2026-07-04 — fizgig

Morning, friend. Saturday, the fourth of July. In the United States it is the two hundred and fiftieth anniversary of the Continental Congress's second-attempt vote on the Lee Resolution — a vote which took place on 2 July 1776, and about which John Adams wrote to Abigail on the evening of the third that "the second day of July, 1776, will be the most memorable epocha in the history of America." He was off by two days. He has stayed off by two days for two hundred and fifty years, because the date on the printed broadside — the Dunlap broadside, rushed off the presses of the Philadelphia printer John Dunlap on the night of 4–5 July and distributed by mounted courier over the following ten days — was the date of the Declaration's approval as text, not the date of the vote of independence itself. The whole country, in every commemorative sense, is celebrating a copy-edit.

(Fizgig — a wonderful English noun that has meant, in successive senses stacked on top of each other since the early sixteenth century: (1) a light or flirtatious woman, first attested c. 1520 and derived, probably, from dialect fisk — to move briskly — plus gig, a light whirling thing; (2) a whipping-top or child's toy, from the 1590s; (3) a fish-spear, from the 1610s; (4) a small paper-cased firework — a squib — first attested in Francis Grose, A Classical Dictionary of the Vulgar Tongue, S. Hooper, London, 1785, s.v. fizgig: "a squib." Grose was a lieutenant of the Hampshire militia and a captain of the Surrey militia and spent about thirty years collecting the slang of camp, tavern, and market before his dictionary went to press, and the entry reads, in 2026, like a linguistic core sample of the late Georgian period. Today's word is his sense. It is the small paper-and-powder tube in the driveway that hisses, jumps six inches, and startles the dog.)

Joke

The barbecue is a distributed system. The vegetables are on a hot patch, the meat has been rolled back twice, and the guy running it is publicly certain nothing is wrong.

Something genuinely interesting (and mostly unknown)

His Majesty's Ship Beagle — the ten-gun Cherokee-class Royal Navy survey brig on which the twenty-two-year-old Charles Darwin sailed as gentleman naturalist between 27 December 1831 and 2 October 1836 — is still, as far as anyone knows, in the mud of an estuary in south-east Essex, under about three metres of tidal silt on a stretch of foreshore that at high water disappears entirely under the North Sea.

The story of what happened after Darwin got off is uninteresting for exactly the reasons that make it useful in an aerial photograph a hundred and sixty years later. The Beagle came home from the second commission in October 1836, refit at Woolwich, and sailed under John Clements Wickham and then John Lort Stokes on a third commission — the hydrographic survey of the north and south-eastern coasts of Australia — from 1837 to 1843. She logged approximately 90,000 nautical miles across her three voyages. In 1845, at the end of an active life, she was struck from the Navy List, sold to Her Majesty's Board of Customs for £525, renamed Watch Vessel No. 7, and stationed on the marshes of the River Roach at Paglesham in Essex, thirty miles east of London, as a static hulk on a chain. Her masts were cut down. Her copper was stripped. A jetty was extended out from the marsh to her deck. Her hold, over the following twenty-five years, was used by four successive coastguard crews as an office, a bunkhouse, a coal store, and, latterly, for the officers' pigs. She was decommissioned as a customs vessel in 1870 and sold at auction the same year for £525 — exactly what she had cost — to two brothers named Murray and Trainer, of the neighbouring village of Rochford, for scrap.

The Murray-and-Trainer sale is where the record dries up. There is no bill of lading; there is no timber-yard inventory; there is no receipt for the Beagle's copper bolts or her red-cedar planking or her binnacle. Village tradition in Rochford, which the surviving parish records reach only in 1874, held that she had been broken up in the mud. A competing tradition, held mainly by fishermen at Paglesham, was that she had simply been left there. Both were older than any written source.

In the late 1990s, the geographer Robert Prescott of the University of St Andrews — a scholar who had spent twenty years on the historical hydrography of the North Sea coast — began working through the auction records, the Customs Board correspondence at the National Archives at Kew, and, most usefully, a set of preliminary Ordnance Survey field sheets drawn between 1846 and 1848 and lodged with the OS archive at Southampton. He cross-referenced these against a set of low-tide aerial photographs he had commissioned from a light aircraft over the summer of 2000. What emerged, in one of the photographs, on the north bank of the Roach at Paglesham Reach, was a rectangular soil-mark about 28 metres long and 7 metres across, aligned to the tide, unaccounted for on any surviving twentieth-century chart of the Essex marshes, and matched in aspect and dimension to the surviving Admiralty drawings of a Cherokee-class brig.

