fwd from paul v. on ZIPI

>In response to your question, I may have been talking about
>ZIPI. Some of the people behind ZIPI were Keith McMillen of
>Zeta Music / Gibson (their address back then was 2560 Ninth St.,
>Suite 212 in Berkeley), and David L. Wessel & Matthew Wright,
>both of CNMAT, (CNMAT's address back then was 1750 Arch St.,
>Berkeley). The Computer Music Journal devoted a lot of ink to
>ZIPI in their Winter 1994 issue (I still have a copy). I made a
>few phone calls approx. six years ago and was informed that ZIPI
>is dead. If I remember correctly, there was some intellectual
>property ownership dispute (don't ask me to elaborate because I
>don't remember) and ZIPI was killed in the cross-fire.
>
>It's funny that you mention this because I have been doing a
>teensy bit of research just this past week and I have the
>feeling that USB midi is starting to supplant midi as we speak.
>It's popping up more and more.
>
>Here's a message I sent one week ago to Paul Berg--the "point
>man" for the USB midi spec and his response:
>
>-----------------------------------------------
>
>At 06:11 PM 7/10/2003, Paul Vandervoort wrote:
>Hello-
>
>I am developing a musical keyboard product.
>
>I have a few questions regarding this document:
>
>http://www.usb.org/developers/devclass_docs/midi10.pdf
>
>Is this the latest version of the protocol?
>
>Will this protocol work with USB II?
>
>Is this protocol in widespread use?
>
>Thank you.
>
>Sincerely,
>Paul Vandervoort
>Daskin Keyboard Co.
>Reno, NV
>----------------------------------------------------
>
>Hello,
>
>Yes, the current version of the "Universal Serial Bus Device
>Class Definition for MIDI Devices" document is 1.0 from November
>1999, and yes, IMHO it is fully applicable to the current USB
>2.0 specification (available at
>http://www.usb.org/developers/docs/usb_20.zip).
>
>To my knowledge this specification describes the design of most
>if not all currently available USB-to-MIDI devices, but as far
>as I can tell neither Microsoft (Windows XP and 2000 and 9x
>versions) or Apple (Mac OS X and OS9) have developed standard
>drivers for devices that implement the MIDI Streaming class. In
>my experience this is related to how common the product is
>relative to the number of OS installations -- I suspect that
>USB-to-MIDI is used in only a very small share of the Windows
>and Mac installations.
>
>It would appear that regardless of your design you will have to
>develop and ship your own drivers.
>
>Still, there would very likely be an advantage in following the
>class specification as it may very well have anticipated issues
>and provided solutions or ways to avoid problems that might not
>be obvious when first considering the design of a USB MIDI
>transport. Certainly at the very least you should fully
>investigate this class specification before you decide to build
>your device in a different way.
>
>I would point out that the class specification uses USB bulk
>transfers -- these provide the highest bandwidth on an otherwise
>unused bus and scale very well from full-speed to high-speed
>designs (as per the success of USB Mass Storage in taking a
>class that was designed to use full-speed bulk transport and
>using the class drivers without change to support high-speed
>devices).
>But using bulk transfers means that you can't run multiple USB
>devices such as standard keyboard, mouse or video camera on the
>same USB host controller without impacting the MIDI throughput.
>
>Further, please note that if you do not fully implement the MIDI
>Streaming class as described in this specification your device,
>even though it offers an interface to a MIDI instrument, is NOT
>a USB Audio class device. If you chose to implement your own
>interface design you must use 0xFF as your class code. This
>will avoid significant problems in the field if any of the major
>OS providers should ever release a driver for the standard class.
>
>--best regards and good luck with your project
> Paul
>
>Paul E. Berg, USB-IF DWG Chair
>pberg@mcci.com www.mcci.com
>direct phone: +1-616-402-2469 or +1-616-850-0742
>
>----------------------------------------------------------