Polaris restaurant at the top of the
Hyatt Regency Hotel (conference headquarters)
Agenda
We need a volunteer to take minutes. First and
foremost, will someone volunteer to take minutes at this
meeting? The same person or someone else can turn the minutes
into a web page later. We hope to post minutes of all meetings
on the VRML-UI site at vag.vrml.org.
Discussion/Endorsement of Co-Chairs.Sascha Becker and Gregory Seidman
have volunteered to be the VRML-UI Working Group's Co-Chairs. For
the time being this also includes Sascha acting as Administrivia
Czar and Greg acting as Webmaster. Let's raise objections to any of
this now or officially endorse them as our Co-Chairs.
Discussion of the Charter. The main goal of this meeting
is to produce the charter. We expect a draft to be sent to the list
by or before March 28 (the day after CHI ends). There will then be
a week of discussion on the list. A final document will be ratified
on April 4, 1997. Charter specifics:
Deliverables. We must agree on what deliverables
we plan on producing. The list below is not necessarily
complete, but additional suggestions should be described at
the same level of abstraction if possible. All further
points of order concerning the charter rely on the list of
deliverables.
Theoretical Architecture of Widgets This
includes naming conventions of fields and events
and the model of the composition of a widget (such
as Pauli's tri-stratification or Greg's layer
model).
Taxonomy of Widgets This is a
taxonomy of some sort, though it may not be a
strict taxonomy, which is a tree, so much as a
DAG due to possible multiple inheritance like
the discrete/continuous in addition to the type
(SFFloat, SFRotation, etc.) of data produced
by the widget. It has been suggested that
classificiation within this "taxonomy" may be
defined by specific syntax of a widget (e.g. a
Slider widget always produces a SFFloat value from
0.0 to 1.0 on an eventOut named fraction_changed,
so anything that produces such an event from user
input is a Slider as well as anything else).
Core Widgets We are looking
for a basic, vanilla set of widget PROTOs
which will probably include at least a
Slider and a Button. They should probably
support pluggable geometry as well. Greg's widgets may be used as a starting point.
Widget Repository The repository will
include our Core Widgets, but we are concerned
with the collection, storage, and retrieval
of the widgets. In particular, the taxonomy
may come into play with searches based on the
taxonomic structure.
Feedback (to VRB and other WGs) on
specification limitations. There are widgets
we might like to have which are not possible
with the current specification. For example,
the current specification has no support for
tracking a 3D input device in more than two
dimensions. We may find limitations in the Java or
Living Worlds specifications which preclude the
use or creation of certain useful widgets. We
will make recommendations of changes to such
specifications a deliverable.
Problem statement(s). Having decided upon the
deliverables, we must compose either a problem statement for
each deliverable or a single unified problem statement which
covers all of the deliverables. At this meeting we hope to not
only decide whether to have a single problem statement or one
per deliverable, but also to compose the statement(s).
Justification(s). Similarly, we need to justify either
the entire Working Group or each deliverable. At this meeting we
hope to make this decision and compose the justification(s).
Timeline. Our timeline begins with the charter we hope
to produce by April 4 and includes the Core Widgets and/or the
Repository by SIGGRAPH '97. Each deliverable must have a deadline.