SATOSHI MASTER FILE

"Everything measured, everyone compared, nothing fragmented."

This file consolidates the published Code DNA forensics (real source code pulled from bitcoin v0.1, bitcoinj, rippled, and stellar-core), the classic suspect board, and the Michael Peirce inquiry (232-page thesis stylometry + the book test + the code hunt) into one comparison. Companion file: the-michael-file.html (full dossier, v3).

Ⅰ · The Grand Matrix — every candidate, every measured dimension

Michael is the only candidate with a double-column strength (knowledge blind-spots + e-cash prerequisites) AND a measured Windows match — and the only one whose weakest column is "unknown" rather than "mismatch." That asymmetry is why he stays interesting: the others fail tests; he mostly hasn't taken them yet. The hyphen tell is his one taken-and-failed test.

Ⅱ · The Classic Seven — original board, summary verdicts

Ⅲ · The Code Exhibits — measured fingerprints on file

bool CTransaction::AddToMemoryPool()
{
    CRITICAL_BLOCK(cs_mapTransactions)
    {
        for (int i = 0; i < vin.size(); i++)
            ...
return error("AcceptTransaction() : ...");

void SHAMap::dirtyUp(...)
{
	while(!stack.empty())
	{
		SHAMapTreeNode::pointer node=stack.top();
		assert(branch>=0);

private InetSocketAddress nextPeer() throws ... {
	if(pnseedIndex >= seedAddrs.length)return null;
	return new InetSocketAddress(...);

void Config::parseNodeID(std::string configStr, ...)
{
    if(configStr.size()<2) throw std::invalid_argument(...);
    retKey= PubKeyUtils::fromStrKey(v.first);

1995, PayMe (TCD-CS-96-21), age ~21, verbatim:
"A prototype was implemented in a C++/Unix
environment on a Sun workstation cluster."
— coins · wallets · banks · signed receipts ·
double-spend DB · PgpTools/RSA/IDEA

2000, thesis:
"The bank maintains a database of the serial numbers
of all spent coins, to prevent double spending..."

[Dai99] W. Dai. Crypto++: a C++ class library...
[Sza96] N. Szabo. The mental accounting barrier...

            ∅

   NO PUBLIC CORPUS EXISTS

Ⅳ · The Shared Tell — one habit, whole field

Every measurable candidate — Hearn, Schwartz, McCaleb — compresses keywords and operators (if(x), node=stack.top()). Satoshi spaces them, every time, without one exception. Michael's code formatting is untested — meaning he is the only live candidate who hasn't failed this test yet. Finding one page of his source settles more than everything else combined.

Ⅴ · Consolidated Standing — the honest ranking

1. MICHAEL PEIRCE — HIGH-INTEREST UNRESOLVED — the classification must separate two evidence classes. Capability evidence (strong): a VERIFIED 1995 C++ e-cash implementation — PayMe: coins, wallets, banks, signed receipts, double-spend detection, "implemented in a C++/Unix environment" (TCD-CS-96-21, verbatim) — plus a full multi-party micropayment system in Java, the encyclopedic survey work, measured British English, and the exact knowledge blind-spot boundary. That is genuine implementation history in the precise problem domain, not just theory. Calling the artifact record "zero" was an overstatement — corrected. Individualized forensics (thin): no attributable personal source corpus to style-match; no 2008-era artifact; the prose round mismatched (sentence rhythm, semicolons, "as long as", "however"); the hyphen tell now stable across THREE of his documents (1995: 5:0 · 1999: 3:0 · 2000: 26:0 unhyphenated) vs Satoshi's hyphenated habit; the Java-era naming and RMI-level serialization read weak-negative. Capability match ≠ identification — profile fit of this quality has produced false positives throughout this case's history. Verdict: high-interest unresolved, not probable. The 1995 PayMe C++ source (or full project report) is the single highest-value artifact that could move this either way.
2. SZABO — best public-board candidate; stylometry + bit gold; blocked by American English and the missing-citation paradox.
3. BRITTO — matches only the silence; unfalsifiable by construction.
4. SCHWARTZ — best C++ caliber; blocked by tabs, philosophy, voice, nationality.
5. BACK — the influence; the 2008 emails and the 2026 denial hold.
6. FINNEY — the first believer; the record reads witness, not author.
7. WHEELER · J. DONALD (added 8 Aug 2026) — the metzdowd room. Lynn Wheeler: domain-perfect payments veteran (IBM, X9.59/AADS) whose garlic.com voice — lowercase, ellipsis-chained, self-citing — and unbroken public posting through the Satoshi window rule him out on style and calendar. James A. Donald: the first public reply to the whitepaper announcement, digital-money toolsmith (Crypto Kong, echeque.com) — but the reply attacks the design's scaling, his prose is Satoshi's opposite, and his cypherpunks pedigree collides with the blind spot. Both: interesting, thin.
8. LE ROUX (added 8 Aug 2026) — the mastermind theory, via a Kleiman v. Wright filing. E4M gives him the only Windows-C++ crypto pedigree among alternative suspects — and the RX Limited empire calendar, the 2012 arrest with untouched coins through every year he needed leverage, and a life of ruthless monetization make Satoshi's fifteen-year restraint unrecognizable in him. Capability yes; character no.
9. McCALEB · HEARN · DAI · SASSAMAN — each cleanly exculpated on behavior, inbox, citation trail, or clock.
10. WRIGHT · DORIAN — closed by court and by facts respectively.

Ⅵ · Prose Round Two — thesis vs whitepaper, with co-author controls (7 Aug 2026)

Four documents, machine-measured: the thesis (113,099 words, solo), the whitepaper (3,518 words), the 1999 joint Peirce–O'Mahony paper (2,554 words), and a 2021 Tewari-group arXiv paper (4,958 words). Burrows-style delta on 48 function words, plus sentence architecture, punctuation, and discourse markers.

Ⅶ · Ian Goldberg — the ingredients without the bridge (11 Aug 2026)

A new falsification-first dossier tests Goldberg's pre-Bitcoin record through 31 October 2008. The primary record now confirms the HINDE collaboration with Ryan Lackey, deep DigiCash protocol work, direct exposure to Hashcash, a distributed pseudonymous-network design, and a 2007 C++ implementation that tolerates incorrect database replies. The same record is also unusually clear about the missing piece: e-cash and client puzzles remain separate anti-abuse tokens; the mint or organizational trust root remains authoritative; and no recovered design uses accumulated work to order conflicting spends or choose a canonical transaction history.

Ⅷ · Open Actions

• [TOP TARGET] The 1995 PayMe C++ source or full B.A. project report — the single highest-value artifact. Candidate locations: TCD undergraduate project archives, Wayback captures of ganges.cs.tcd.ie/mepeirce/ (manual — blocked from my tooling), WWW4 conference archives, PgpTools-era Unix software archives.
• [MANUAL — YOU] Wayback Machine: browse web.archive.org for ganges.cs.tcd.ie/mepeirce/. Look for PayMe pages, software links, code samples, "last updated" stamps after 2002.
• Prose round two — DONE (7 Aug 2026): see section Ⅵ. Net result: did not corroborate. Delta nominally favorable; every individualized habit (sentence rhythm, semicolons, "as long as", "however", hyphen tell) mismatched. Caveat: no solo O'Mahony text located — the joint-99 control is half-Michael, and the Tewari control is post-Bitcoin American-formatted.
• [ME, ON REQUEST] The 1997 first edition book test (Google Books id rCYsAQAAIAAJ) — was the canon boundary already in place pre-1999?
• [STANDING RULE] The bar for ever going further than this file: evidence that survives the Dorian test — not a compelling profile.