[openbeos] Re: openbeos Digest V2 #4
- From: philippe.houdoin@xxxxxxx
- To: openbeos@xxxxxxxxxxxxx
- Date: Fri, 04 Jan 2002 15:00:32 +0100 (MET)
> >Michael Phipps wrote:
>
> There are two sides to the coin. How to make an OBOS/BeOS/NewOS driver
> and how to talk to hardware device foo.
> Obviously, we can teach/document how to make a driver from our side.
> How to talk to the device? That needs to come from
> 1) the maker of the device (best case)
> 2) reverse engineered from an open-source driver
> 3) from the maker of the "main" chip - for example, the bt848 video
> sampling cards
> 4) reverse engineered from other sources
Well, I would re-order like this:
1) from the maker of the chip(s) (best case)
2) from the maker of the device (but he often rely on openess-mind of the
chip(s) maker... unless he's the chip maker himself, like nVidia, ATI, Matrox,
Adaptec...)
3) from the *consortium* who wrote the standard hardware API, IF ANY !
Examples: USB controllers (UHCI, OHCI, EHCI aka USB 2.0), i1394/FireWire
controllers (OHCI), IDE controllers, etc)
4) reverse engineered from an open-source driver :-(
5) reverse engineered from other sources :-( too, but even worse
Mostly, chip makers who are technology leaders (nVidia, ATI, WebCAM chipsets
like Philips...) don't release datasheets until they have a better products
line to offer. Sad, but true.
Philippe.
Other related posts:
