← Back to journal

When a Smaller Supported Surface Is the Faster Route

Why product focus means reducing ambiguity before expanding coverage.

Evidence basisKalu, Korpus, and frozen Ushman source/docs snapshot reviewed September 17, 2026; no new runtime qualification performed
DisclosureEngineering or methodology note based on a September 17, 2026 snapshot. Source-level behavior is not a public availability or universal compatibility claim. Any scenarios are illustrative, not customer outcomes.
Related to Kalu KodeExplore Kalu Kode →

Scope: Engineering or methodology note based on a September 17, 2026 snapshot. Source-level behavior is not a public availability or universal compatibility claim. Any scenarios are illustrative, not customer outcomes.

Supporting fewer things can look like moving slowly. In an evidence-producing tool, it can be the route to moving responsibly: fewer hidden assumptions, clearer refusal behavior, and a smaller set of installed journeys that need to be understood end to end.

Kalu’s current supported host is macOS. KodeCapture’s early capability floor is deliberately narrower than the full browser platform. Those are product boundaries, not statements about what JavaScript or browser automation can theoretically do.

Breadth multiplies the questions

Adding an operating system is not only a packaging task. Process ownership, cleanup, path behavior, browser installation, and failure reporting all need to retain their meaning. A runtime that starts successfully does not establish that a timed-out browser was terminated safely or that recovery behaves correctly after interruption.

Adding a browser capability creates a similar chain. A new storage mechanism may affect session setup, capture identity, replay, and the interpretation of a checkpoint. Treating it as one checkbox can hide several missing owners.

A narrow path should still be demanding

Small scope is not an excuse for a weak demonstration. The selected path needs positive evidence, explicit bounds, and negative cases that show the system does not quietly fall back to something less controlled.

For example, using the expected browser executable matters. So does refusing an unsupported measurement rather than substituting a different browser and keeping the same claim. The result should stay connected to the conditions under which it was produced.

Expansion needs an independent decision

The frozen Ushman research gives us useful designs and failure cases for broader work. It does not remove the need to qualify the Kalu owner that receives a harvested capability. We can reuse knowledge without inheriting every old support promise.

Our active focus on KodeBack and KodeCapture is part of the same choice. KodeProof development is frozen rather than being described as equally active simply because it shares a repository and product-family diagram.

The goal is not permanent narrowness. It is an understandable expansion path. A product surface that is small enough to explain can become strong enough to extend. A broad promise that cannot be audited is much harder to repair after someone has relied on it.