%% SCRAP: papers/FINAL_REPORT/appendix_glossary
%% SOURCE: docs/working/papers/FINAL_REPORT/appendix_glossary.adoc
%% STATUS: CURRENT
%% FITS: ssrn/app-glossary, vol3-research/app-glossary
%% EDITORIAL: lifted — AsciiDoc syntax translated to LaTeX; adoc stem[] → $...$,
%%            |=== tables → tabular with booktabs, [glossary] → description env

\chapter*{Glossary of Terms}
\addcontentsline{toc}{chapter}{Glossary of Terms}
\label{app:glossary}

Precise definitions for all terminology used throughout this work.
Terms are organized alphabetically within three groups: core concepts,
feedback loop taxonomy, and deprecated terminology. A mathematical
notation table follows.

\section*{Core Concepts}

\begin{description}

\item[Adaptive Heartbeat]
Time-driven coordination mechanism that orchestrates feedback loop execution
at dynamically-adjusted intervals. The heartbeat thread executes
\texttt{vm\_tick()} at frequency $f_{\text{tick}}$, where
$f_{\text{tick}} \in [f_{\min}, f_{\max}]$ adapts based on system stability
metrics.

\textit{Measurement}: tick period in nanoseconds, configurable via
\texttt{HEARTBEAT\_TICK\_NS}.

\textit{Implementation}: background pthread executing
\texttt{heartbeat\_thread\_main()}.

\item[Attractor]
Stable equilibrium point or region in phase space toward which execution
trajectories converge. Formally, a fixed point $\mathbf{x}^*$ where
$F(\mathbf{x}^*) = \mathbf{x}^*$ for dynamical system
$\mathbf{x}_{t+1} = F(\mathbf{x}_t)$.

\textit{Measurement}: coordinates in $(w, \lambda, \sigma^2)$ phase space.

\textit{Empirical observation}: StarForth exhibits a stable fixed-point
attractor across 90 experimental runs.

\item[Coefficient of Variation (CV)]
Normalized measure of dispersion, defined as the ratio of standard deviation
to mean:
\begin{equation*}
  \mathrm{CV} = \frac{\sigma}{\mu}
\end{equation*}

\textit{Convergence criterion}: $\mathrm{CV} \to 0$ indicates deterministic
convergence.

\textit{Application}: quantifies variance reduction in steady-state metrics.

\item[Decay Coefficient ($\lambda$)]
Rate parameter controlling exponential reduction in execution frequency over
time; units $[1/\text{time}]$.
\begin{equation*}
  f(t) = f_0 \cdot e^{-\lambda t}
\end{equation*}

\textit{Measurement}: derived via exponential regression on rolling window
data; stored as \Qtype\ fixed-point.

\textit{Typical range}: $\lambda \in [10^{-6}, 10^{-3}]$ per microsecond.

\item[Deterministic Convergence]
Property whereby repeated executions of identical workloads produce
statistically indistinguishable steady-state metrics. Formally:
\begin{equation*}
  \forall\, i, j\colon \frac{|\text{metric}_i - \text{metric}_j|}{\sigma} < \varepsilon
\end{equation*}
where $\varepsilon \to 0$ as $t \to \infty$.

\textit{Empirical result}: 0\% algorithmic variance across 90 runs
($\mathrm{CV} < 0.001\%$).

\textit{Significance}: enables reproducible performance characterization.

\item[Execution Frequency]
Count of times a dictionary entry has been executed since VM initialization,
optionally adjusted by temporal decay. This is the \emph{primary measurable
quantity} in the adaptive runtime.
\begin{equation*}
  f = \sum \text{executions} - \int \text{decay}(t)\,dt
\end{equation*}

\textit{Implementation}: unsigned 64-bit integer (\texttt{uint64\_t
execution\_heat}).

\textit{Note}: ``heat'' is a metaphorical naming convention; the actual
quantity is an execution count.

\item[Exponential Decay]
Mathematical function modeling reduction in execution frequency proportional
to current value:
\begin{equation*}
  f(t) = f_0 \cdot e^{-\lambda t}
\end{equation*}
where $f_0$ is the initial frequency.

\textit{Metaphorical parallel}: mathematically similar to radioactive decay
or thermal dissipation (conceptual metaphor only; no physical process implied).

\textit{Application}: applied periodically by the heartbeat system to reduce
the weight of stale frequency counts.

