> On Wed, Apr 02, 2003 at 02:17:03AM -0500, rm wrote: > > i looked around a bit and it seems like AMDTP/IEC61883-6 is starting > > to be used and is in the linux 1394 drivers. it appears to include > > multi-channel audio and midi which would be quite cool. (it also > > appears not to be mlan). > > actually it sort of is mlan.
yep... > > from what i've read, mlan seems to deal with finding and connecting > resources rather than actually pushing packets. so mlan lives on top > of AMDTP. yep...completely on top... > > and while mlan connectivity would be nice from other perspectives, i > would just as soon use a non-compatible (open) method as any. I can understand this desire, but I have to ask what you want to connect? If it is Linux PCs and little boxy things developed in the future using Linux, then I think we would be well served by having a separate, new protocol in Open Source. However, if your desire is to hook together studio instruments and devices that exist today, then you need to implement mLAN as it is specified. We have recently gone through a review of the mLAN Licensing Agreements, and one of the questions I asked our lawyer was whether a company could legally direct the development of an Open Source version of an mLAN implementation? His answer was 'yes'. If there were developers operating under NDA with a company that had access to an mLAN spec, or developers that had access to the spec itself directly from Yamaha, then he did not believe that the current licensing agreement required the actual Linux source code that implemented mLAN to be confidential. However, the licensing agreements do call on the developer/company that does this work to properly implement the spec with no changes, and possibly to submit samples to Yamaha for testing and verification, and should Yamaha find problems, to correct them, or remove the 'product' from the marketplace immediately. So, I think for now that there is no real need to have a separate, new protocol. However, owing to the nature of Open Source, I am quite concerned that if there was a problem code source, we would not be able to 'remove it' from the marketplace. I also note that the general nature of Open source developers seems to be to do what they want to do, and not necessarily what they are required to do as per a spec like this. I think that could be a problem. Probably the best way to make this happen would be for an individual developer who is interested in this area to take a license with Yamaha himself, do the implementation, submit it for testing BEFORE it becomes Open Source, get it up to par, and then release it. At least that's my best idea right now. > > rob > > PS nothing seems terribly novel (irreplaceable) in the patents i've > seen (viz, 5,825,752). > PPS if i see the word 'plurality' again i'm going to swear repeatedly. I agree! It's a horrible word these days...