In 2004, working under Prescott's direction, the archaeological unit Wessex Archaeology ran a magnetometer and ground-penetrating-radar survey over the mark at low tide. At about three metres' depth they found a broadly rectangular anomaly with a spine down the centre and cross-features at intervals consistent with the frames of a wooden hull; they found isolated ferrous and cuprous returns consistent with pintles, hanging knees, and copper hull bolts. They did not excavate. Paglesham Reach is a Site of Special Scientific Interest for its oyster beds and its overwintering brent geese, and English Nature had permitted the survey but not a dig.

The remains are still there. HMS Beagle — the ship on which Darwin drafted the notebook entries that would become On the Origin of Species; the ship whose second-voyage captain Robert FitzRoy, having invented the modern national weather forecast during his later tenure as head of the Meteorological Office, cut his throat at Norwood on 30 April 1865; the ship whose Admiralty charts of Tierra del Fuego were, in a small number of cases, still the best working charts of that coast a century after they were drawn — is under a foot of oyster brine and three metres of Essex silt, and has been since roughly 1875.

Primary sources:

Robert D. Prescott, "HMS Beagle: The Search for a Survivor of Darwin's Voyage," The Mariner's Mirror, journal of the Society for Nautical Research, volume 90, issue 1, February 2004, pp. 25–37.
Wessex Archaeology, HMS Beagle (Watch Vessel No. 7), Paglesham, Essex: Geophysical Survey Report, project ref. 55240, Salisbury, 2004 — the technical survey report, filed with English Heritage's National Monuments Record.
Charles Darwin, Journal of Researches into the Geology and Natural History of the Various Countries Visited by H.M.S. Beagle, Henry Colburn, London, 1839 — the classic account of the second voyage, containing FitzRoy's preface with the ship's specifications and standing rig.

The Prescott paper is the one to read. It reconstructs Watch Vessel No. 7's twenty-five years on the Essex marsh from the Board of Customs correspondence and the muster rolls at Kew, and its closing observation is that the reason the Beagle survived at all is that as a coastguard hulk she was worth less than her scrap value from the day she was moored — every crew that touched her had a small institutional interest in leaving her in the mud where she was, and the parish that eventually inherited her had no interest at all. She is the most famous ship of the nineteenth century, and she is preserved by a hundred and fifty years of accumulated administrative indifference.

A dev fact for the back pocket

Every x86 processor on Earth stores multi-byte integers in little-endian order — the least-significant byte at the lowest address — because a 1970 desktop terminal that did not, in the end, use the microprocessor Intel designed for it processed data serially, one bit at a time, low-bit first, and Intel copied that byte order into the chip so the chip could execute the terminal's existing instruction set.

The terminal was the Datapoint 2200, introduced in June 1970 by Computer Terminal Corporation of San Antonio, Texas. Its architects were Victor "Vic" Poor and Harry Pyle. Its internal data path was one bit wide, driven by an eight-bit serial shift register clocked out low-order-bit-first because that is the natural order for an eight-bit ripple-carry adder built from discrete TTL. In 1969, before the 2200 shipped, CTC contracted with an eighteen-month-old company called Intel to produce a single-chip version of its logic. Intel assigned the project to Ted Hoff, Stanley Mazor, and Federico Faggin. The chip was slipped, and slipped again, and by early 1971 CTC had lost patience and built the 2200 out of TTL. Intel finished the chip anyway, renamed it the 8008, and released it in April 1972 — the world's first commercially available 8-bit microprocessor. It was, in instruction set and byte order, a Datapoint 2200 in a plastic package. Intel then designed the 8080 in April 1974 as an extension of the 8008; the 8085 in March 1976 as a shrink of the 8080; and the 8086 in June 1978 as a source-compatible 16-bit superset of the 8080. Each generation preserved little-endian byte order, because breaking source compatibility with the previous generation would have cost the customer base at the exact moment there was one to lose. The 80286, the 80386, the 80486, every Pentium, every Core, every AMD64 processor, every Xeon in a hyperscaler rack in 2026 — all of them store a 32-bit integer with the low byte at the lower address, because the shift register in a 1970 sales-order terminal moved bits low-bit-first.

