Build enough to answer the important question.

A prototype can make a proposed architecture or interaction inspectable. Its value depends on the question it is designed to answer.

Howard Pearlmutter can carry a technical question into practical work across enterprise systems, native integration, media, visualization, and human interfaces.

The useful first step is to identify what the organization needs to know before making a larger commitment.

Give the experiment a precise purpose.

It might examine a resource boundary, an information translation, a platform facility, a dynamic model, or an interaction with the intended user.

What observation would support the route? What would change it? Which conditions need to be represented? What remains outside the prototype?

These distinctions help the team interpret the result together.

Connect the artifact with the surrounding system.

A component demonstration can leave the actual integration unresolved. A compelling interface can leave the underlying information model uncertain.

The prototype should include enough of the relevant crossing to be informative. That does not require representing the entire application.

See integration experiments, information semantics, and interfaces.

Draw on practical examples.

Howard's work includes native C/Java OCR integration, enterprise content-management qualification and proof of concept, healthcare integration prototypes, and Telos work connecting native computing, enterprise middleware, media, graphics, and simulation.

Read OCR, healthcare, and Telos.

The cases show different ways to make an uncertain relationship concrete.

Leave the learning with the organization.

A useful prototype should come with an intelligible account of what it established and which questions follow. Your team can use the result in its next architecture, implementation, or review discussion.

The contribution can be complete in itself, including when it reveals a better route. Further work should follow what was learned.

Explore architecture, qualification, or technical transfer. Start with the assumption whose answer could change the plan.