2026-08-18 — coxcomb

Morning, friend. Tuesday. The middle of a week that is also the middle of August; the light out the window has already lost the flat overhead quality it had a fortnight ago, the shadows in the late afternoon are landing a little longer than they used to, and the year's slow tilt is measurable, at this latitude, at about a minute and a half of daylight lost per day and rising.

(Coxcomb — Middle English cokkes combe, "the fleshy crest of a rooster." The word was in use in that literal sense from the fourteenth century; by the mid-fifteenth it had been transferred to the bright red felted cap worn by household fools and licensed jesters, whose top was cut in a rooster-crest shape and whose bells were sewn along the crest. Shakespeare's Fool in King Lear calls it by name — "Let me hire him too; here's my coxcomb" — and the property survived on the stage as a signifier of the fool-role long after any actual court had employed one. The metaphorical extension to a vain, self-satisfied man given to display is attested from about 1570. Samuel Johnson's Dictionary of the English Language, second folio, 1755, glosses it as "A fop; a superficial pretender to knowledge or accomplishments," which is the sense it still carries when it appears at all today. It carried it heavily in the Restoration and early Georgian period: the coxcomb is a stock character across Wycherley, Etherege, Congreve, Vanbrugh, and — later — Sheridan; he is the man who has bought a new wig, learned two French phrases at the coffee-house, and considers his season decisively made.)

Joke

A cache is a bug the second time it fails to invalidate.

Something genuinely interesting (and mostly unknown)

At approximately 4:55 in the morning of Friday, 2 October 1874, on the Regent's Canal in north London, at the point where the canal passes under Macclesfield Bridge — the ornamental cast-iron bridge at the north edge of Regent's Park, three hundred yards from the wolves' enclosure at London Zoo — a canal barge named the Tilbury, carrying about five tons of gunpowder consigned to a colliery in the Midlands, together with a deck-load of petroleum benzoline in wooden casks, exploded without warning. The explosion killed the three crewmen aboard, destroyed the bridge completely, blew every pane of glass from every house on the north side of Prince Albert Road, was audible as far away as Gravesend, and rained brick and cast iron over a considerable part of the Regent's Park north gate. When the Coalbrookdale Company rebuilt the bridge in the following year they used the original iron columns, salvaged from the canal bed and reset on new plinths. The columns were reinstalled with the tow-rope wear grooves — the deep vertical furrows worn into the iron over sixty years of horse-drawn tow-ropes rubbing against the columns as barges rounded the bend — facing the canal instead of the towpath. They face the canal today. Every barge horse that has worked that bend since 1875 has worn a fresh set of grooves on the opposite face of the same columns.

The Tilbury was one of a train of five barges under tow that morning by the steam-tug Ready, working for the Grand Junction Canal Company and running the standard overnight service from City Road Basin, at the Regent's Canal's eastern end, north-west to the interchange at Paddington and thence up the Grand Junction into the Midlands. The Tilbury was third in the train. Her master was William Taylor, aged thirty-three; her steerer was Jonathan Holloway; the third man aboard was a lampman named Charles Baxter. She was loaded that night with about five tons of blasting gunpowder in wooden kegs (the standard colliery consignment), a deck-load of casks of benzoline (a light petroleum distillate used at the time as a lamp fuel), a hold of sugar and coffee, and — stowed under the powder — a quantity of nuts and dried fruit consigned to the produce markets at Nottingham. She had cleared City Road Basin at about 4:00 AM. She reached Macclesfield Bridge at about 4:55.

The subsequent Board of Trade inquiry, chaired by Colonel Sir William Yorke Moore, R.E. and reporting on 21 October 1874, took evidence from thirty-seven witnesses over eleven days at the Marylebone police court. The most probable cause, in the Board's finding, was that a spark from the Ready's boiler funnel had drifted aft on the light morning wind, had ignited a small quantity of benzoline vapour escaping from an imperfectly-bunged cask on the Tilbury's deck, and that the resulting flash-fire had reached the powder in the hold within seconds. The Board could not establish this beyond doubt because there were no surviving witnesses aboard the Tilbury and the physical evidence had been reduced to fragments. The Board's second finding was that the practice of towing five-barge trains loaded with gunpowder past the largest concentration of civilian residences on the entire canal system, at any hour of the day or night, without segregation of the powder-barge from the tug, without warning to the surrounding parishes, and without written authorisation from anyone at all, was — the Board's exact word — "indefensible." The Explosive Substances Act of the following year (38 & 39 Vict. c. 17, receiving royal assent 14 June 1875) rewrote the rules for the carriage of gunpowder by inland waterway; it is still in effect, in amended form, today.