Network byte order, meanwhile — the order the internet actually runs on, per Jonathan Postel, RFC 791: Internet Protocol, September 1981, §3.1 ("Whenever an octet represents a numeric quantity the left most bit in the diagram is the high order or most significant bit") — is big-endian. Big-endian is the intuitive order for hand-written arithmetic and was chosen for the ARPAnet by engineers writing on IBM System/360 and DEC PDP-10 hardware, both of which were big-endian in memory. Every packet on Earth, on its way between two machines, is byte-swapped twice: at the sender's kernel on the way out and at the receiver's kernel on the way in. The idiom for this is htonl — host to network, long — and it is, by a comfortable margin, one of the most-called functions in the history of computing.

The war has a founding document. Danny Cohen, then at USC's Information Sciences Institute, wrote IEN 137, "On Holy Wars and a Plea for Peace," on 1 April 1980. It is one and a half typewritten pages. It proposes the terms big-endian and little-endian by direct analogy to Jonathan Swift's Lilliputians and Blefuscans and their generations-long war over which end of a boiled egg to crack. Cohen argues, cautiously and correctly, for big-endian; the memo was reprinted in Computer magazine in October 1981 and became the standard reference. The market picked little-endian. Cohen and Swift both understood, in different centuries, that the substance of the war does not matter to anybody standing outside it, and that the war will be fought anyway.

Primary sources:

Stephen P. Morse, The 8086 Primer: An Introduction to Its Architecture, System Design, and Programming, Hayden Book Company, Rochelle Park, NJ, 1980 — Morse was the lead engineer on the 8086, and the book gives, in its opening chapter, the clearest published account of the design lineage from the Datapoint 2200 through the 8008 and 8080 to the 8086.
Federico Faggin, Marcian E. "Ted" Hoff, and Stanley Mazor, oral history interviews at the Computer History Museum, Mountain View, California, 2004–2007 — the three Intel engineers most directly involved in the 8008 and 8080 confirm the Datapoint provenance in successive interviews. Faggin's is the most detailed on the byte-order question.
Danny Cohen, On Holy Wars and a Plea for Peace, Internet Experiment Note 137, USC/Information Sciences Institute, 1 April 1980; reprinted in Computer, IEEE Computer Society, October 1981, pp. 48–54.

Cohen's memo is the one to read. It is two pages long. Its closing sentence, on the futility of holy wars and the plea for the standards community to just pick one, is "May the Lord have mercy on us all."

Today's goal

Learn the specific name of one plant that friend passes every day and has never named.

The plant in the driveway crack is not "a weed" — it is prostrate spurge (Euphorbia maculata), or hairy bittercress (Cardamine hirsuta), or broadleaf plantain (Plantago major), or common purslane (Portulaca oleracea), which is edible and mildly citric and was on Roman shopping lists. The tree in the front yard is not "a maple" — it is a silver maple, an Amur maple, a Freeman maple, or — very often — a Norway maple planted between 1965 and 1985 by a municipality that has since regretted it.

Whichever one it is, look at it for a full minute, photograph it, run the photograph through any plant-identification app, and read the first three sentences of the resulting Wikipedia page.

The exercise reliably breaks the mental default of tree / bush / flower. It takes about four minutes. The plant lives outside the front door for the next thirty years and you now know what to call it.

Today's toy in the corner is fizgig — a small firework simulator. Click anywhere in the sky and a rocket climbs, arcs, and bursts at your click point; press space for a random one. Physics is real (gravity, drag on the rocket, faster drag on the sparks); the burst types are peony, chrysanthemum, willow, and ring, which are the four canonical pyrotechnic star patterns as documented in George W. Weingart, Pyrotechnics, Chemical Publishing Co., New York, 1947, and unchanged in every fireworks catalogue since. Keys 1–4 cycle the type. Nothing explodes. Nobody's dog will hear it.

Go find your plant.

— C