FY00: Object Web Technologies for Education and Computing Portals

NPAC's activities fit well into the Alliance road map and in our 99 activities,
we built a prototype computing portal for nanomaterials illustrating some of the
features highlighted in the common portal architecture (CPA) and consistent
with grid architecture roadmap. We also developed the portal collaboration service.

Computing Portal Architecture and Services
In our activity in FY00, we propose to work within the emerging Alliance CPA in
the areas of basic architecture, secure object brokers, visual authoring tools
built around our WebFlow technology, modifying our current system to support
Alliance standards. We will do this both for our original nanomaterials portal and
the lead AT team chosen by the Alliance -  chemical engineering. This should lead
to operational 3 tier portals in two application areas supporting dataflow and a
general object based programming model. We will make this available with both our
JWORB and commercial CORBA object broker middle tier infrastructure. We will
develop this activity synergistically with our DoD funded work on the so called
Gateway project. 
We anticipate a substantial Alliance planning activity in all aspects of portal
development and we will participate in most areas as we have quite a bit of
relevant experience. In particular we will continue leadership of Java Grande
activity which has already a Computing Portal working group. Another major
activity identified at Oak Brook concerns XML standards for scientific data where
a working group was identified. This needs to set standards and also either
develop or recommend existing XML management tools. We have already work of this
type with NASA and DoD and will participate in these discussions.

Portal Collaboration Service
Collaboration is both a key part of the access grid and a major service identified
in the Alliance portal architecture. We will work with the Alliance to develop a
good definition of collaboration for portals viewing collaboration as implying a
general synchronous shared object facility. We will continue to develop support
for EOT projects including the Biology workbench, which is expected to be an early
example of the new Alliance portal architecture. This is likely to require
significant work as both the workbench and the underlying document object model
are expected to change significantly. This work will be done using where possible
the new IMS messaging standards for which Tango Interactive has been provisionally
selected as a reference implementation. We will deploy a version of Tango2 that
supports Microsoft explorer and the new version 5 browsers (currently only
released for Microsoft). We expect to be able to support Macintosh's with this
version of Tango. These steps imply enhancing Tango2 to share the new W3C Web
document object model.

We will continue our work on the implications of universal access for
the access grid. The goal is allow convenient specification of customized client
side rendering of a shared object. Technically systems like Tango can
share the XML or similar abstract definition of an object state and then
customize the style sheet on each client to cope with users who could be
visually impaired (and so require sonified rendering) or have just a low
bandwidth line (and so require the low resolution version of an image). 
We can produce a modest illustration of this but we will only be able to
implement a significant demonstration of this with additional funds.

Our architecture work will make use of a new XML/Jini experimental
collaboration system being developed inside NPAC. We expect this to fit and
support the CPA very well. We will also use our operational XML link to
Tango2 in the context of sharing information resources in a web-linked
databases using XML templates. This can be extended to include the set
of XML based information (stored both in and outside databases) expected to
be developed as part of the CPA. (A working group in this area was
established at Oak Brook). It is difficult to be precise at this stage
given the uncertainties in the implementation of the Alliance plans.
Collaborative visualization is likely to be important for the Alliance and
we will share our expertise here based on the integration of a Java3D
visualization system SV2 into Tango Interactive.
Finally we remember that Tango Interactive is living in a time of great
change in core software infrastructure. We therefore need an ongoing effort
to maintain Tango at leading edge.