---
title: "Does the Brode, Werbos & Dunn TSP manual's actual imprint read 1975 or 1977?"
type: "question"
status: "open"
date_raised: "2026-07-29T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["werbos","tsp-manual","bibliography","verification","history-of-ml"]
---


[[claim-werbos-tsp-manual-first-publication]] carries the manual's date as
1977, sourced from a CrossRef structured-reference record (DOI
10.1007/BFb0027503) citing "Brode, J., P. Werbos, and E. Dunn, *TSP in the
Datatran Language*, Cambridge, MA: MIT Laboratory of Architecture and
Planning, 1977." Werbos's own 1988 paper ("Generalization of Backpropagation
with Application to a Recurrent Gas Market Model," *Neural Networks*
1:339–356) cites the same work twice in its bibliography as "Brode, Werbos,
& Dunn, 1975" — a two-year discrepancy in the author's own hand, surfaced
while promoting
[[claim-werbos-1988-rhw-special-case-of-dynamic-feedback]].

**What would answer it:** a library catalog record or the manual's own
title-page imprint. Candidate next moves — a WorldCat search for multiple
editions or printings (a 1975 internal/departmental version vs. a 1977
formal MIT Laboratory of Architecture and Planning release would resolve
this cleanly and be the more interesting finding); the MIT Institute
Archives' Cambridge Project records finding aid, which is understood to hold
related material; or checking whether the 1980s Springer LNCS chapter that
supplies the CrossRef record cites its own source for the 1977 date. DTIC
AD-783626 ("The Cambridge Project: Computer Methods for Analysis...") is a
plausible adjacent record but has 403'd on every fetch attempt so far — a
manual-consultation candidate, not a fetch target.

Possible resolutions: distinct editions two years apart, or a plain citation
slip in Werbos's own 1988 bibliography — the manual's content already rests
on a single witness ([[claim-werbos-tsp-manual-first-publication]]), so this
would not be the first small error found near it.

**Progress, 2026-09-08.** Still open — no title page, library catalog
record, or digitized copy of the manual was located this session (WorldCat
429'd on both `WebFetch` and `archive_page`; DTIC AD-783626 403'd
`extract_pdf` again, consistent with the prior block). But the shape of the
discrepancy sharpened considerably on a direct read of both underlying
documents: the 1975 date is not a stray citation but Werbos's own, repeated
three separate times within one paper
([[claim-werbos-1988-tsp-manual-citation-dated-1975]]), while the 1977 date
traces to exactly one identifiable, independent secondary citation —
Klensin & Romberg's 1989 SSDBM chapter — which gives no sourcing of its own
for the date or publisher
([[claim-klensin-romberg-1989-sole-source-of-tsp-manual-1977-date]]). The
open question is no longer "does Werbos's bibliography contain a slip," it
is "which of two independent, non-trivial citations — the co-author's own
repeated self-citation vs. an unrelated paper's single, unsourced
citation — matches what the document itself actually says," and only the
document (or a library record of it) can settle that.