\item[Feedback Loop]
Self-referential process in which system output influences future input.
Classified as:
\begin{itemize}
  \item \emph{Positive} (amplifying): output reinforces input
  \item \emph{Negative} (stabilizing): output opposes input
  \item \emph{Neutral} (monitoring): no direct influence
\end{itemize}

\textit{Example}: Loop~1 (Execution Heat Tracking) is positive feedback:
\begin{center}
  Execution $\to$ Frequency$\uparrow$ $\to$ Cache Rank$\uparrow$
  $\to$ Lookup Speed$\uparrow$ $\to$ More Execution
\end{center}

\item[Hot-Words Cache]
Fixed-size array storing pointers to the $K$ most frequently executed
dictionary entries, enabling O(1) lookup acceleration.

\textit{Selection criterion}:
\begin{equation*}
  e \in \text{Cache} \iff \text{rank}(e) \leq K
\end{equation*}
where $\text{rank}(e) = |\{e' \in \text{Dictionary} : f(e') > f(e)\}| + 1$.

\textit{Performance impact}: reduces average lookup time by 70--95\%
(workload-dependent).

\item[Levene's Test]
Non-parametric statistical test for homogeneity of variance across groups.
Tests $H_0\colon \sigma_1^2 = \sigma_2^2 = \cdots = \sigma_k^2$ (equal
variances).

\textit{Test statistic}: F-statistic with associated $p$-value.

\textit{Application}: used in window width inference (Loop~5) to detect
variance changes when adjusting window size.

\textit{Decision threshold}: typically $\alpha = 0.05$ (5\% significance
level).

\item[Phase Space]
Multi-dimensional coordinate system where each axis represents a system state
variable. For StarForth:
\begin{equation*}
  \mathcal{S} = \{(w, \lambda, \sigma^2) \mid w \in \mathbb{N},\;
  \lambda \in \mathbb{R}^+,\; \sigma^2 \in \mathbb{R}^+\}
\end{equation*}

\textit{Dimensions}: $w$ (window size, execution events retained);
$\lambda$ (decay slope, frequency reduction rate);
$\sigma^2$ (variance, metric dispersion).

\textit{Analysis technique}: execution trajectories in phase space reveal
attractor basins.

\item[Rolling Window of Truth]
Circular buffer recording recent execution history for deterministic metric
seeding. Guarantees identical initial conditions across runs.

\textit{Data structure}: ring buffer $B[i] = \text{word\_id}$ at execution
event $i \bmod |B|$.

\textit{Buffer size}: configurable (default: \texttt{ROLLING\_WINDOW\_SIZE =
4096}).

\textit{Purpose}: enables reproducible variance calculations by providing
consistent historical context.

\item[Steady-State Equilibrium]
Condition where adaptive system metrics stabilize within bounded oscillation.
Formally:
\begin{equation*}
  \exists\, t_0\colon \forall\, t > t_0,\; |x(t) - x^*| < \delta
\end{equation*}
for small $\delta$.

\textit{Empirical criterion}: variance $\mathrm{CV} < 0.1\%$ over a
1{,}000-tick window.

\textit{Physical analogy}: similar to thermodynamic equilibrium in that
macroscopic properties cease changing (conceptual metaphor only).

\item[Thermodynamic Metaphor]
Conceptual mapping between thermodynamic quantities and execution metrics.
This is a \emph{metaphorical framework}, not literal physics.

\textit{Mappings}:
\begin{itemize}
  \item Heat $\leftrightarrow$ Execution Frequency
  \item Temperature $\leftrightarrow$ Normalized Rank
  \item Cooling $\leftrightarrow$ Exponential Decay
  \item Equilibrium $\leftrightarrow$ Steady State
\end{itemize}

\textit{Academic usage}: must be qualified as metaphor in formal writing.

\item[Transition Probability]
Conditional probability that word $B$ is executed immediately after word $A$.
Maximum likelihood estimate:
\begin{equation*}
  P(B \mid A) = \frac{\text{count}(A \to B)}{\text{count}(A)}
\end{equation*}

\textit{Implementation}: stored as \Qtype\ fixed-point in the
\texttt{transition\_metrics} structure.

\textit{Application}: used for speculative execution (prefetching the
likely-next word).

\item[Variance Inflection Point]
Window size $w^*$ where variance begins to increase when window shrinks below
$w^*$. Represents the optimal trade-off between sample size and temporal
locality.
\begin{equation*}
  w^* = \arg\min_{w \in [w_{\min},\, w_{\text{current}}]} \mathrm{Var}(w)
\end{equation*}

\textit{Search method}: binary search with Levene's test validation.

\textit{Purpose}: adaptive window size tuning (Loop~5).

\end{description}

\section*{Feedback Loop Taxonomy}

\begin{description}

\item[Loop 1: Execution Heat Tracking]
\textit{Type}: positive feedback (amplifying).
\textit{Mechanism}: increment frequency counter on each word execution.
\textit{Effect}: more executions $\to$ higher rank $\to$ more cache hits
$\to$ more executions.
\textit{Implementation}: \texttt{physics\_execution\_heat\_increment()} in
\texttt{vm.c}.

