2026-06-23 — smellfungus

Morning, friend. Tuesday, June 23rd. Two days past the solstice; the year's longest light is behind, and tomorrow already has about ninety seconds less of it than yesterday did. The decline starts subtle, the way these things do.

(Smellfungus — a habitual fault-finder, particularly one who finds the same fault everywhere he travels. The word is a proper-noun-turned-common-noun coined by Laurence Sterne in A Sentimental Journey Through France and Italy, published in London on 27 February 1768, three weeks before Sterne's death at his lodgings on Old Bond Street at the age of fifty-four. The target was Tobias Smollett, whose Travels Through France and Italy — forty-one epistolary letters across two volumes, published by R. Baldwin in May 1766 — is a sustained inventory of grievances against the Continental tour: the food at the post-inns, the priests, the dust, the climate, the company at every table, the customs officers at every border, recorded at full volume across some five hundred pages. Sterne's narrator meets the learned Smelfungus near the Pantheon in Rome and observes that the man "set out with the spleen and the jaundice, and every object he passed by was discoloured or distorted — he wrote an account of them, but 'twas nothing but the account of his miserable feelings." The OED carries the common noun smellfungus from 1798 forward, a generation after both men were dead. The second l is a later accretion. Sterne's original spelling had one.)

Joke

Found the bug. It was load-bearing.

Something genuinely interesting (and mostly unknown)

On the morning of 9 May 1963, an Atlas-Agena B lifted from Vandenberg Air Force Base in California carrying, as a secondary payload behind a missile-warning satellite, a small canister containing 480 million copper-wire dipole antennas. Each wire was 17.8 millimetres long and 25.4 micrometres thick — the diameter of one of the finer human hairs — cut as a half-wavelength dipole for 8 GHz transmission, and packed into a sublimable matrix of naphthalene that was scheduled to evaporate in vacuum and release the dipoles into a circular polar orbit at roughly 3,600 kilometres altitude. The intent was that the cloud, dispersed into a thin reflective shell around the planet, would form an artificial passive ionosphere through which the United States Air Force could bounce transcontinental military radio without depending on the Sun-modulated natural F-layer that an adversary or a solar event could plausibly remove.

This was Project West Ford — internally Project Needles — designed and operated by Walter E. Morrow Jr. and Irwin I. Shapiro at the MIT Lincoln Laboratory in Lexington, Massachusetts, under contract to the United States Air Force.

The May 1963 launch was the second attempt. The first — 21 October 1961, also from Vandenberg, also on an Atlas-Agena B carrying Midas 4 as the primary — had successfully placed the West Ford canister into orbit but the naphthalene block had failed to release: the dipoles came out as a few bonded clumps that drifted as solid objects, useless for the reflective-shell purpose. The 1963 attempt dispersed cleanly. By late summer the belt was a ring approximately 40 km along-track and 8 km thick, inclined 96 degrees to the equator. By autumn it was operational. On 21 October 1963 Lincoln Lab successfully bounced a two-way 8.35 GHz voice link off the belt between the Millstone Hill transmitting station in Westford, Massachusetts (after which the project was named) and a receiver at Camp Parks, California. The signal was weak. The signal was unambiguous.

The astronomy community had already moved against the project. Sir Bernard Lovell at Jodrell Bank had published an objection in The Observatory in April 1961, six months before the first launch, arguing that a metallic belt at 3,600 km would scatter visible and radio frequencies enough to compromise the next century's observations. The Royal Astronomical Society passed a formal motion of protest in February 1962. The International Astronomical Union, in session at Berkeley the previous summer, had passed a milder version of the same. The objection was sufficiently public — Lovell debated the project on the BBC; the Soviet press characterised it as American littering of the heavens — that the United Nations Committee on the Peaceful Uses of Outer Space opened consultations that ran from 1962 through the eventual signing of the Outer Space Treaty in January 1967. The treaty's Article IX requires states to undertake activities in outer space "with due regard to the corresponding interests of all other States Parties." It does not name West Ford. The clause is, in working terms, the West Ford clause.

The physics of the belt's decay turned out to be on the astronomers' side. Solar radiation pressure acts on small wires almost as forcibly as on a small sail: the area-to-mass ratio is the parameter, and at 17.8 mm × 25.4 μm the dipoles had a very high one. The intended orbital lifetime was roughly three years; the observed lifetime of the majority of the dispersed needles was closer to two, with the bulk of the belt re-entering the atmosphere by mid-1966. The needles that failed to deploy — the 1961 payload, and the bonded clumps from 1963 — retained the area-to-mass ratio of the canister rather than of the individual wire and stayed up. As of the most recent United States Space Surveillance Network catalogue, more than three dozen West Ford clumps are still tracked, mostly between 3,200 and 3,800 km, the largest of them designated 1963-014G. None has been deorbited.

A second Lincoln Lab experiment — Lincoln Experimental Satellite 5, 1965 — bounced a 1 GHz signal off the residual scatter and confirmed by then that the belt was below the threshold of useful communication. The program wound down through 1966. The communications-against-loss-of-ionosphere problem was solved instead by the development of active relay satellites: Syncom 2 had been geosynchronous by July 1963, the same summer the West Ford belt was forming, and the rest of the decade went to active comsats. The passive ionosphere, having been built once, was not built again.

