trying different things before settling on current d-scape stack.

Although the way the wireless stuff was handled by the kernel developers didn't help at all.

But the bigger reason probably is because the Broadcom hardware is much much much more common as being sold with most laptops and other devices. Getting the broadcom driver into the kernel probably more then doubled or tripled the amount of different 802.11g devices that Linux supported in one big swoop.

I imagine getting it so all those ndiswrapper users to wean themselves off of depending on crappy windows drivers was a pretty high priority. Combine that with nvidia propriatory drivers people are starting to get the notion that open source model doesn't work for hardware (which itself is laughable idea.. but the impression is often has much more impact then actual reality)

Why the Intel Video drivers are significant?

BECAUSE IT'S THE BEST WE GOT.
That's all. Intel isn't a saint, they aren't going to save the world. They are mearly doing what they should be doing all along.

But it's still the best we got. And it's MUCH MUCH better then binary drivers, I don't care what anybody says. Even though the binary drivers can be reverse engineered having open source drivers from the get-go saves people soooooo much work and time.

They are open source. They are Free software drivers. They fuffill the basic requirements and even though documentation is great to have this is what we have to work with.

Realy, it's fundamental. Do you want to buy a ATI card or a Intel card?

One company provides code, the other one doesn't. There is no other company. There is not a single video card you can go out and buy for a modern machine that has documentation for their devices. Not a single one.

You cannot not buy a video device. Your desktop is worthless without one. Your money is going to go to Intel or Nvidia or ATI. You don't have a choice. So you take the best you can get. That's all.

That said, if it is a choice between a binary driver and a free software driver I'd choose the free software driver -- at least it reveals one way of using the hardware. This might be enough to guess how the device works.

Personally, I think source code is some of the best documentation you can have.

Yeah, especially since any halfway responsible coder of a driver intended to be open source for a device with no manual would put all the needed information in comments.

I've always thought what's worth even more than specs is technical support at the driver level. E.g. you can email the company with, "I wrote Q to Register Z and nothing happened," and someone responds, "We found a bug in that version. To work around it, write Q twice."

In contrast, all the technical support you can usually get is, "what version of Windows is that, and which button did you click?"

Personnally I think source code is some of the best documentation you can have. This is especially true when the code was written by someone who knew the hardware (as apposed to being reverse engineered).

Your argument seems to be that if there is a bug in the code, you can't fix it if you don't have documentation for the hardware. The problem is this logic assumes that your documentation has no bugs.