What good is a powerful simulation if people cannot explore what it means?

An atmospheric model asks extraordinary things of a computer. It also asks something of the person trying to understand the result.

Howard Pearlmutter's parallel-computing work for NASA Ames brought him into the machinery of that problem: computing architecture, programming approach, and the representation of a demanding scientific model.

He asked another question. How would human beings interact with such computational power? Would the results arrive as a tall stack of numerical printouts?

The question led to Interactive Computer Graphics: The Human Interface to Dynamic Simulation, a 700-page NASA Ames study drawing together computing, human factors, input and output technologies, hardware, and software.

The study examined a relationship that remains central to modern digital work. A result has a technical representation. A person has a purpose. The interface has to make a useful connection between them.

That connection can involve a change in medium. A table may reveal one kind of information. Animation may make a dynamic relationship visible. Interaction may let a person explore a question the original presentation did not anticipate.

The appropriate form depends on what the audience needs to inspect, understand, or do.

Howard's subsequent practice carried this concern through graphics, simulation, commercial applications, and human interfaces. The animated factory-design signoff system needed by GM and Ford users connected specialized models with executives. Modeling and rendering software contributed animation to Ecological Design: Inventing the Future. Current work brings simulation, graphics, media, enterprise software, and native computing into practical prototypes.

These are different applications of an enduring question: how can a technical capability expand human ability?

For an organization, the question can be very immediate. A team has a complex proposal nobody outside the specialty understands well enough to question. A demonstration is convincing but leaves different people imagining different results. A dashboard gives an overview while making the important relationship hard to inspect.

The solution begins with the audience and its responsibility. Which distinction should become visible? What should the person be able to vary or examine? What context must remain available? How will the group know whether the representation helps?

A working session can clarify the question. A prototype can test the interaction. Architecture and implementation can make the result part of a useful system.

Howard's wider experience includes computational-fluid-dynamics research at MIT, commercial software, enterprise integration, and advanced technical learning. That range can help keep the representation connected to the machinery and the human purpose.

Explore interactive explanation, media prototyping, or the factory story. The most interesting next medium may be the one that lets someone ask a better question.