%% SCRAP: papers/ONTOLOGY
%% SOURCE: docs/working/papers/ONTOLOGY.md
%% STATUS: CURRENT
%% FITS: ssrn/ch-ontology, vol3-research/ch-ontology
%% EDITORIAL: lifted — prose rewritten to press voice; thermodynamic-as-metaphor framing preserved throughout

\section{Ontology, Taxonomy, and Lexicon}
\label{sec:ontology}

This section provides the formal conceptual framework for the StarForth
adaptive runtime. Its purpose is to eliminate ambiguity, distinguish
metaphorical from literal components, and enable precise academic discourse.
Reviewers should read this section before evaluating any terminology appearing
elsewhere in the work.

\subsection{Conceptual Framework}

\subsubsection{Core Concepts}

The adaptive runtime system is organized around four conceptual domains:

\begin{description}
  \item[Execution metrics] Execution frequency (the primary measurable
    quantity), temporal decay (derived), and transition probability (derived).
  \item[Adaptive mechanisms] Frequency-based caching, window-based inference,
    and decay-based pruning.
  \item[Convergence properties] Deterministic behavior, steady-state
    equilibrium, and variance reduction.
  \item[Analysis frameworks] Dynamical systems view, statistical inference
    view, and control theory view.
\end{description}

\subsubsection{The Thermodynamic Metaphor}

Using execution frequency as a proxy for thermal energy, the system maps
thermodynamic quantities to implementation quantities as follows:

\begin{table}[h]
\centering
\caption{Metaphorical mapping. Column ``Thermodynamic'' lists the borrowed
concept; column ``Implementation'' lists the literal quantity or function.
The mapping is conceptual, not physical.}
\begin{tabular}{ll}
\toprule
\textbf{Thermodynamic concept (metaphor)} & \textbf{Implementation (literal)} \\
\midrule
Thermal energy & Execution frequency (integer count) \\
Heat dissipation & Exponential decay: $f(t) = f_0 e^{-\lambda t}$ \\
Thermal equilibrium & Steady-state convergence (stable metrics) \\
Temperature & Normalized frequency rank \\
Cooling rate & Decay coefficient $\lambda$ \\
\bottomrule
\end{tabular}
\end{table}

The literal implementations do not depend on the metaphor: a frequency
counter is an integer increment, and decay is a multiplication. The metaphor
provides intuition; the mathematics stands independently.

\subsection{Taxonomy}

\subsubsection{Component Hierarchy}

\begin{description}
  \item[Measurement layer]
    \begin{description}
      \item[1.1 Execution frequency tracking] Per-word integer counter,
        incremented on each execution.
      \item[1.2 Temporal recording] Rolling Window of Truth---circular buffer
        of execution events.
      \item[1.3 Transition tracking] Word-to-word transition matrix.
    \end{description}

  \item[Transformation layer]
    \begin{description}
      \item[2.1 Linear decay (Loop~3)] $\Delta f = -k \cdot \Delta t$
      \item[2.2 Exponential decay (Loop~6 inference)] $f(t) = f_0 e^{-\lambda t}$
      \item[2.3 Frequency ranking] Sort by decayed count; assign ordinal rank.
    \end{description}

  \item[Inference layer]
    \begin{description}
      \item[3.1 Window width inference (Loop~5)] Levene's test; binary search
        for variance inflection point.
      \item[3.2 Decay slope inference (Loop~6)] Exponential regression;
        least-squares fit on rolling window data.
    \end{description}

  \item[Actuation layer]
    \begin{description}
      \item[4.1 Hot-words cache (Loop~1)] Top-$K$ selection; O(1) fast-path
        lookup.
      \item[4.2 Speculative execution (Loop~4)] Transition probability
        calculation; prefetch decision.
    \end{description}

  \item[Coordination layer]
    \begin{description}
      \item[5.1 Adaptive heartbeat (Loop~7)] Time-driven tick generation;
        loop orchestration; adaptive tick rate.
    \end{description}
\end{description}

\subsubsection{Feedback Loop Classification}

\begin{description}
  \item[Positive (amplifying)] Loop~1 (Execution Heat Tracking): more
    executions $\to$ higher rank $\to$ more cache hits $\to$ more executions.
    Loop~4 (Pipelining): more transitions $\to$ better prediction $\to$ more
    prefetch hits.

  \item[Negative (stabilizing)] Loop~3 (Linear Decay): high frequency
    $\to$ faster decay $\to$ lower frequency. Loop~5 (Window Width
    Inference): high variance $\to$ smaller window $\to$ lower variance.
    Loop~6 (Decay Slope Inference): unstable metrics $\to$ steeper decay
    $\to$ faster stabilization.

