Journal of Systems Thinking · Volume 6, Issue 1 · 2026-09-08
Object Oriented Programming Requires Object Oriented Thinking
From Information to Organization in the Age of AI
Dave Silberman, Sree Kumar Valath Bhuan Das, Derek Cabrera
https://doi.org/10.54120/jost.v26i1.4p7aait
Abstract
Artificial intelligence is changing what it means to program. Syntax, APIs, framework knowledge, boilerplate, and implementation can increasingly be supplied on demand, yet the harder problem remains: deciding what should be built and how the problem should be organized before it becomes code. This paper argues that programming has historically conflated two distinct factors: Information (I), the content available to a programmer, and Organization (O), the structure imposed on that information to form a mental model (). Object-oriented programming (OOP) developed powerful mechanisms for implementing object models but left the upstream cognitive work of identifying the right objects, boundaries, part–whole structures, relationships, and perspectives largely tacit. We call that missing activity Object Oriented Thinking (OOT). Drawing on O-Theory and DSRP Theory, we propose OOT as an onto-epistemic bridge between the Organization of mental models and the Organization of reality, using the shared structural grammar of identity–other, part–whole, action–reaction, and point–view. We examine the historical development and structural asymmetries of programming paradigms, distinguish model correctness from implementation correctness through the O-check, and show how explicit O can improve the conditions for model–world alignment without claiming omniscience or guaranteed correspondence. We then develop implications for AI-assisted programming, programming education, hiring, governance, and software design; propose a three-part OOT curriculum; and point to Feature-Controlled Programming (FCP), examined in a companion paper, as an observational case showing how structural commitments can be embedded in a production workflow. Finally, we specify five empirical studies to test the paper's central predictions. As AI makes implementation increasingly abundant, the relative constraint shifts upstream from knowing how to express a model to organizing, testing, and revising the model itself. OOT makes that capability explicit, inspectable, teachable, and testable. Programming today stands where chemistry stood before Kekulé's structural formulas — rich in working knowledge, poor in structural notation — with the difference that its candidate grammar is already on the table.
© 2026 Journal of Systems Thinking (JoST) · CC BY-NC-ND 4.0
