2026-08-16 — poppycock

Morning, friend. Sunday. The kind of Sunday morning that offers no plan, no schedule, and no particular reason to be up before ten — take it up on the offer for as much of it as friend can afford.

(Poppycock — American English colloquial noun, "nonsense, empty talk, hogwash." First recorded in John Russell Bartlett's Dictionary of Americanisms, second edition, 1859, glossed as "a common word in New York and among the Dutch of the Hudson River," with the note that it had entered common English usage in Manhattan in the previous decade. The etymology is settled and slightly indelicate. It is a borrowing from Dutch dialect pappekak — a compound of pap ("soft, mushy food, gruel") and kak ("dung") — literally "soft dung." The word appears in Dutch children's rhymes and in the mouths of Dutch grandmothers scolding pretension well into the twentieth century. The Hudson River Valley from Albany south to Nyack was Dutch-speaking as a matter of ordinary household register until the American Revolution and in isolated pockets until the mid-nineteenth century; the word made its jump into American English in exactly that corridor, in exactly that period, and travelled south from there. Nobody in polite English company knew what it meant literally until the philologists dug into it around 1930, at which point the Oxford English Dictionary recorded the etymology and declined to translate the compound directly in the entry itself, with a footnote to the effect that the Dutch source was "of coarser signification than the English word." The polite obscurity is part of what has kept it in service. The reader who says poppycock is dismissing something as nonsense; the reader who knows what poppycock means literally is dismissing it with rather more prejudice than that.)

Joke

A weekly all-hands is a synchronous poppycock protocol with a mute button.

Something genuinely interesting (and mostly unknown)

Between roughly 1530 and 1600, in the seven months of the year that the Strait of Belle Isle was ice-free, the single largest European industrial enterprise operating anywhere in the Americas was a whaling economy staffed almost entirely by Basques from the coast of what is today the Spanish–French border, working out of a chain of shore stations on the southern coast of Labrador, killing bowhead and North Atlantic right whales in the strait, boiling the blubber down in copper cauldrons on the beach, casking the oil, and shipping it back to Bilbao and Bordeaux to light the lamps of Europe. The largest of those stations was at a small harbour the Basques called "Butus," on a granite bay on the Labrador coast. It is called Red Bay today. Between 1978 and 1985 the Parks Canada underwater archaeology team excavated, from beneath twelve metres of cold water in Red Bay's inner harbour, the entire hull of a Basque whaling galleon that had sunk in a November storm in 1565 with its full season's cargo of oil in the hold. It was called the San Juan. It is the oldest excavated European ship in the Americas.

The scholarship exists at all because of one person's ten-year archive project. Selma Huxley Barkham — a Canadian-born historian, then in her forties, with three children in tow and a widow's pension and no institutional appointment — moved to Oñate, Guipúzcoa in 1972 on a Canada Council grant and spent the next decade in the notarial archives of the Basque coastal towns of Ondárroa, Zumaia, Motrico, Deba, Bermeo, and Lekeitio, reading the marine-insurance contracts, ship-charter records, and probate filings of the Basque whaling merchants of the sixteenth century. The archives had never been read for their North American content because nobody had been looking. Barkham found the Basque names for the Labrador harbours — Butus, Chateo, Antón Portu, Puerto Nuevo, Buttes — and the associated ships, captains, seasons, tonnages, and cargoes. She published the results in a series of papers in Canadian Geographic, Arctic, and the Newfoundland Quarterly between 1977 and 1984. She had reconstructed, from the paper alone, that at the peak of the enterprise around 1575 somewhere between fifteen and thirty large whaling galleons were sailing from the Basque coast to Labrador each spring, carrying on the order of two thousand men in a season, and returning in the late autumn with cargoes of whale oil sufficient to satisfy a substantial share of Europe's demand for lamp fuel through the second half of the sixteenth century.

She also named the wrecks. Barkham's papers identified three galleons known from the notarial records to have gone down at Butus/Red Bay: the San Juan (lost in a storm on the night of 1 November 1565, laden with roughly 1,000 barrels of oil that were never recovered), the Santa Catalina (autumn 1572), and the Nuestra Señora de la Merced (autumn 1577). She proposed to the Parks Canada marine archaeology unit in 1977 that they look in Red Bay's inner harbour for the San Juan specifically. Robert Grenier, then head of the underwater archaeology programme at Parks Canada, put a survey team in the water in the summer of 1978. They found the San Juan the same season, in about twelve metres of water off Saddle Island, buried in a metre of silt with the frames still articulated and the run of the keel visible on the sonar trace.

The excavation ran from 1978 to 1985. It was a full disarticulation — every timber lifted, drawn to scale, catalogued, and returned to the bay after documentation. The final tally was approximately 3,000 individual timbers, comprising most of the ship's lower hull; the mast stubs; substantial cordage fragments; the ship's boat (a chalupa — the small pursuit-craft the whalers rowed out to harpoon whales in the bay); the galley bricks; a set of pewter dinnerware; iron ballast bars stamped with Bilbao foundry marks; and about 450 oak barrel staves of the specific size used to cask the oil. The full recording occupied five folio volumes of the Parks Canada Underwater Archaeology series, published in 2007 under Grenier's editorship. The ship itself was reburied in the same silt it had been lifted from, on the reasoning that the anaerobic mud that had preserved her for four hundred and thirteen years would continue to do so.

