Papers on the Compiler Compiler System: what it is, how the self-hosting bootstrap works, what it measures, and what it means for cross-language translation and knowledge representation.
The system these papers describe is covered by US Patent 8,464,232 — Compiler Compiler System with Syntax-Controlled Runtime and Binary Application Programming Interfaces, granted 11 June 2013 to A. F. Urakhchin.
| document | what it covers | |
|---|---|---|
| Three Bets Article Public Draft 0.1 | position paper: the three bets on the future of software | |
| Two Domains Article Public Draft 0.1 | companion paper: cross-language translation + knowledge representation | |
| Three Bets arXiv Draft 0.2 | the position paper in arXiv form (latest draft) — Zenodo preprint, DOI 10.5281/zenodo.22547440 | |
| CCS Compiler Compiler arXiv Draft 0.1 | flagship technical overview: self-hosting bootstrap, multi-language runtime emission, JSON benchmarks — Zenodo preprint, DOI 10.5281/zenodo.22546174 | |
| Cross Language Translation arXiv Draft 0.1 | the Python-to-TypeScript RealWorld translation case study — Zenodo preprint, DOI 10.5281/zenodo.22548018 | |
| Adaptive Programming High School Edition | plain-English explanation of Adaptive Programming | |
| CCT Adaptive Programming with Claude Code Draft 0.1 | the Adaptive Programming conceptual framework | |
| CCT RoundTrip Testing as Oracle Draft 0.1 | round-trip testing as an oracle: definition, properties, application | |
| Big Picture Small Use Case | end-to-end CCS walkthrough on a bank-account grammar | |
| CCS Technical Specification | the complete CCS reference: parsing model, runtime and binary representations, SGDL, generator subsystem, operational use — cross-referenced to the patent's claims and figures | |
| A Second Kind of Secret Draft 0.1 | position paper for cryptographers: grammatical-inference hardness as a complement to cryptographic confidentiality, and the primitive-design problem it opens | |
| MeTTa on CCS Draft 0.1 | MeTTa language support, what the -D flag builds for a consumer, and the external conformance corpus | |
| b3ubot workflow | Adaptive Programming explained for a newcomer, with b3ubot as the worked example: the three documents, the two human gates, and how to read a published step | |
| Adaptive Programming p2t Use Case Draft 0.1 | Adaptive Programming applied to one project: how a coupled, intractable translation problem was partitioned into ten fragment families, and the record of building on that partition — including where the estimates missed by an order of magnitude | |
| p2t Design Notes | The eight design documents p2t was built from, combined: the upstream service's entities, routes, contracts, flows and layers, then the partition into ten fragment families that made the translation tractable — written before any machinery existed, and frozen since | |
| b3ubot design | the live authoritative design document a step file's Parent: cites (snapshot) | |
| b3ubot end to end | the live implementation plan and ledger — where a unit of work comes from (snapshot) | |
| MeTTa Bulk Load Benchmark Draft 0.1 | producing a CCS binary once vs reading it many times: the crossover, both ratio families, and what the result does not establish | |
| ccs java hl7 | technical report for the HL7 provider market: CCS fundamentals, Java as the fifth SCB runtime, the five-runtime HL7 ORU^R01 demonstration with published code, JVM JSON benchmarks, a distributed-deployment pattern, and the good/bad/ugly assessment | |
| ccs x12 | technical report for the X12 EDI provider market — the HL7 report's sibling: the five-runtime envelope demonstration on X12 syntax, the one-day second-format transposition, the compliance-clean answer to X12's AI/IP licensing rules, a clearinghouse deployment pattern, and the good/bad/ugly assessment | |
| ccs cobol data | technical report on the mainframe DATA layer — the trilogy's third: the copybook as a compiled schema read byte-identically by five runtimes, the real-copybook corpus verdict, a double-oracled EBCDIC/COMP-3/zoned/REDEFINES decode witness, why the data layer and deliberately not translation, and the good/bad/ugly assessment | |
| why not protobuf | positioning one-pager answering "why wouldn't I just use protobuf?" — schema-first serialization vs the grammar-level engine, the non-canonical-encoding point, an honest side-by-side and decision guide, what is not claimed (no Go/C#, no RPC), and the one-file free-pilot offer with the docs/usecases map | |
| AI Limits For CCS Parsing Model | plain-English account of a clean-room experiment: five fresh AI sessions given only the seven requirements CCS satisfies, a sixth as blind scorer — all five converged on a flat lossless concrete syntax tree and none produced the CCS four-entity model; which of the six departures from the YACC lineage the requirements force and which they cannot, why (behavioral requirements vs an ontological choice), the experiment's known weaknesses, and the two follow-up rounds — Zenodo preprint, DOI 10.5281/zenodo.22548361 | |
| ccs web3 | report addressed to web3 infrastructure builders — the published-results mapping (five-runtime byte-identity, binary load speed, the ECIES/AES-GCM obfuscation layer, the second-kind-of-secret position) onto client diversity, bridge/oracle payloads, and indexer ingestion; Draft 0.2 adds the multisig custody demonstration (five-runtime byte-identity + per-record dual-recipient ECIES storage law, gate results, MIT-era-pin provenance and independence statement); the Go gap named, self-built-demo honesty, and a design-partner offer (Draft 0.2 re-gated same day as 0.1) |