  \item[Neutral (monitoring)] Loop~2 (Rolling Window): records execution
    events; provides historical context for inference.

  \item[Meta-loop (coordination)] Loop~7 (Adaptive Heartbeat): stable
    system $\to$ slower ticks $\to$ less overhead.
\end{description}

\subsection{Lexicon}

Terms are listed alphabetically. For each term, the formal definition is
given first, followed by implementation details and category. Where a
thermodynamic metaphor is in use, it is explicitly flagged.

\paragraph{Adaptive Heartbeat.}
Time-driven coordination mechanism executing \texttt{vm\_tick()} at
dynamically-adjusted frequency $f_{\text{tick}} \in [f_{\min}, f_{\max}]$
based on system stability. Implementation: background pthread.
Category: coordination mechanism.

\paragraph{Attractor.}
Stable equilibrium point $\mathbf{x}^*$ in phase space where
$F(\mathbf{x}^*) = \mathbf{x}^*$ for dynamical system
$\mathbf{x}_{t+1} = F(\mathbf{x}_t)$. Measured in
$(w, \lambda, \sigma^2)$ coordinates. Category: dynamical systems.

\paragraph{Decay Coefficient ($\lambda$).}
Rate parameter in $f(t) = f_0 e^{-\lambda t}$; units $[1/\text{time}]$.
Derived via exponential regression; stored as \Qtype\ fixed-point.
Category: transformation parameter.

\paragraph{Deterministic Convergence.}
Property: $\forall i, j\colon |\text{metric}_i - \text{metric}_j| / \sigma
< \varepsilon$ as $t \to \infty$. Measured as CV $\to 0\%$.
Category: convergence property.

\paragraph{Execution Frequency.}
Count of dictionary-entry executions since VM initialization, adjusted
by decay: $f = \sum \text{executions} - \int \text{decay}(t)\,dt$.
Stored as \texttt{uint64\_t execution\_heat}.
\emph{Note}: ``heat'' is a metaphorical label; the quantity is a count.
Category: primary measurable.

\paragraph{Exponential Decay.}
$f(t) = f_0 e^{-\lambda t}$, applied periodically by the heartbeat system.
\emph{Metaphorical parallel}: similar to radioactive decay or thermal
dissipation in mathematical form only. Category: transformation function.

\paragraph{Hot-Words Cache.}
Fixed-size array of pointers to the $K$ most frequently executed dictionary
entries. Membership criterion: $e \in \text{Cache} \iff \text{rank}(e) \leq K$.
Provides O(1) lookup. Category: frequency-based optimization.

\paragraph{Levene's Test.}
Non-parametric test for homogeneity of variance; $H_0\colon \sigma_1^2 =
\cdots = \sigma_k^2$. Used in Loop~5 to detect variance changes as window
size varies. Category: statistical inference.

\paragraph{Phase Space.}
$\mathcal{S} = \{(w, \lambda, \sigma^2) \mid w \in \mathbb{N},\;
\lambda \in \mathbb{R}^+,\; \sigma^2 \in \mathbb{R}^+\}$.
Execution trajectories in this space reveal attractor basins.
Category: dynamical systems representation.

\paragraph{Rolling Window of Truth.}
Circular buffer $B[i] = \text{word\_id}$ at execution event $i \bmod |B|$.
Default size: 4{,}096 entries. Ensures identical initial conditions for
reproducibility. Category: temporal recording mechanism.

\paragraph{Steady-State Equilibrium.}
$\exists\, t_0\colon \forall t > t_0,\; |x(t) - x^*| < \delta$.
Criterion: CV $< 0.1\%$ over a 1{,}000-tick window.
\emph{Metaphorical parallel}: analogous to thermodynamic equilibrium
in the sense that macroscopic properties cease changing.
Category: convergence property.

\paragraph{Thermodynamic Metaphor.}
Conceptual mapping (Table~above). This is not a physics claim; it is a
modeling tool. All academic writing must qualify thermodynamic language
as metaphor. Category: conceptual framework.

\paragraph{Transition Probability.}
$P(B \mid A) = \text{count}(A \to B) / \text{count}(A)$. Stored as
\Qtype\ in the transition matrix. Category: derived metric.

\paragraph{Variance Inflection Point.}
$w^* = \arg\min_{w \in [w_{\min},\, w_{\text{current}}]} \text{Var}(w)$,
found via binary search with Levene's test. Optimal window size for stable
metrics. Category: inference target.