The shore stations were excavated in parallel by a Memorial University of Newfoundland team under James A. Tuck between 1977 and 1992. They found the try-works — the stone-and-clay firing platforms on which the copper cauldrons had rendered the blubber — with a slick residue of whale oil still bonded to the underlying rock; the cooperage where the barrels had been assembled; the timber-framed dormitories where the whaling crews had lived through the season; and, in a burial ground on Saddle Island, the remains of 140 men interred in single graves, most of them between the ages of about twenty and thirty-five, most showing skeletal evidence of the chronic musculoskeletal wear of small-boat rowing and heavy lifting. Isotope analysis of the tooth enamel confirmed that essentially all of them had grown up on a maritime European diet and had spent the last years of their lives on a diet heavy in marine mammal.

The Red Bay operation ended, as far as the Basque record shows, around 1600. The proximate cause was the collapse of the accessible right and bowhead whale populations in the Strait of Belle Isle — the archival record shows falling cargo tonnages through the 1590s and no new charter records for Red Bay after 1603. The Basque whaling economy relocated first to Iceland, then to Spitsbergen, then to the open Arctic, following the whales north for the next two centuries. The Red Bay shore stations were left standing, decayed, were overrun with alder and spruce, and vanished from European memory by the mid-eighteenth century. The site was designated a UNESCO World Heritage Site on 22 June 2013, on the joint nomination of Parks Canada and the Government of Newfoundland and Labrador.

Primary sources:

• Barkham, Selma Huxley. "The Basque Whaling Establishments in Labrador 1536–1632 — A Summary." Arctic vol. 37, no. 4, December 1984, pp. 515–519. The compact statement of the archival case, with the list of identified ships and stations and the estimate of season tonnages.
• Grenier, Robert, Bernier, Marc-André, and Stevens, Willis (editors). The Underwater Archaeology of Red Bay: Basque Shipbuilding and Whaling in the 16th Century. Ottawa: Parks Canada, 2007. Five volumes, folio. The full excavation record of the San Juan, with timber-by-timber drawings and reconstruction plans.
• Tuck, James A., and Grenier, Robert. Red Bay, Labrador: World Whaling Capital A.D. 1550–1600. St. John's: Atlantic Archaeology, 1989. The site report for the shore excavations, including the Saddle Island cemetery and the try-works platforms.

A dev fact for the back pocket

In the winter of 1969, a small terminal manufacturer in San Antonio, Texas — Computer Terminal Corporation, later renamed Datapoint — approached two chip companies independently with a request to build a single silicon chip that would replace the roughly two hundred discrete TTL logic packages in the programmable terminal they were about to ship. One of the two companies was Intel. Intel delivered the chip in the second half of 1971; CTC had decided months earlier that it was too slow, and shipped the terminal with the discrete TTL implementation instead. Intel productised the leftover design as the 8008 microprocessor in April 1972, added a stack and a larger address space in the 8080, added segmentation in the 8086, and every x86 CPU shipped since — every desktop PC, every laptop, every rack-mounted server in every data centre on Earth not on ARM — descends directly from a 1969 specification for the register file of a video terminal that never used it.

The company was Computer Terminal Corporation (CTC), founded in San Antonio in 1968 by two former Convair engineers, Phil Ray and Gus Roche. Their product concept was a small, programmable, tape-cartridge-loaded terminal that could act as a smart replacement for the IBM 029 keypunch and the ubiquitous glass-teletype terminals of the period; the machine they eventually shipped as the Datapoint 2200 in June 1970 looked like a small typewriter with an integral 12-line CRT and executed programs loaded from a compact cassette. Internally it was built from roughly one hundred TTL small- and medium-scale integrated circuits on two large printed-circuit boards, implementing an 8-bit accumulator-plus-seven-general-register architecture designed by CTC's Vic Poor and Harry Pyle in the summer of 1969.

Poor and Pyle wanted to consolidate the CPU logic onto a single chip. In December 1969 CTC took the block diagram of the 2200's processor — the register file, the ALU, the instruction decoder, the small internal subroutine stack — to two vendors on the same trip: Texas Instruments in Dallas, and Intel in Santa Clara. Both companies agreed to attempt a single-chip implementation on a fixed-price contract. TI's proposal was assigned to Gary Boone, whose team produced the TMX 1795 in 1971 — a working single-chip realisation of the Datapoint architecture that consumed too much power and ran too slow for CTC's target production. Intel's proposal was assigned first to Ted Hoff and Stan Mazor for architecture and, once Federico Faggin joined Intel in April 1970 from Fairchild, to Faggin for silicon layout. Intel's chip — internally the 1201 — slipped its schedule repeatedly and was delivered as an engineering sample to CTC in the second half of 1971, benchmarked at roughly one-tenth the speed of the discrete-TTL implementation of the same instruction set that CTC had by then already built and validated.

