xml-dev - RE: Less Stupid XML

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

I agree completely but as XLANG Orchestration will out, 
simplicity can be the layer over very complex but 
coherent processes.  It takes imagination and 
grit, but without imagination, grit gets the same 
thing and that won't get it done.

If I worked for Microsoft or any other company 
not doing the life-numbing things we do here, 
I'd design interfaces made to humble a 
P500 so the user would want a P1000. I would:

o  Push the 2D interface designers to the 
bottom of the stack and hire skilled animators.

o  Push the "Click here:" box designers to 
the bottom of the stack and replace them with 
script writers and some sound shape designers 
who know how to work with ADSR envelopes.

Then integrate the presentation of the process 
inside 3D characters who act.  This isn't that 
far out.  You have most of what you need in 
X3D, VoiceXML, and maybe SMIL (I say maybe 
because the syncing isn't all that good).  
You see, Mr Costello, a video game isn't what 
I had in mind.  I see a rockin' revolution 
that changes the entire face of the web by 
giving them a much easier interface.   If 
Dvorak wants to play, give him Claudia.

The key, as always, is to fuse the information 
already there into a coherent visualization 
that is easy to use.  Why did want XML?  We 
don't want to write a new parser for all of the 
different information we need to fuse.  

Let's use a simple example we can scope 
quickly: MP3.Com.  Artists create playlists 
for stations and their own site.  Let's 
say that MP3.COM keeps the servers for the 
music, but allows a "federated page" to 
be created that takes advantage of several 
services.  One of these can be a DJ service 
that specializes in:

o  Design of 3D talking heads with extensible 
behaviors; for 3D com, digitalWolfmanJack or 
for Dvorak, ClaudiaWithCurves and a nice soprano.

o  Sound shaping for the VoiceXML grammars (VoiceXML 
may need work for this; I don't know what is in 
the grammar yet, but emotive markup is required that 
is independent of the particular character).

When the user gets the federated page, it sends a 
SOAP message to MP3.COM to return the comments that 
artists enter.  These are used to feed VoiceXML.  
They may be dressed up a little, or a savvy artist 
enters the emotive markup.  Because a surfer may 
be getting songs from several genre, different 
characters have to be able to use the same 
markup for say, <phrase tone="scarcasm' /> 
or whatever.   Of course, I am depending on a 
voice generator with a lot of oompah, or prerecorded 
snippets in which case, the emotive markup has 
to select a snippet.  How much interaction we 
do depends on how good we are at using props 
and scripting dialog (spoken, not square with 
beveled edges).

Complicated to script, sort of.  Certainly easier 
with XML and a good editor.  Impossible or unlikely? 
Well, in 1991, we were told that using SGML to 
define and drive enterprise processes was lunacy.  
In 1995 we were told that using information containers 
to sequence processes was bad design.  Here it is 
2000 and damm, all of the new markup languages look 
like drum sequencers with crummy interfaces.  Listening 
to them is a fine way to find innovations in someone 
elses invention.  BTW: the XLANG guys should take a hard 
look at how midi system designers do it.  Same gig; 
different contents in the macros.

You see, I'm not all that excited about a semantic 
web.  It is a "let the machine think for me" wetdream.  
You can do nice discovery stuff with arcs, but the 
game is smart interactivity.  The game is an interface 
that is seamless and mouseless.  The semantic stuff 
can be done but it isn't that exciting.  Interacting 
with the network transparently is, and a receptionist is a 
better index.mimeThing than a top level form.

The excitement of milllenium hypermedia is just about to begin. 
The Dvorak's can bare their teeth or wait until we 
can hide complex production behind transparent designs. 

Then we rock this house.

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Joshua Allen [mailto:joshuaa@microsoft.com]
Sent: Thursday, October 05, 2000 2:41 PM
To: Bullard, Claude L (Len); xml-dev@lists.xml.org
Subject: RE: Less Stupid XML


Maybe two others:

* Well-defined and limited scope.

* Simplicity -- I repeat smplicity because I do not think it has to be all
that subjective.  Things like XML exist for developers to use.