Corroboration · An audit of the record

The Classification Record

The paper trail behind the Classification of Finite Simple Groups, censused with respect

Integrity catches

Seven fixable things

A census reads a bibliography more literally than any human reader ever would, and small defects fall out. Here are the ones we are confident in. Each is repairable from public records. None of them touches the mathematics, and none of them is a criticism of it. This is the errata-culture gesture the series itself models, returned by a stranger who happened to read very closely.

Offered first to the people who made the proof

The same findings, verified and with their evidence attached, have been offered privately to the series' editors first. They are the kind of thing a maintained ledger retires the moment they are known, and a frozen page keeps forever.

Seven defects in the canonical reference list

Each is fixable from public records. None touches the mathematics. They are the kind of thing a maintained ledger retires automatically, and a frozen one keeps forever.

[Lyo72]does not touch the maths

One tag, two different papers

The printed book reuses [Lyo72] for two unrelated Lyons papers.

[Lyo72] labels "A characterization of the group U3(4)", Trans. AMS 164 (1972) in the GW-type section, and "Evidence for a new finite simple group", J. Algebra 20 (1972) in the characteristic-2 section. Legitimate within section-scoped citation; a hazard for any tool that reads the bibliography as a graph; the reason the census has 506 unique tags across 507 records.

Fixable from: public bibliographic records

[Hig72]does not touch the maths

Notes that were already a published paper

Cited in 2011 as mimeographed notes; a same-title published version had existed since 1974.

Higman's Oxford notes are cited as unpublished, but the same material appeared as Symposia Mathematica XIII (1974), 531–540, Academic Press, thirty-seven years before the citation. The resolution was available the whole time; nothing recorded it.

Fixable from: public records

[Man79]does not touch the maths

A year that disagrees with the university's own record

Printed 1979; Berkeley's official record says 1977.

Berkeley's own record gives the thesis title "The Standard Form Problem Associated with the Sporadic Simple Group ·2C" and the date 1977. The bibliography prints 1979.

Fixable from: the awarding institution's record

[Pag69]does not touch the maths

Probably the wrong author

Attributed to "D. Page"; the evidence points to Peter Rowlinson.

A published Springer citation (Liebeck, Math. Z. 168 (1979), reference 13) and Stirling University's staff profile both attribute the same-titled Oxford D.Phil to Peter Rowlinson. The deep resolution pass upgraded this from a possible misattribution to one that should be corrected.

Fixable from: a published reference list and a university profile

[Pat72]does not touch the maths

A thesis year the series disagrees with itself about

1972 here, 1974 in GLS volume 3, 1973 in a biographical record.

Surveys 172 and the ATLAS give 1972; GLS volume 3's endmatter gives 1974; a biographical record gives 1973. The reference series carries three different years for the same thesis in its own volumes. Best reconciliation from the Cambridge deposit is 1973.

Fixable from: the series' own volumes plus the Cambridge deposit

[McC82]does not touch the maths

An approximation the AMS could have removed

Printed "roughly ... about 1982"; the exact title and year sit in the AMS's own 1980 records.

The entry is printed with an approximate title and year. It resolves to McClurg, "Classification of F1-pairs in Chevalley groups over fields of even order", UCSC 1980, from the AMS's own November 1980 Notices dissertation list; the publisher's own record held the exact entry all along.

Fixable from: the AMS's own dissertation list

[Ega80]does not touch the maths

A forthcoming-manuscript note next to its own successors

"New MS to be submitted soon", printed in 2011, lines from the entries that superseded it.

The entry's note points to a manuscript to be submitted; the same 2011 bibliography already lists that work's published successors ([Ega81], [EY82]) a few entries away. What the note itself refers to was never traced.

Fixable from: the same page of the same book

Digit-level disagreements between the two bibliographies

Where the 1986 Bulletin survey and the 2011 Surveys 172 cite the same work, seven page or volume numbers disagree. Almost all resolve in the 2011 list's favour. These are transcription findings about the printed pages, verified against the scans, not corrections imported from later metadata.

ReferenceGorenstein 1986Surveys 172Nature
[4] / [Asc74]pages 87–223pages 87–112range disagreement
[20] / [AGL81]pages 335–455pages 335–456end page off by one
[91] / [Jan76]pages 564–496 (reversed)pages 564–596reversed and mistyped range; verified by eye against the 1986 scan
[111] / [Pha77a]pages 66–76pages 67–77range disagreement
[122] / [Smi79b]pages 251–280pages 251–281end page off by one
[141] / [Tim75a]pages 75–135pages 75–134end page off by one
[155] / [Pet84]volume 229volume 299volume digits transposed