Reconstructing an undocumented system from partial evidence

The challenge

We were engaged to complete the user interface of an industrial data logger, on the understanding that the embedded software beneath it was finished. It was not. Underneath sat partial and largely untested firmware, no design documentation, incomplete schematics, and no test framework. The stated scope — a UI — was a small part of the real one.

What we did

• Reconstructed the architecture by inference and experiment, building a model of what the hardware and firmware actually did rather than what they were assumed to do.
• Established, with evidence, that the real scope was far wider than the interface — and made that case to the client rather than papering over it.
• Restructured at system level — across firmware, communications and application architecture — so the product had a foundation that could be built on.

The result

• The product was brought to a working, coherent state.
• The client had, for the first time, an accurate picture of what their system was and did — the basis for continuing development with confidence.
• Development continues from a clean architecture rather than an unknown one.

Recovering a stalled or undocumented development starts with telling the client the truth about its scope — then rebuilding the parts that can’t carry the product as they are. That your situation?