\item[Loop 2: Rolling Window History]
\textit{Type}: neutral (monitoring).
\textit{Mechanism}: record execution events in circular buffer.
\textit{Effect}: provides historical context for inference.
\textit{Implementation}: \texttt{rolling\_window\_record\_execution()} in
\texttt{rolling\_window\_of\_truth.c}.

\item[Loop 3: Linear Decay]
\textit{Type}: negative feedback (stabilizing).
\textit{Mechanism}: reduce frequency proportional to current value.
\textit{Effect}: high frequency $\to$ faster decay $\to$ lower frequency
$\to$ slower decay.
\textit{Implementation}: \texttt{vm\_tick\_slope\_validator()} applies linear
decay.

\item[Loop 4: Pipelining Metrics]
\textit{Type}: positive feedback (amplifying).
\textit{Mechanism}: track word-to-word transitions, predict next word.
\textit{Effect}: more transitions $\to$ better prediction $\to$ more
prefetch hits.
\textit{Implementation}: \texttt{transition\_metrics\_record()} in
\texttt{physics\_pipelining\_metrics.c}.

\item[Loop 5: Window Width Inference]
\textit{Type}: negative feedback (stabilizing).
\textit{Mechanism}: shrink window if variance increases (Levene's test).
\textit{Effect}: high variance $\to$ smaller window $\to$ lower variance.
\textit{Implementation}: \texttt{find\_variance\_inflection()} in
\texttt{inference\_engine.c}.

\item[Loop 6: Decay Slope Inference]
\textit{Type}: negative feedback (stabilizing).
\textit{Mechanism}: increase decay rate if metrics are unstable (exponential
regression).
\textit{Effect}: unstable metrics $\to$ steeper decay $\to$ faster
stabilization.
\textit{Implementation}: \texttt{infer\_decay\_slope\_from\_trajectory()} in
\texttt{inference\_engine.c}.

\item[Loop 7: Adaptive Heartbeat]
\textit{Type}: meta-loop (coordination).
\textit{Mechanism}: adjust tick rate based on system stability.
\textit{Effect}: stable system $\to$ slower ticks $\to$ reduced overhead.
\textit{Implementation}: \texttt{heartbeat\_thread\_main()} in \texttt{vm.c}.

\end{description}

\section*{Deprecated Terminology}

The following terms should be avoided in formal academic writing.

\begin{description}

\item[``Physics-based optimization'']
\textit{Use instead}: ``thermodynamically-inspired metaphor for frequency
decay.''
\textit{Reason}: implies literal physics; the implementation uses integer
counters and exponential functions only.

\item[``Execution heat'' (formal context)]
\textit{Use instead}: ``execution frequency with temporal decay.''
\textit{Reason}: ``heat'' is a metaphorical label; the precise term avoids
confusion in academic writing.

\item[``AI-driven'' or ``ML-based'']
\textit{Use instead}: ``statistically-inferred'' or ``adaptive via Levene's
test.''
\textit{Reason}: no neural networks or machine learning are involved.

\item[``Learning'']
\textit{Use instead}: ``adaptive inference'' or ``parameter convergence.''
\textit{Reason}: not supervised or unsupervised learning; this is statistical
convergence.

\item[``Quantum-inspired'']
\textit{Use instead}: N/A.
\textit{Reason}: no quantum mechanics or superposition is involved.

\end{description}

\section*{Mathematical Notation}

\begin{table}[h]
\centering
\caption{Mathematical symbols used throughout this work.}
\label{tab:notation}
\begin{tabular}{ll}
\toprule
\textbf{Symbol} & \textbf{Definition} \\
\midrule
$f$ & Execution frequency (count with decay) \\
$\lambda$ & Decay coefficient $[1/\text{time}]$ \\
$w$ & Window size (number of events) \\
$K$ & Cache size (constant) \\
$\sigma$ & Standard deviation \\
$\mu$ & Mean value \\
$\mathrm{CV}$ & Coefficient of variation $= \sigma / \mu$ \\
$P(B \mid A)$ & Transition probability (word $B$ after word $A$) \\
$w^*$ & Variance inflection point \\
$\mathcal{S}$ & State space $= \{(w, \lambda, \sigma^2)\}$ \\
$\mathbf{x}^*$ & Attractor (fixed point) \\
$F$ & State transition function \\
$f_0$ & Initial frequency \\
$t$ & Time (ticks or microseconds) \\
$r(t)$ & Execution rate $[\text{executions}/\text{second}]$ \\
\bottomrule
\end{tabular}
\end{table}

\section*{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}