> capable of taking on the job? We can certainly pay a reasonable sum for your > efforts. The driver would not be released open source, but binaries would > likely be made available on our website. Sometype of non-disclosure contract > might need to be signed - I'm trying to get the VP to lighten up :). Well if you are going to do a binary only driver you are going to have to get your VP and lawyers to review the product and the question of it being a derivative work. The GPL prohibits derivative works that are not GPL and most copyright holders have not given permission for binary only modules that are in any way a derivative work. You must also understand that there is no binary ABI for Linux kernel modules. You'd need to build for each kernel version and update your customer did, even a critical security fix might need a recompile. A much better approach is to figure out _what_ your company actually cares about being a secret. There is a lot more to a webcam than just the interface to talking to it, and the value of your IP for the communications is probably < 1 months engineering time (because thats the engineering time probably needed to reverse it and nobody has yet...). Some vendors realise that actually their goal is truthfully "sell more cameras" rather than "hush hush this is secret". Others genuinely feel they have some "magic" advantage - The pwc driver for example doesn't support some of the clever compression features for the camera it drives because the vendor considers that bit a secret while not considering the rest a problem, presumably because they've got a clever proprietary encoding scheme which will run on a microcontroller. One or two vendors won't tell us stuff, often sad to say because it would reveal their $60 big brand name camera is a $20 .tw camera and a paint job. If I wanted to clone a webcam I can clone the other published ones, I can use the imaging device spec, I can invent a perfectly workable spec over lunch. Alan