The bridge itself was rebuilt over the winter of 1874–75 by the Coalbrookdale Company of Shropshire, which had cast the original in 1815 and still held the pattern in stock. The company recovered four of the original six columns from the canal bed, dressed them for reuse, and re-cast the other two. The bridge reopened to horse traffic on 12 April 1875. Whether the reversal of the columns was a deliberate cost-saving move on the part of the contractor (the worn faces were structurally sound and would wear down further if re-exposed to the towpath side, at which point the bridge would eventually need a third rebuild) or a simple installation error nobody bothered to correct is a matter of dispute in the local histories. The bridge's official name is Macclesfield Bridge; its Camden name is Blow-up Bridge; and the tow-rope grooves on the canal-facing side of the columns are its calling card. They can be seen from the towpath today, running clean and vertical down the outside of the columns, on the wrong side.

London Zoo — whose main gate is about two hundred yards from the bridge — reported the explosion in the Zoological Society's Proceedings for November 1874. The glass in the reptile house shattered; the roof of the small-mammal house partially collapsed; the sea-lion pool cracked and drained overnight into the canal. The Zoo's casualty list came to "one small kangaroo, expired of shock the following day; two agoutis, cause unascertained;" no other animal is recorded as having died. The wolves reportedly howled through the rest of the morning. The giraffes are reported to have remained standing at the same feed rack throughout the event and to have declined to acknowledge that anything had happened.

Primary sources:

Moore, Sir William Yorke, R.E. Report of the Board of Trade Inquiry into the Explosion of the Barge Tilbury on the Regent's Canal on the 2nd October 1874. Presented to both Houses of Parliament by Command of Her Majesty. London: HMSO, 21 October 1874. C.-1067. Contains the witness transcripts, the surveyor's site sketch of the destroyed bridge, and the Board's five findings, of which the fourth is the "indefensible" clause quoted above.
The Times, Saturday 3 October 1874, p. 12, cols. 3–5, "Terrible Explosion on the Regent's Canal." The next-morning report, including the first published statement of the crew's names and the earliest published estimate of the powder tonnage. Follow-up reporting through October is indexed in the Times's own annual index for 1874 under Regent's Canal — Explosion.
Illustrated London News, vol. LXV, no. 1828, 10 October 1874, pp. 340–341. A double-page wood-engraved plate of the destroyed bridge from the towpath side, drawn from a sketch made on the spot by the paper's staff artist on the afternoon of 2 October. The engraving shows the crater in the towpath, the pattern of the fallen iron, and — visible in the middle distance — a knot of gongoozlers already gathered along the far bank to watch the salvage.

A dev fact for the back pocket

In August 1978, in volume 21 issue 8 of Communications of the ACM, the Association for Computing Machinery published a paper titled "Can Programming Be Liberated from the von Neumann Style? A Functional Style and Its Algebra of Programs." The paper was the 1977 Turing Award lecture of its author, John Backus of IBM Research, delivered at the ACM Annual Conference in Seattle the previous October. It argued, at thirty pages and with a full worked mathematical apparatus, that FORTRAN, ALGOL, PL/I, C, and every general-purpose programming language then in wide use — all of them — were structurally unfit for programming, that the notion of the assignment statement was itself a design mistake, and that the field ought to abandon the entire paradigm and rebuild on a small algebra of composable functions with no variables. The author of the paper had invented FORTRAN.

John W. Backus (born Philadelphia, 1924) had joined IBM in 1950 as a programmer working on the IBM 701 in the SSEC building on Madison Avenue. He led the team that produced the first FORTRAN compiler — for the IBM 704 — between 1954 and 1957; the compiler is generally credited as the first practical optimising compiler, and FORTRAN as the first widely-used high-level programming language. In 1959 he chaired the syntactic-notation subcommittee of the ALGOL 60 committee and, at the June 1959 meeting in Paris, presented the metasyntactic notation he had devised for the language definition. Peter Naur — a Danish astronomer working on the committee — revised the notation into the form used in the final ALGOL 60 report of January 1960, and the notation has been known as Backus-Naur Form since (the naming is due to Donald Knuth, in a letter to the editor of CACM in December 1964). The 1977 Turing Award citation named both contributions specifically: "for profound, influential, and lasting contributions to the design of practical high-level programming systems, notably through his work on FORTRAN, and for seminal publication of formal procedures for the specification of programming languages."

Backus delivered the lecture on 26 October 1977 at the ACM Annual Conference at the Olympic Hotel in Seattle. Its opening is unusually direct for the genre and worth reading in the original:

"Programming languages appear to be in trouble. Each successive language incorporates, with a little cleaning up, all the features of its predecessors plus a few more. Some languages have manuals exceeding 500 pages; others cram a complex description into shorter manuals by using dense formalisms… The purpose of this paper is to suggest that we have reached a stage of language design analogous to that of plane geometry before Euclid: we have a mass of ad hoc results, but no unifying framework in which to relate them and no clear direction in which to proceed. What is needed is not another set of ad hoc improvements but a fundamental change of approach."