CTC was under pressure to ship the Datapoint 2200 into a growing order book and could not wait. It abandoned both single-chip contracts and shipped the 2200 with the TTL implementation. Because CTC had paid a fixed price rather than committing to a per-unit royalty, and because the design contracts contained no exclusivity clause, both TI and Intel were legally free to sell the resulting chip on the open market. TI's TMX 1795 was announced in trade press in 1971, drew no interest, and was written off; TI has occasionally advanced the claim, since, that the 1795 was in fact the first commercial single-chip microprocessor, predating the Intel 8008 by several months. Intel's 1201 was productised, renamed the 8008, and shipped to catalogue customers on 1 April 1972 at an initial unit price on the order of $120.

The 8008's instruction set, register model, and byte orderings were not designed by Intel. They were the Datapoint terminal's, transcribed. The register naming A, B, C, D, E, H, L — where H and L form the sixteen-bit HL memory-pointer pair used for indirect addressing — is the Datapoint 2200's register model, from Poor and Pyle's 1969 specification. The little-endian byte ordering of multi-byte values in memory is the Datapoint 2200's memory model. The port-based I/O instructions (IN port, OUT port) are the Datapoint 2200's I/O model.

Intel's follow-up chip, the 8080 — shipped in April 1974 on a slightly larger die with an extended register file, a full 16-bit address space, and a proper external stack — was upward-compatible with the 8008 through a mechanical assembler-level translation table. The 8086, shipped in June 1978 as Intel's response to Motorola's 68000, was designed by Stephen P. Morse under the explicit constraint that 8080 assembly-language programs should be translatable to 8086 assembly with a straight one-to-one substitution. Morse's architectural paper on the 8086 (IEEE Computer, vol. 11, no. 6, June 1978, pp. 18–27) is explicit about carrying the register naming, the byte ordering, and the port I/O model forward from the 8080, and thereby from the 8008, and thereby from the Datapoint 2200. Every x86 processor since — the 80286 (1982), 80386 (1985), 80486 (1989), Pentium (1993), through Core, Xeon, Ryzen, and EPYC on into the current generation — inherits that lineage. The Datapoint 2200 register set is still there, promoted from A/B/C/D/E/H/L to AX/BX/CX/DX/SI/DI/BP to RAX/RBX/RCX/RDX/RSI/RDI/RBP, at the bottom of every x86 disassembly on every machine everywhere.

CTC itself renamed to Datapoint Corporation in 1972, expanded into networking (they invented the ARCnet local-area network in 1976, more than a decade before Ethernet reached the desktop), and rode the small-business terminal market until the arrival of the IBM PC in 1981 — a machine based on the Intel 8088, a cost-reduced 8086, running an operating system (Microsoft's PC-DOS/MS-DOS) whose author (Tim Paterson at Seattle Computer Products) had written it explicitly to run 8080 code that had been mechanically translated up to 8086 assembly. That is: the IBM PC ran, on a chip descended from Datapoint's spec, an operating system compatible with an instruction set descended from Datapoint's spec, and it destroyed Datapoint's business inside four years. Datapoint filed for bankruptcy protection in 2000.

Primary sources:

• Faggin, Federico, Hoff, Marcian E., Mazor, Stanley, and Shima, Masatoshi. "The History of the 4004." IEEE Micro, vol. 16, no. 6, December 1996, pp. 10–20. The four principals' joint account, with the parallel narrative of the 8008 project running alongside the 4004. Contains Faggin's account of the CTC contract and the delivery slip.
• Wood, Lamont. "Forgotten PC history: The true origins of the personal computer." Computerworld, 8 August 2008. Long-form interview with Vic Poor and Harry Pyle of CTC; the definitive published statement of the Datapoint side of the story, including the December 1969 vendor trip and the register-file design.
• Morse, Stephen P. "The Intel 8086 Microprocessor: A 16-bit Evolution of the 8080." IEEE Computer, vol. 11, no. 6, June 1978, pp. 18–27. The 8086 designer's own architectural paper, explicit on the 8080-compatibility constraints that carried the Datapoint architecture forward into every subsequent x86 chip.

Today's goal

Sit somewhere in friend's own home for fifteen minutes that friend does not normally sit. A different chair, a different corner of the couch, the second step of the staircase, the floor beside a bookshelf, the kitchen counter, the wide window sill in the room that gets the afternoon light. Any spot that is not the usual spot, for fifteen minutes, without a screen. Notice, before the fifteen minutes are up, one thing that is visible from this vantage that is not visible from the usual one.

A room learned from one seat is only ever the room's shape as it appears from that seat. Move eight feet, and the shelf friend has walked past for a year suddenly has depth in it; the crack in the plaster is a different crack, or the same crack going the other way; the light lands on a different half of the wall. Costs nothing, takes fifteen minutes, and is the only Sunday errand today that will still register on Monday morning.

Today's toy is poppycock press — a procedurally-composed front page of a small Victorian broadsheet, with a masthead, a dateline, three columns of news, a woodcut, and a classified-ads column at the foot. Set a new edition, or click any single item to reset just that item. Every edition is nonsense; the compositor is asleep. Lives in the corner.

Go and sit somewhere new, friend.

— C