Message-Id: <9302051639.AA23459@Athena.MIT.EDU>
Date: Fri, 5 Feb 93 11:39:25 EST
From: Toshiharu Harada <harada@ceci.mit.edu>
To: jud@ceci.MIT.EDU, pbailey@ceci.MIT.EDU, aybee@ceci.MIT.EDU,
        aybee@Athena.MIT.EDU, lynne@Athena.MIT.EDU, schluss@Athena.MIT.EDU
Cc: sigmund@nora.hd.uib.no
Subject: my version of summary

Last night I had a dream.  In it, can you believe it, I was thinking
about AM2 media classes in it.  Oh well...

After the yesterday's meeting in my office, now I am understanding
things better (Jud, I finally could get your points ;-).
To make today's meeting (ah..I'm scared), I'm going to "dump"
the contents of my brain (as little as 64KB) as a short series of
e-mail.

					
				--- Toshi
---

What I had proposed in Tuesday's meeting:

abstract
	o joint services (Galatea/QuickTime/MediaServer) to AM2
	  media classes in the very minimum way
	o force AM2 application writers to specify the name of the service
	  (explicitly) and also arguments for services in *ADL* without
	  trying to make any disguise.  Those arguments
	  will be passed to services *without* any interpretation
	  by AM2 media classes

	... or ...
	o treat each services as black boxes including even arguments
	  for them
merits
	o minimum set of abstract media types should make ADL statements
	  (not definitions) more readable
	o whole media classes are loosely coupled with other modules
	  (just receive anonymous STRINGS as arguments), in particular,
	  control object doesn't have to understand (or to be hardcoded)
	  the meaning of service (Galatea/QuickTime/MediaServer) dependent
	  specification
	o adding new service is relatively easy and can be done without
	  requiring recompilation of control object (basically)
	o probably the simplest approach to implement
demerits
	o ADL will be reflected (polluted) by the names of services
	o ADL will not be portable (explanation: if two platforms support
	  both the same service, probably ADL will be portable,
	  however, there's absolutely no mechanism that bridges
	  different services, which means ADL description for
	  one service will never be used (portable) for other
	  service)

---

What was discussed last evening by Adam, Schluss, Lynne and me
(Adam, thank you adam for making a summary) --- this is similar
to what Jud proposed in Tuesday's meeting

abstract:
	o roughly speaking, insert a new layer, "format", between
	  media clip and C++ wrapper in my scheme
merits
	o now some different services can be specified as an unique
	  virtual service.  for instance, analogVideo probably can be
	  used for both Galatea and MediaServer
demerits
	o a number of different media types (formats) will be
	  increased and AM2 application writers have to be
	  differentiate FileStill from AnalogStill.

---

[ to be continued ... ]