Journal of Systems Thinking · Volume 6, Issue 1 · 2026-10-09

Feature-Controlled Programming

Embedding structural commitments in AI-assisted software development: an industrial case

Dave Silberman, Sree Kumar Valath Bhuan Das

https://doi.org/10.54120/jost.v26i1.ywrqbnt

Abstract

AI-assisted programming has made implementation abundant. It has not made structural judgment abundant. A programmer can move from an idea to substantial working code extremely quickly [1]. A companion paper argues that this abundance relocates the scarce human capability upstream, from Information (I) such as syntax, stacks, and APIs, to Organization (O): the structural decisions that determine what should be built, where its boundaries lie, what its parts and relationships are, and from whose perspective it is conceived. This paper presents an observational industrial case in which that argument is built into a working development environment. Feature-Controlled Programming (FCP) is a method used in the production development of a commercial systems thinking mapping platform built by a small distributed team using AI-assisted programming. FCP's rules — include a clarity gate that refuses vague feature cards, single-programmer checkout, a component map with overlap alerts, persistent serial identity, declared relationships, and a short release cycle. Together, these rules operationalize the DSRP-484 structural grammar directly in the workflow, upstream of code generation. Read structurally, the method forces Distinctions, makes part–whole Systems countable, makes Relationships writable, makes the active Perspective point visible, and closes the model–reality alignment loop. Because the rules are enforced procedurally, ordinary work produces a behavioral event stream — gate refusals, overlap alerts, checkouts, rework, releases — that supports longitudinal study [2]. We present the case as evidence of operationalizability, not comparative effectiveness, and we state its evidentiary limits plainly.

Download PDF

© 2026 Journal of Systems Thinking (JoST) · CC BY-NC-ND 4.0