xml-dev - IAML v1.0 alpha

[da - dan ancona <da@vizbang.com>]

hi there,

after a somewhat long hiatus away from this community and from vrml, i'm
back.  i'm another person working on 3d data representation, and i've
been in serious larval/hermit mode for the past six months or so.  my
focus is on 2d/3d integration, especially with regards to text and
linking.  i've come up with a syntax that (as simply as possible, i
hope) associates data (for now, text, images and objects) with spatial
information, and allows some preliminary kinds of links to be drawn. 
i'm calling it IAML, information architecture markup language.  think
storyspace nodes, but in a 3d space.  i've also written a good sized
chunk of a client that renders it...the client is still a bit rough
(writing 3d code has been an interesting challenge).  it's demoable now,
and i'll have screenshots up soon.

the IA of the IAML is, at this point, a bit of a stretch.  it's based on
the hypothesis that the process of information architecture is, in some
way, a process of locating information in space, whether the end user
experiences it as such or not.  i realize that this is a) altogether
unproven (but, i hope, interesting) and that b) i'm overloading the
phrase "information architecture" here a bit.  i hope the
interestingness of work along these lines precludes my hubris in doing
so.

the basic idea behind the syntax is to take the hard parts of 3d (i.e.
the scene graph) and hide as much of it as possible.  the snippet below
does render in the client i've written, and shows off a goodly chunk of
it's capabilities; however, it is not finalized and will remain in flux
as the client continues to develop.  ergo, i have not developed a dtd or
schema for it as yet.  i've also tried to take the other hard part of 3d
spaces (free navigation...not hard for the quake/doom crowd, but not
very practical in a presentation), and hide that by the use of easily
navigable visible linking structures.

see below for the kind of feedback i'm looking for...this chunk renders
the first slide in a slide show (one of the applications i have in mind
for the syntax), as well as a bit of other gadgetry...

<?xml version="1.0"?>
<scene xmlns="urn:vizbang.com:iaml">

<!-- vzPlane's are sort hierarchical basket for ideas...parts of a
presentation, chunks of an outline, 
related subgroups, etc.  they render as horizontal planes -->

<vzPlane w="40" d="40" position="0 -2 30" color="1 0.18 0.18" title="the
technology" closed="FALSE"/>
<vzPlane w="40" d="40" position="-30 -2 80" color="0.38 0.78 0.56"
title="applications" closed="FALSE"/>
<vzPlane w="40" d="40" position="30 -2 80" color="0.18 0.59 1"
title="context" closed="FALSE"/>

<!-- vzText, quite simply, associates a bit of text with a point in
space -->

<vzText position="0 0 0" title="intro" ID="1" fontName="Arial"
fontSize="18" xPos="10">Vizbang Alpha 0.001 Screenshot Demo

        16 November 02000

        Da - da@vizbang.com</vzText>

<!-- vzImg and vzShape load in images and shapes, where a shape can be
just about anything expressible
in a .x file (for now, this is microsoft's semi-proprietary format for
3d shapes), from terrain to
building like shapes to protein structures to ... -->

<vzImg position="0 0 50" src="fodder/logo128.bmp" size="8 8 0"/>
<vzShape file="fodder/teapot.x" position="-10 0 25" scale="0.9 0.9
0.9"/>

<!-- vzLink(s) draw an easily navigable link structure from one part of
the scene to another -->

<vzLink fromPos="0 0 0" toPos="-20 -2 10" startColor="0.18 0.59 1"
endColor="0.18 0.59 1" fromID="text1" toID="text2"/>
<vzLink fromPos="0 0 0" toPos="-50 -2 60" startColor="0.18 0.59 1"
endColor="0.18 0.59 1"/>
<vzLink fromPos="0 0 0" toPos="10 -2 60" startColor="0.18 0.59 1"
endColor="0.18 0.59 1"/>

</scene>

...i realize this is a little bit of a specialized weirdo syntax and
that (especially without a demo or at least some screenshots) it's a
little hard to tell what i'm chasing.  but what i'm looking for now
mainly is intersections.  i'm looking at interoperability possibilities
with the following standards and products or systems: topic maps/rdf,
xLink, VRML, OPML, and existing presentation/authoring systems like
PowerPoint.  other systems that i'd like to hook up with are blogs, and
possibly other content management systems in general.

i hope to be in a position to attend some of the upcoming conferences in
this space soon; please to forgive if i match up name tags to folks who
post about one or the other of these topics a lot and ask for a rundown
or pester you with questions regarding your work.  (i find that other
people's syntaxes usually end up making my head spin, but i'm
nonetheless fascinated by the great scope of the work going on here)

on a related note, a bit of encouragement: it seems like the overlap
between what might loosely be called "syntax people" and
"interface/usability people" is kind of small.  this could just be my
i've got a hammer so everything looks vaguely nail-ish perspective, but
it seems like few syntaxes are really developed with end users in mind. 
i hope this isn't interpreted as flamebait (please spool the thread in
the subject line if there's further interest in this), i'm really
curious as to what people think on this.  possible medium term goal:
come up with some best practices for this, for developing human facing
(as opposed to application-facing) clients?  is this even a useful
distinction to make?

the bottom line: any feedback greatly appreciated.  while the client is
still not quite ready for any kind of a release (pending possible
development of a business model, among other things), i am able to
schedule appointments for demos in the bay area now (or possibly
eleswhere with enough warning).  i'll be posting screenshots and more
info about the client as it becomes available.

also, this a bit of a tangential request for this list, but the company
that i've been developing this under the auspices of (vizbang software),
is currently post-lawyers but pre-revenue and pre-funding.  any advice
or contacts about changing the status of this (especially the
pre-revenue part, i'm hoping for a small vc or other seed roung and
quick-to-some-revenue bootstrapy kind of model) would also be greatly
appreciated.

thanks so much for your time and happy holidays to those who are so
inclined,
da