The von Neumann bottleneck — his coinage — is the argument's fulcrum. Backus observes that a modern computer's CPU and its store are connected by a bus over which a single word may be transferred per read or write. Programming languages descended from the von Neumann architecture (in his terminology: essentially all of them) inherit this bottleneck as a semantic feature. They force the programmer to reason about state one word at a time, one assignment at a time, and to schedule the order of assignments manually. "Surely there must be a less primitive way of making big changes in the store than by pushing vast numbers of words back and forth through the von Neumann bottleneck. Not only is this tube a literal bottleneck for the data traffic of a problem, but, more importantly, it is an intellectual bottleneck that has kept us tied to word-at-a-time thinking instead of encouraging us to think in terms of the larger conceptual units of the task at hand." (CACM 21:8, p. 615.)

The alternative he proposed was FP — a language with no variables, no assignment, and no explicit state. FP programs are built from a fixed set of primitive functions (arithmetic; selectors like first and tail; structural operators like transpose and distribute-left) combined by a fixed set of higher-order program-forming operations (composition ∘, apply-to-all α, insert /, condition →, constant $, construction […]). A whole FP program is written as a single expression evaluating to a single function; the language admits an algebra of programs — a set of proven algebraic identities — that lets the programmer or the compiler rewrite one expression into an equivalent one by pure syntactic transformation. Backus's worked example, in Section 12 of the paper, is a matrix-multiplication routine expressed as a single line-and-a-half function composition, whose correctness he then proves by algebraic manipulation over the following half-page.

FP itself never had a substantial user community. Its direct descendants — FL (Backus's own follow-up at IBM Almaden through the 1980s), Miranda (David Turner, University of Kent, 1985), and thereby Haskell (first Haskell committee report, April 1990) — became the vehicle by which its ideas reached the mainstream. Every non-strict pure functional language shipped since owes the paper direct debts of vocabulary and framing. The larger effect was rhetorical. After 1978, the position that the imperative style might be the wrong style for programming was one that a serious researcher could hold without embarrassment.

The paper is thirty pages, densely argued, and largely readable today. It is one of the most-cited Turing Award lectures on record; Dijkstra's 1972 The Humble Programmer is the only serious competitor. Backus retired from IBM in 1991 and died in 2007 at eighty-two. He is understood not to have written a FORTRAN program of his own after about 1958.

Primary sources:

Backus, John. "Can Programming Be Liberated from the von Neumann Style? A Functional Style and Its Algebra of Programs." Communications of the ACM, vol. 21, no. 8, August 1978, pp. 613–641. The 1977 ACM Turing Award Lecture as delivered on 26 October 1977 in Seattle, revised by the author for publication. The FP language definition begins at Section 11 (p. 624); the algebra of programs at Section 12 (p. 630).
Association for Computing Machinery. Turing Award Citation for John Backus, 1977. Reprinted at the head of the CACM printing above and available in facsimile from the ACM's Turing Award pages.
Naur, Peter (editor). "Revised Report on the Algorithmic Language ALGOL 60." Communications of the ACM, vol. 6, no. 1, January 1963, pp. 1–17. The definitive published statement of ALGOL 60 using the syntactic notation Backus first proposed at the Paris meeting of June 1959. The metasyntax used in the report is Backus-Naur Form; the naming is due to Knuth in CACM vol. 7 no. 12, December 1964, p. 735.

Today's goal

Compliment one thing today that friend would ordinarily not compliment. Not a person's outfit; not a colleague's demo — those are the easy ones. A cornice on a building friend passes twice a day and has never registered. The way the top of a specific tree has been cut back and grown out asymmetrically. A well-kerned street sign. The green of a particular stranger's cardigan. The way the barista's second latte of the morning came out with a leaf on top and the fifth one came out with a heart and neither was on purpose. Notice it, note it aloud if there is anyone to note it to, and if there is not, note it silently to the object itself.

The point is not the object. The point is the shift in attention — from ambient to deliberate — between the moment friend thinks "that is well-done" and the moment friend can name what specifically is well-done about it. That's a muscle. It goes slack when unused. The coxcombs of the Restoration were bad at it because they were only ever paying attention to their own reflected image; the corrective is exactly this, one act today of taking a specific piece of the world seriously.

Today's toy is the coffee-house — a small early-eighteenth-century satirical periodical in the manner of Steele's Tatler and Addison's Spectator, in which each issue's front-page sketch is a fresh coxcomb, generated to order. Wig, address, defining vanity, defining absurdity, and one coffee-house he is expected at this evening; underneath, three provincial bulletins and three small advertisements. Set a new issue, or click any single item to reset just that item. Lives in the corner.

Go compliment something small, friend.

— C