SPI-CILLUSTRATIVE
Tracks quality-adjusted software production economics across comparable software classes.
SPI · Software Production Intelligence
The Software
Production Index.
Measuring the economics of software creation.
Measure everything. Learn from everything. Improve everything. An evidence-led foundation for better software-production decisions.
RESEARCH PREVIEW · NO OFFICIAL BENCHMARK PUBLISHED
Frontier artwork: illustrative scenarios, not validated observations or software-class rankings.
12Index families
S1–S5Software classes
Evidence-ledResearch framework
01 — How SPI works
From production to better decisions.
Measure what a build required. Check what it delivered. Learn what could make the next decision better.
- 01
Produce
A software build and its resource use.
- 02
Preserve evidence
Traceable artifacts, costs and quality checks.
- 03
Review & measure
Independent review before internal calculation.
- 04
Compare
Comparable classes, scope and evidence.
- 05
Inform the next build
Advisory learning—not an automatic action.
The target feedback loop is not live. Current public figures are illustrative; source telemetry and adaptive advice remain future work.
02 — Index universe
Twelve canonical indexes.
One evidence-led foundation.
Each index has a defined purpose, inputs, method and limitations. Only the SPI Composite carries values today — synthetic and illustrative. The other 11 are methodology only and await validated data.
SPI Composite
115.3
Illustrative · synthetic · not an official benchmark
Arileus Production Score
Not established
Published only once validated observations exist.
Production Alpha
Not established
Performance against the class frontier — awaiting validated data.
12 canonical indexes
QAPEMETHODOLOGY ONLY
Measures qualified software output relative to the resources required to produce it.
BCXMETHODOLOGY ONLY
Measures the fully loaded cost of developing and delivering software that meets defined quality standards.
TTXMETHODOLOGY ONLY
Measures the time required to transform approved requirements into accepted, deployable software.
CAPXMETHODOLOGY ONLY
Measures qualified software production output relative to the capital deployed.
RUXMETHODOLOGY ONLY
Tracks validated component reuse and its measured effects on development cost, time, and quality.
AIXMETHODOLOGY ONLY
Measures observable AI participation in software production, distinguishing automation from demonstrated productivity gains.
HLXMETHODOLOGY ONLY
Measures qualified software output per human hour, including oversight, corrections, and rework.
RQXMETHODOLOGY ONLY
Evaluates software against defined reliability, security, maintainability, and operational quality standards.
CRXMETHODOLOGY ONLY
Evaluates whether software meets the requirements for secure, supported commercial deployment.
PFXMETHODOLOGY ONLY
Tracks leading validated software production outcomes across comparable projects and complexity classes.
RCXMETHODOLOGY ONLY
Estimates the fully loaded current cost of reproducing equivalent software functionality and quality.
In focus · SPI-C
SPI Composite
Tracks quality-adjusted software production economics across comparable software classes.
Headline demonstration of index construction: a single series summarizing movement across all five software classes.
ILLUSTRATIVEEvidence: UNVALIDATED
115.3
ILLUSTRATIVE · synthetic · not an official benchmark
- Methodology
- v0.1-provisional
- Confidence
- U · unvalidated
Measurement relationships
Different questions. Shared input concepts.
Registry-based context only — not measured correlations or calculation dependencies.
SPI-C
Methodology & evidence requirements
Provisional formula
SPI Composite(t) = (S1(t) + S2(t) + S3(t) + S4(t) + S5(t)) / 5
Required inputs
- Class index levels S1–S5 (synthetic demonstration series)
Each class enters at equal weight; class levels are normalized to 100 at the base period before averaging.
- Equal 20% weights are a provisional demo convention, not an empirically validated weighting methodology.
- Underlying class series are synthetic; the composite is not a measured industry trend.
Canonical means part of the SPI registry, not validated or officially published. Only the Composite carries synthetic demonstration values. Proposed families have no index definitions or measurements.
03 — Adaptive intelligence
CONCEPTUAL ARCHITECTURE · NOT A LIVE CONNECTIONMeasure. Learn.
Improve the next decision.
ILLUSTRATIVE SCENARIO · NO LIVE CONNECTIONS · NO AUTOMATIC ACTIONS
New outcome → fresh evidence → independent review
01 / 06 — Phronesis
A build becomes a traceable outcome.
Illustrative scenario: an S3 enterprise workflow is delivered. Build artifacts, resource use and acceptance checks would accompany the outcome—not just a claim that AI made it faster.
