[Sam Hunting]
| Maybe a little explanation of what an ilink is and how a web (?) of them
| might be set up would be of benefit to the list. TIA.
OK, Sam. But I have to do it in SGML/HyTime terms which may
not be understood in this forum. I shall post it separately
so folks can *kill* the post if they aren't interested.
[Maribeth Back]
| Just saw your last post, Len. Does MIDI answer your criteria?
Yes. MIDI can be pretty bad on some cards, but that is SMOM
(Simply A Mattter of Money). OTH, there is so much MIDI and it
is a lingua franca for sound designers as well as an inexpensive
sound source. Umm, .wav is also an alternative particularly for speech.
[Maribeth Back]
| I find this notion of semantic relationships utterly fascinating and would like
| to follow up on it. I'm wondering which Minsky work you're referring to? And if
| you feel like elaborating -- to me or to the list -- I for one would be delighted.
Let me post to the subject with the ilink discussion.
[Mark Waks]
| Something to consider, for those who have their heads into the URI
| stuff: are our assumptions different from those that have come
| before? The HTML side of the Web, which formed most of the assumptions
| underlying URI's to date, has a subtle assumption that "standard"
| objects are rare.
The HTML linktype is limited but for its initial environment, that wasn't
too bad. It gets work done. For a more robust set of linktypes, consider
ISO 10744 HyTime. Eight years of work went into it and the current ISO
WG8 work on it will produce a better and clearer definition. (Should be ready
about midsummer). Some of the "generic object" stuff will be there.
[Anthony Parisi]
| You raise some good points. This is why I would rather go for a simple
| AmbientSound idea. This gives us basic functionality without a heavy spec
| or development commitment. Why ram sound through in a few weeks when we
| gave 3D geometry the better part of a year?
Thank you. The knot in my stomach just relaxed. You, Nadeau, Marrin, and
Behlendorf have a good grasp of the complexity of this. The best suggestion
I've seen so far is a simple pointer out to a notation. OTH, when I post the
answer to Maribeth and Sam, I'll list the solutions we worked out for the
US Navy MID project for behaviors/scripts and external notation reference
and locations. Interested parties can look at it to see what can be culled.
[Gavin Bell]
| What, exactly, do you propose we do for sound? If you're going to complain, at
| least suggest something better...
Complain? Me? ;-) The other posts are excellent descriptions of the agony
and ecstasy of creating a sound notation. Reference one for now and consider
with the luxury of time the 3-D local values that must be passed to the notation engine.
[Brian Behlendorf]
| I don't see why the scene description language and the media types
| of inlined data aren't completely orthogonal -a binding might have to be
| described, as Tony has proposed for sound... think orthogonally.
Yes.
[Dave Nadeau]
| I recommend we *not* ram any part of sound through right now. Sound
| synthesis and processing is at least as difficult and rich a field as is
| graphics, and deserves the same serious careful consideration we have
| lavished on geometry descriptions. Remember that once you add something
| to a spec, you can almost never remove it. If we make a mistake now, we
| have to live with it for many years to come.
you betcha. that "live with it" sure describes my experiences in the CALS
arena. Something like carpentry: measure twice cut once.
[Mark Waks]
| My opinion, having looked over a bunch of alternatives, is that the
| best option is *probably* creating a hybrid: a simple language that
| also acts as an API.
Yes. But, the problem of needing *all* of a language is quite real
and this is a comment I've heard on several projects when scripting
was required. Also, creating the language "sui generis" as they say,
creates problems. For MID, we used a subset of C and HyTime.
It works for the application we designed.
[Robert Weideman]
| The best *protection* that VRML can get would be to fold it into a non-business
| structure, like W3O, ASAP. We support that approach strongly and will
| contribute resources to putting into the public domain VRML 2.0 that
| we ALL agree on.
Yes. Consider ISO as well. There are rumors that the IETF and ISO have
discussed alliances to their mutual benefit. Yes, I know that ISO carries
considerable *overhead* and *control* issues, and if you don't get in the
right group, you will struggle politically more than technically as well as having
other *domain* issues thrust upon you. The benefit is that ISO is an experienced
international standards organization with policies and infrastructures that
are useful. I've worked with ISO, and if you get the right combination
of people and companies, the results are good. Since you post that your
company has experience in ISO, you understand all of this. W3O probably
does good work. I've never been to a W3O meeting, but, there were horrible
mistakes made in the way the WWW came into being (IMNSO), and these
can be prevented if ISO, the W3O and the IETF align. International resources
should be controlled by international groups. Yes, it always comes down
to the *dedicated few*, or that is my experience in SGML, but if the policies
are defined well, and you have confidence in your skills, that is not an obstacle.
[Anthony Parisi]
| FILE FORMAT/DEFAULTS
| WWWAnchor {
| name "" # SFString
| description "" # SFString
| map NONE # SFEnum
| }
|
| What do people think?
Yes. Simple, effective, and common practice.
Len Bullard