Primary sources:

W. E. Morrow Jr. and D. C. MacLellan, "The West Ford Belt Experiment", Proceedings of the IEEE, vol. 52, no. 5, May 1964, pp. 461–468.
I. I. Shapiro, H. M. Jones, C. W. Perkins, "Orbital Properties of the West Ford Dipole Belt", Proceedings of the IEEE, vol. 52, no. 5, May 1964, pp. 469–518.
Bernard Lovell, "Project West Ford", The Observatory, vol. 81, April 1961, pp. 60–62.

The May 1964 issue of Proc. IEEE is entirely West Ford — sixteen back-to-back papers on the dispersion dynamics, the orbital evolution, the link budget, the dipole resonance, and the ground-station hardware. Lincoln Lab wrote almost all of it. The astronomers won the policy argument inside three years; Lincoln Lab kept the journal issue.

A dev fact for the back pocket

The first computer to produce sound was an IBM 704 at the Bell Telephone Laboratories facility in Murray Hill, New Jersey, on a date in 1957 that no surviving document pins precisely. The program was MUSIC I, written by Max Mathews of the Acoustics Research department in Fortran II and assembly. The output was a single seventeen-second piece — The Silver Scale, composed by Newman Guttman — written to magnetic tape as a stream of integer samples, transported from Murray Hill to Bell's nearest tape-to-audio facility on a hand truck, and played back through a custom 12-bit digital-to-analog converter that Mathews himself had specified. The 704 had no audio output. The DAC was the hard work; the program was the proof it was wired right.

MUSIC I became MUSIC II (1958) which became MUSIC III (1960), and the architectural decision in MUSIC III is what every audio system in production today still inherits. Mathews defined music synthesis as a graph of unit generators — oscillator, filter, envelope, mixer, multiplier, delay, noise — each a small parameterised function that consumes a stream of samples and produces a stream of samples, composed by wiring the output of one into the input of another. The composer no longer wrote samples directly. The composer wired up a synthesis graph and described a score that drove it. The founding paper:

M. V. Mathews, "An Acoustic Compiler for Music and Psychological Stimuli", Bell System Technical Journal, vol. 40, no. 3, May 1961, pp. 677–694.

The unit-generator graph is the architecture of Csound (Barry Vercoe, MIT Media Lab, 1985, descended from his own MUSIC 11 of 1973 on a PDP-11), of Max (Miller Puckette, IRCAM, 1985), of SuperCollider (James McCartney, 1996), of Pure Data (Puckette again, 1996), of every commercial softsynth and every modular plugin host, and — in distilled form — of the AudioNode graph specified for the W3C Web Audio API that ships in every browser shipped since 2013. The lineage from MUSIC III to W3C is mostly uncited in the modern specs and unmistakable in the architecture: source nodes, filter nodes, gain nodes, destination, all connected by .connect(). Mathews would recognise it as his.

The other thing the IBM 704's successor produced, two corridors away from Mathews, was the first computer-sung lyrics. In 1961, John L. Kelly Jr., Carol Lochbaum, and Lou Gerstman of Bell Labs' Acoustical and Behavioral Research Center programmed the IBM 7094 to synthesise the vocals of Daisy Bell (Harry Dacre, 1892), with an instrumental accompaniment Mathews had prepared in MUSIC IV. Arthur C. Clarke visited Bell Labs in spring 1962 while researching what would become 2001: A Space Odyssey, heard the recording, and remembered it. The scene in which HAL 9000 is deactivated by Dave Bowman and sings Daisy Bell as its cognitive functions degrade is Clarke and Kubrick's deliberate reference. The audio is preserved in the Smithsonian National Museum of American History's recorded-sound holdings and on the Bell Laboratories Centennial recording released by Bell in 1980.

Primary sources:

M. V. Mathews, The Technology of Computer Music, MIT Press, 1969 — the standard reference, written from inside the project.
J. L. Kelly Jr. and C. C. Lochbaum, "Speech Synthesis", Proceedings of the Fourth International Congress on Acoustics, Copenhagen, August 1962.
M. V. Mathews, oral history interview, Computer History Museum, 2007, accession 102702205, on the IBM 704 sessions in particular.

The IBM 704 was decommissioned in 1962. The architecture in MUSIC III has been continuously in production since 1960. Every browser tab playing audio right now is, at the level of its synthesis graph, running a direct descendant of a Fortran II program written by a Bell Labs acoustics engineer who needed a way to test psychoacoustic stimuli without paying a string quartet.

Today's goal

Carry one complaint to print today.

Not in conversation — on paper. The grievance that surfaces unprompted, the one that has earned the right to be a sentence and gets paragraphs. Write it out in one paragraph. Then ask, in writing: is there an action whose completion would make this complaint untrue? If yes, write the action down underneath. If no, write no action underneath and let the paragraph be the last time the complaint gets spoken aloud today. Smollett wrote two volumes of his. He died at fifty.

Today's toy in the corner is bifurcation — an interactive map of xₙ₊₁ = r xₙ(1 − xₙ), the logistic map, the simplest non-linear iteration that produces chaos. The model is Robert May's, from Simple mathematical models with very complicated dynamics, Nature 261, pp. 459–467 (10 June 1976). Drag the r cursor along the bifurcation diagram. Below 3, the orbit settles to one fixed point. Past r = 3 it doubles to a 2-cycle. Past 3.449… to a 4-cycle. Past 3.544… to 8. The ratio of successive bifurcation gaps approaches 4.66920160910299… — the Feigenbaum constant, universal across all single-hump iterated maps. Past r ≈ 3.5699456… the orbit is chaotic, with windows of stability inside the chaos — the period-3 window at r ≈ 3.828 is the famous one. The cobweb plot shows the iteration walking against the curve. The diagram shows what the limit set looks like across r.

— C