Overall Discussion on Last Day

First Panel: Bob Lucas Chair

Written Suggestions of Future Activities

* CFDRC/John Rice:

- Geometry

* Minnesota:

- Preconditioners,
- expert systems,
- pre and post processing standards
- MDO PSE

* Argonne,MSU:

- software engineering standards (object-oriented)

* Purdue:

- Virtual Environments,
- Multidisciplinary and domain specific PSE's,
- Browsers/Agents to find useful material

Bob Lucas Again

* Design $10M/year program over next 6 months
* In same science/technology area as now
* Remember speakers at Arpa PI meeting from industry said software integration is key
* Package proposals as DoD area
* Metacomputing Programming Tools

- such as compilers and algorithms for World Wide Virtual Machine
- components and large complex systems
- legacy systems

Boisvert's Ideas --I

* Emerging Trends

- Computing is getting better
- Systems are getting more complex and flexible and this is a PSE
- global networks are increasing vision of what computer is

* High Fidelity Multi-Component Distributed Simulation
* Software development environment is bottleneck

Boisvert's Ideas --II

* High cost of developing a PSE

- GUI
- Expert Systems
- Ad Hoc Integration and distribution of components
- they are heterogeneous

* Need dynamic interoperable problem solvers on the network
* Lucas suggests eigensolvers on the network
* Components of Program

- distribution object management
- metadata to discover services
- coarse grain software integration
- need prototypes, PSE middleware

Discussion -- I

* Lucas says no two sparse solvers have the same interface and no two users use the same package

- Boisvert says object-oriented techniques solve
- But Lucas says this is a mature field; in real world can't afford performance loss of such wrappers
- Lucas notes real application people are cleverer than computer scientists think! They often use excellent innovative algorithms

* Sameh emphasizes expert systems to help choice of method

Discussion -- II

* Heath says grid generation and human time more important than PDE solution time

- Lucas counters that PDE solver purchase is done in "isolation" as a separate competition

* Need to trap and understand error messages from say NASTRAN and detect that say one needs to give it more memory
* Barry Smith says Web Technology is just TCP/IP!
* Lucas suggests partnering with either aerospace industry or government laboratory.

- But Lucas's Arpa's interest is mainly electronic not mechanical application

Discussion -- III

* I added following points

- In multidisciplinary PSE's need to do more than just CFD and structures -- ASOP needs 10,000 programs
- Shared memory and distributed memory metacomputer two architecture model
- Coarse grain and Fine grain software/problem models
- Exploit software engineering opportunities of the web

Discussion -- IV

* Lucas can't persuade its management to support MOSIS

- How can he persuade them to support software systems which are less clearly generally valuable