The Origins of AVS - Putting the Record Straight

[Ian Reid @stardent]

I would like to respond to a number of the comments you made in
your email of June 13th to Jeff Hanson.

'AVS was built at Stellar (now Stardent) by Craig Upson and his
crew..'

Easy-to-use Visualization Toolkits or Environments were under
discussion at Stellar in 1987 prior to Craig Upson's arrival. Craig
was brought in to contribute his wide experience in scientific user
requirements and rendering techniques. He was certainly a key
contributor to defining those requirements and establishing what
scientists wanted to do and the kind of environment they needed.
His position was that of Visualization Scientist in the Marketing
group. He worked alongside the AVS engineering team discussing
functionality and working with various prototypes of AVS.

The development team was led by Rob Gurwitz and then Dave Kamins.
Craig Upson played a limited role in AVS implementation. With the
exception of the VBUFFER module for volume visualization, there is
almost no code written by Craig in AVS.

'...using an interconnection mechanism invented by SGI's own Paul
Haeberli'

Both ConMan and AVS are implementations of visual programming
interfaces for data flow execution models, a concept which has been
described and implemented in numerous forms over the past 15 - 20
years. AVS is a released product, ConMan, so far as we are aware,
is not. Neither can claim to have 'invented' this approach. 

Prior work of which we are aware includes: a project called EOM by
Paul Pangaro for the MIT Architecture Machine in the mid 1970's;
Shane Robison (now at Apple) wrote his master's thesis project on
a graphical network editor for the Data Flow project for Al Davis
(1981); the PS300 Function Network Editor written in 1983 for E&S
by Dave Schlegel (a member of the AVS development team). Many of
these earlier implementations, including ConMan, were referenced
in the AVS team's paper 'The Application Visualization System: A
Computational Environment for Scientific Visualization (IEEE CG&A,
July 1988). See Nan Shu's book, "Visual Programming", pp 173, Data
Flow Diagrams. (Van Nostrand Reinhold, N.Y.,1988) for further
information on data flow approaches to visual programming and Brad
Myers' "Taxonomies of Visual Programming and Program
Visualization", Journal of Visual Languages and Computing (1990),
1, 97-123. 

Customers, other vendors and co-workers in the visualization field
have recognised AVS as the most innovative, functional and
extensible heterogeneous visualization environment available today.
It has been ported onto five different systems and further ports
are underway. The AVS development team is expanding rapidly and
working with AVS licensees and customers in extending the platform
support and functionality of the product. 


Ian Reid

Stardent Computer