\subsection{Avoided and Deprecated Terms}

\begin{table}[h]
\centering
\caption{Deprecated terms and their replacements.}
\begin{tabular}{ll}
\toprule
\textbf{Avoid} & \textbf{Use instead} \\
\midrule
``Physics-based'' & ``Thermodynamically-inspired metaphor'' \\
``Execution heat'' (formal writing) & ``Execution frequency with decay'' \\
``Temperature'' & ``Normalized frequency rank'' \\
``Quantum-inspired'' & N/A (no quantum mechanics involved) \\
``AI-driven'' & ``Statistically-inferred'' \\
``Learning'' & ``Adaptive inference'' \\
``Training'' & ``Convergence to steady state'' \\
\bottomrule
\end{tabular}
\end{table}

\subsection{Mathematical Formalism}

\subsubsection{Execution Frequency Evolution}

Continuous-time model:
\begin{equation}
  \frac{df}{dt} = r(t) - \lambda f(t)
\end{equation}
where $r(t)$ is the execution rate [executions/second] and $\lambda$ is
the decay coefficient [1/second]. Solution:
\begin{equation}
  f(t) = e^{-\lambda t}\!\left[f_0 + \int_0^t r(\tau) e^{\lambda\tau}\,d\tau\right]
\end{equation}

\subsubsection{Hot-Words Cache Selection}

Cache membership:
\begin{equation}
  e \in \text{Cache} \iff \text{rank}(e) \leq K, \quad
  \text{rank}(e) = \bigl|\{e' \in D : f(e') > f(e)\}\bigr| + 1
\end{equation}

\subsubsection{Window Width Inference}

\begin{equation}
  w^* = \arg\min \bigl\{\text{Var}(w) : w \in [w_{\min}, w_{\text{current}}],\;
        p_{\text{Levene}}(w) < \alpha\bigr\}
\end{equation}

\subsubsection{Decay Slope Inference}

Given $\{(t_i, f_i)\}_{i=1}^N$ from the rolling window, log-transform
$\ln f_i = \ln f_0 - \lambda t_i$ and fit by least squares:
\begin{equation}
  \lambda^* = \arg\min_\lambda \sum_{i=1}^N \bigl[\ln f_i - (\ln f_0 - \lambda t_i)\bigr]^2
\end{equation}

\subsubsection{Convergence Metric}

\begin{equation}
  \text{CV} = \frac{\sigma}{\mu}, \qquad
  \text{convergence achieved when } \text{CV} \to 0
\end{equation}

\subsection{Ontological Commitments}

\paragraph{Foundational assumptions.}

\begin{enumerate}
  \item \emph{Frequency as proxy for importance.} Frequently executed words
    are most important to optimize. Justified empirically by Zipf-law
    execution distributions (measured $\alpha \approx 1.1$).
  \item \emph{Decay models temporal relevance.} Recent executions are more
    informative than distant past. Justified by the temporal locality
    principle.
  \item \emph{Determinism through convergence.} Adaptive systems can converge
    to deterministic steady states. Justified by fixed-point theorems for
    contractive mappings.
  \item \emph{Statistical inference validity.} Execution patterns are
    statistically analyzable. Justified by the Central Limit Theorem for
    $n \geq 30$ samples.
\end{enumerate}

\paragraph{Scope.}
The ontology covers execution frequency measurement and decay, adaptive
caching and inference, dynamical systems characterization, and statistical
convergence properties. It does not cover actual thermodynamic processes,
machine learning, quantum computing, or biological neural systems.

\subsection{Usage Guidelines for Academic Writing}

In abstracts and titles, use mathematical language:
\emph{``thermodynamically-inspired adaptive runtime''} is acceptable;
\emph{``physics-based virtual machine''} is not.

In technical sections, use literal descriptions: \emph{``frequency counter
incremented on execution,''} not \emph{``temperature increases when word heats
up.''} In results sections, report statistics: \emph{``CV = 0.00\%
($p < 10^{-30}$)''}, not \emph{``perfect thermodynamic equilibrium.''} In
discussion sections, exploratory language is permitted with explicit
qualification: \emph{``one interpretation is that\ldots however, causation
is not established.''}

\subsection{References}

\begin{itemize}
  \item Strogatz, S. (2015). \emph{Nonlinear Dynamics and Chaos}. Westview Press.
  \item \r{A}str\"{o}m, K. \& Murray, R. (2008). \emph{Feedback Systems}.
    Princeton University Press.
  \item Casella, G. \& Berger, R. (2002). \emph{Statistical Inference}.
    Duxbury Press.
  \item Bolz, C.\ et al.\ (2009). ``Tracing the Meta-Level: PyPy's Tracing
    JIT Compiler.'' \emph{ICOOOLPS}.
  \item Ertl, M.A.\ (1996). ``Stack Caching for Interpreters.'' \emph{SIGPLAN
    Notices}.
\end{itemize}