- Receives
- Approved requirements and an authorized production decision
- Passes forward
- Build artifacts and recorded resource use
Authority & limits
Source-side production telemetry is not connected to this demonstration.
1 / 6 · CONCEPTUAL
Advice is not evidence. An AI hypothesis must never become a validated measurement or benchmark by repetition. Phronesis cannot determine the rules used to judge its own performance.
04 — Index Factory
PROPOSED · NOT OPERATIONAL
A definition today.
A governed index tomorrow.
The future Index Factory will connect a measurement purpose to its inputs, versioned method, evidence and approval requirements.
- 01
Define the measurement
- 02
Identify required evidence
- 03
Review before publication
Explore three proposed designs with synthetic calculations and simulated review. The demonstration creates no official indexes or internal measurement records.
Index intelligence
Economics, in perspective.
SPI CompositeILLUSTRATIVE
115.3
+15.3 vs synthetic base 100.0
Equal 20% weights across S1–S5 — a provisional demonstration convention, not official methodology. Synthetic data; not measured industry performance.
Composite measurement
- SPI Composite
Tracks quality-adjusted software production economics across comparable software classes.
Status: ILLUSTRATIVE · Synthetic demonstration
Evidence: UNVALIDATED
Synthetic Composite rises from base 100 to 115.3 ILLUSTRATIVE; this is a demonstration, not measured industry growth.
ILLUSTRATIVE · Synthetic series · Provisional equal 20% class weights
02 — The production frontier
Cost efficiency moves.
Not equally for every class.
Index of quality-adjusted output per unit of cost, base 100. Toggle classes to compare trajectories.
Illustrative · synthetic data
Software class measurements
- S1 · Utility Software
Narrow functionality, limited integration, low operational criticality.
Examples: A file-conversion utility; a standalone calculator.
Complexity, integrations and operational demands change production effort and quality requirements; compare like-for-like software within its class, not raw costs across classes. Examples illustrate scope, not automatic class assignments.
Status: ILLUSTRATIVE · Synthetic scenario
Evidence: UNVALIDATED · Verification not performed
- S2 · Commercial Applications
Customer-facing products, persistent data, defined commercial workflows.
Examples: An online booking application; an e-commerce storefront.
Complexity, integrations and operational demands change production effort and quality requirements; compare like-for-like software within its class, not raw costs across classes. Examples illustrate scope, not automatic class assignments.
Status: ILLUSTRATIVE · Synthetic scenario
Evidence: UNVALIDATED · Verification not performed
- S3 · Enterprise Platforms
Multiple roles, enterprise workflows, integrations, operational controls.
Examples: An enterprise resource-planning platform; an integrated customer-management platform.
Complexity, integrations and operational demands change production effort and quality requirements; compare like-for-like software within its class, not raw costs across classes. Examples illustrate scope, not automatic class assignments.
Status: ILLUSTRATIVE · Synthetic scenario
Evidence: UNVALIDATED · Verification not performed
- S4 · Critical Infrastructure
Elevated security, reliability, governance or regulatory requirements.
Examples: A payment-processing system; an operational infrastructure-control system.
Complexity, integrations and operational demands change production effort and quality requirements; compare like-for-like software within its class, not raw costs across classes. Examples illustrate scope, not automatic class assignments.
Status: ILLUSTRATIVE · Synthetic scenario
Evidence: UNVALIDATED · Verification not performed
- S5 · Composite Systems
Multiple integrated products or services with substantial orchestration.
Examples: A multi-product enterprise ecosystem; an orchestrated logistics network.
Complexity, integrations and operational demands change production effort and quality requirements; compare like-for-like software within its class, not raw costs across classes. Examples illustrate scope, not automatic class assignments.
Status: ILLUSTRATIVE · Synthetic scenario
Evidence: UNVALIDATED · Verification not performed
These hypothetical paths rise at different rates. They demonstrate possible efficiency changes, not a measured production frontier or an industry forecast.
Two views of production efficiency: cost per qualified unit decreases as efficiency improves, while qualified output per dollar increases. Both charts are illustrative scenarios, not measured industry trends.
Observed benchmark series will be shown in a distinct style with provenance and confidence bands once real data is published.
03 — Five software classes
Compare like
with like.
Class membership follows published criteria across functional complexity, operational criticality and architectural integration — never terminology alone.
S1
Utility Software
Narrow functionality, limited integration, low operational criticality.
