The Evidentiary Record
in progress

Applied thermodynamic principles by analogy to the problem of architectural selection under legacy constraints. Introduced an entropy-based formulation treating architectural desirability as a function of position: the flexibility of the current architecture, the desirability of available transitions, the value of the new state, and the forward-looking position it enables. Proved three theorems demonstrating that the formulation has the right qualitative properties. Named the concept of option lock-out — the tendency to lose access to potentially desirable regions of the architectural space when exercising a transition.

Open research questions

Does the entropy-based formulation of architectural desirability survive derivation from first principles — and does it become a testable prediction of UST?

What is the relative entropy differential between within-family architectural adaptation and cross-family architectural transition? Does that differential follow a consistent pattern across domains?

These questions are noted here in the spirit the evidentiary record is intended to carry: open, inquisitive, and honest about what is not yet known.

Relationship to UST

Laid the groundwork for thinking in terms of entropy. Preliminary application of entropy-based principles by analogy.

Framed a normative comparative study to assess the validity of what the paper calls the system design community's deep-rooted assumption: that a system is a collection of interconnected parts that interact to produce some behavior. Evaluated two viewpoints against criteria for successful system design — assurance of the design process, and correctness and completeness of the solution.

The prevailing viewpoint, grounded in the constructionist assumption, was shown to become increasingly unsatisfactory as systems exceed the capability and capacity of the practitioner to comprehend them. The analysis further suggests the community has been actively compensating for the insufficiency of its models — adding complexity to manage complexity — which may itself generate the perception of complexity as significant to design rather than reflecting a property of the systems themselves.

An alternative viewpoint — one that takes the system's own point of view, treating it as a unified whole motivated to satisfy its own value structure — demonstrated theoretical agreement with the evaluation criteria regardless of scale and scope, and showed the potential for systematic solution derivation.

An unexpected finding: the analysis uncovered evidence for four distinct viewpoints — need-forming, solution design, construction, and interaction — that appear to work collaboratively. The prevailing viewpoint conflates all four, which may be a further source of the observed insufficiency.

We committed to our reviewers to develop the theoretical foundations the argument required. The Skeptical Systemist is the fulfillment of that commitment.

Relationship to UST

Questioned a foundational assumption the field had not examined, and committed to developing the theoretical alternative the argument pointed toward.

Reframed the processor security problem from defense against known attacks toward derivation of necessary and sufficient conditions for trusted execution. The governing specification — the Trusted Application Pattern Space (TAPS) — defines what must be true: each application executes in a private, distinct, non-overlapping instruction and data space that no other application can read, infer, or modify. Collaboration between applications occurs only through explicitly scoped shared spaces, accessible solely to the designated parties. These conditions must hold throughout the full execution lifecycle.

To enforce TAPS, the architecture asserts exclusive and immutable control over precisely the mechanisms that are necessary and sufficient to establish and maintain the self/other distinction — no more, no less. The operating system and hypervisor are treated as de-privileged applications, not trusted components. Everything that could enable or disrupt trusted execution is owned by a federation of roles and agents — each with exclusive and immutable authority over a precisely scoped, non-overlapping domain — forming a chain of dependency across the full stack. Everything else is left unrestricted.

An early instantiation of UST principles and an application of the approach to hardware security, compensating for the self/other architectural flaw inherent in most processor designs.

Relationship to UST

Reframed the computing platform security problem from the external observer's viewpoint to the system's own coherence: what must be true to execute the application lifecycle with proper distinction of self and other? The application produced a citable public outcome.