Class index 128.4 · illustrative
S2
Commercial Applications
Customer-facing products, persistent data, defined commercial workflows.
Class index 119.7 · illustrative
S3
Enterprise Platforms
Multiple roles, enterprise workflows, integrations, operational controls.
Class index 112.6 · illustrative
S4
Critical Infrastructure
Elevated security, reliability, governance or regulatory requirements.
Class index 106.2 · illustrative
S5
Composite Systems
Multiple integrated products or services with substantial orchestration.
Class index 109.8 · illustrative
Complexity · Low / Medium / High
Criticality · Standard / Elevated / Critical
Delivery state · Prototype / Validated Demo / Production / Qualified Operational
SPI Research Terminal
Investigate the production frontier.
Indexes, class comparisons, evidence and methodology — one connected research environment.
08 — Methodology & evidence
Independent by design.
Checkable by evidence.
Origin and validation are separate facts. Public series are synthetic. The original 12 definitions remain authoritative; no official benchmark has been established.
Research framework, quality gates & methodology
01 — What SPI measures
Speed is not production.
Qualified output is.
How much high-quality, commercially useful software can be produced for a given amount of capital and time?
Core relationship
Qualified Software Output
Production Cost
Lines and commits measure activity, not delivered function.
Raw velocity ignores defects, rework and unshippable output.
AI coding benchmarks measure task capability, not production economics.
04 — Quality
Quality changes everything.
Output only counts once it passes qualification gates. Cheap software that fails in production is not efficient — it is deferred cost.
- G1Functional verificationDelivered functionality is tested against specification.
- G2ReliabilityBehaviour holds under realistic load and failure conditions.
- G3SecurityStructural and dependency risks are assessed, with material weaknesses remediated or addressed under defined acceptance criteria.
- G4MaintainabilityCode can be changed safely by someone other than its author.
- G5Commercial readinessThe software meets defined operational, support and release-readiness requirements.
06 — Research & methodology
Trusted because
it can be checked.
SPI draws on established work — function points, COCOMO II, DORA, ISO/IEC 25010 and 5055, ISBSG — and is designed against benchmark-governance principles.
Published definitions
Every metric, class criterion and adjustment is documented before it is used.
Reproducible changes
Index movements can be recalculated from versioned inputs and rules.
Visible confidence
Each measurement carries provenance and a data-confidence grade.
Conflict controls
Arileus submissions are identified and subject to the same published qualification rules, with safeguards against influencing their own benchmark comparisons.
Status: provisional methodology. Formula calibration awaits a real pilot dataset. Research publications forthcoming.
Full methodology for all 12 indexes — Index Registry →05 — The compounding thesis
Beat the index.
EVERY BUILD SHOULD MAKE THE NEXT BUILD BETTER.
Organizations that reuse validated components should see marginal production cost fall faster than the class benchmark. Production Alpha measures that gap — whether it is positive or negative. If a producer underperforms, SPI publishes that too.
07 — The SPI platform
One system for
production intelligence.
01
Index
Composite and class-level SPI series.
Open in terminal →02Frontier
The quality-adjusted cost frontier over time.
Open in terminal →03Organizations
Production Score and Alpha against peers.
In development · preview →04Projects
Classified project observations with evidence.
In development · preview →05Classes
S1–S5 qualification criteria and comparisons.
In development · preview →06Research
Methodology notes, papers and data releases.
In development · preview →08 — Commercial use
Enterprise Benchmarking
“Where are our production costs and inefficiencies?”
Portfolio Intelligence
“How does production performance compare across holdings?”
Due Diligence
“What evidence supports build-cost and replacement-cost assumptions?”
Research
“How is the industry production frontier changing?”
09 — Arileus ecosystem
Part of a wider
intelligence system.
PhronesisScalarSyndesisNeuroTraderEventEdgeCapital IntelligenceHiveSPI
Connected intelligence
Production. Measurement. Evidence.
Build measurements
Phronesis
Planned source-side capture of cost, timing, human effort and AI participation.
Economic intelligence
SPI
Synthetic-tested ingestion adapters and internal experimental calculations.
Evidence & provenance
Scalar
Planned evidence attachment; independent human review remains the calculation gate.
Source systems currently expose simulations, not live production telemetry. Live capture awaits source-side installation and actual builds.
10 — Participation & research
Start with the evidence.
Explore the research direction and the requirements for credible measurements. Research publications and participation workflows remain in development.