Eight Films

F1 · The Stone That Remembered: done. The module is at /home/claude/afterlight/storyboard/film1.js. node --check passes and shoot.py renders 5 canvases with no errors. The frames are film1_k1.png through film1_k5.png in the same folder.

Medium: multiplane cut paper with Norstein-style glass planes, over photographed window plates. Matte paper cannot glow, so temperature has to be shown by hands and posture. The stone stays plain speckled grey in every frame.

Keyframes:

1. The cold room (t=0). Wide shot. The stone sits on the sill in the beam from the first frame, with dust in the light. The elder waits with cupped hands, and the child stands at the window.
2. The borrowed patch (t=14). The child angles a hand mirror, and a bright patch lands on the elder's knee beside her hands.
3. The patch goes out (t=26). Dusk. The patch is gone. The elder's hands have become fists pulled to her chest, gripping her shawl.
4. Second touch (t=38). Close-up. The child's palm rests flat on the stone, the other hand is drawn back at her chest, and her face looks down at the stone.
5. The hands close (t=52). The stone, wrapped in linen, is in the elder's lap. Her hands hold the bundle and the child's hand slides away. The film ends on the hands.

After the first render, frames 3 to 5 were too dark, the touch in frame 4 did not land on the stone, and the bundle in frame 5 was hidden. One fix pass corrected all three, and all five frames now read.

References: every reference was opened with WebFetch before I set verified:true. Sources used:

• Prelinger films on Internet Archive: Sleep for Health (1950, public domain), The Face in the Mirror (1940, public domain), New England: Yesterday and Today, and Hope Is in Your Hand.
• NREL "Building a Better Trombe Wall" (1998) and the DOE Consumer Guide to Passive Solar Home Design.
• Hammershøi, Dust Motes Dancing in the Sunbeams (via its Wikipedia page) and SMK KMS3696, "Interior in Strandgade, Sunlight on the Floor" (public domain per SMK's API).
• Soapstone warmers: Smithsonian NMAH 318370 (CC0), Maine Memory Network 17255, and Shaker Museum 23043.

Could not verify or fetch:

• Wikimedia Commons: every file and search page failed to fetch, so I cited the Wikipedia article and SMK open data instead.
• movingimagearchive.com: its search could not be run, from either WebFetch or curl. I confirmed that a source page for Sleep for Health exists there (public domain, links to archive.org/details/Sleepfor1950), but I could not pull a clip ID for the window shot. All footage refs therefore point to archive.org items. Shot timecodes come from the Prelinger shot lists; I did not watch the footage.
• Library of Congress: the item page (loc.gov) returned an error, so it is not used.
• No footage of a sunset seen through a domestic window: the closest are an outdoor sunset pan in New England: Yesterday and Today and a "sunlight streams through window" shot in Sleep for Health.

Rights flagged as "unknown — not granted":

• New England: Yesterday and Today and Hope Is in Your Hand: their archive.org pages state no licence.
• Maine Memory Network: the record says rights have not been evaluated.
• Shaker Museum: the page claims copyright without giving terms.

Treatment claims:

• Thermal storage: supported. NREL says masonry stores the sun's heat and releases it for hours after sunset.
• DOE guide: it defines thermal mass but does not mention Trombe walls, so the release-later claim rests on the NREL sheet.
• Stone size: a hand-sized sill stone holds heat for minutes to tens of minutes, not hours. The gap between sunset and the touch has to stay short to be honest.
• Historic warmers: the documented soapstone warmers were heated on stoves, not by the sun. The sun-warmed stone is the film's fiction built on the same physics, and the basis array says so.
• The stone's "memory": the treatment already calls this a metaphor. A basis entry marked verified:false records that the drama is invented.

I wrote film2.js (F2, "The Picture You Could Feel", role "The bridge to HCI and making"). It passes node --check, and shoot.py renders all 5 canvases with no errors, each in well under 150 ms. I looked at all five PNGs and made one fix pass, as the contract allows. Two frames still have small problems: the hands in KF3 are cut off at the bottom of the frame, and the lamp-off sheet in KF5 is still easier to read than "barely visible".

Files are in /home/claude/afterlight/storyboard/:

• film2.js
• film2_k1.png to film2_k5.png

Medium: embossed cut-paper relief on a tabletop under one lamp. The picture is computed as a height field, and its spread mimics heat diffusion: thin branches stay shallow and the solid sun balloons. It is then shaded by the actual lamp position, so raking light and no light look clearly different.

Cast: two equal makers.

• A: ochre sleeve, darker skin.
• B: slate sleeve, lighter skin.
• In KF4 the roles cross: A reads by touch and B revises the print.

Beats:

1. Lamp on. A draws; B traces with eyes closed and finds nothing.
2. B rejects the thin branches. A spreads paste on the back of a reprint with thicker lines while B holds the corner.
3. Under raking light the sun over-swells, the hill reads clearly and the branches barely rise. Both wait the 30 s cool-down from the tutorial.
4. Both read the relief by touch. B stipples the sun on the next print so it will rise less.
5. Lamp off, cool ambient light. Two fingers follow the hill from opposite ends toward each other.

Research basis (verified):

• The paper: Sosuke Ichihashi, Noura Howell and HyunJoo Oh, "Swell by Light: An Approachable Technique for Freeform Raised Textures", TEI '25 (ACM), Article 45, pp. 1–16, March 2025, DOI 10.1145/3689050.3704420. Ichihashi is the first author.
• Lab page: https://sites.gatech.edu/futurefeelings/?p=474
• Department record: https://lmc.gatech.edu/publications/pub/7751
• What it claims: a dark pattern is printed on the front of the paper and a heat-expanding paste is spread on the back. Visible light heats the dark areas and the back swells. The paper includes a print-pattern optimisation algorithm based on thermal simulation: it darkens thin lines and lightens large solid areas to compensate for heat dissipation. Its applications are tactile picture books, elevation maps, touch and pressure sensors, scented braille and textured origami.
• Tutorials: the lab's tutorials add that "thin lines don't puff well", that the light should be 100 W or more (or sunlight through a magnifier), and that bumps need about 30 s to cool before touching.

Correction to the treatment: the heat-diffusion work is not separate "accompanying work". It is part of the same paper, as the thermal-simulation optimisation. I found no separate heat-diffusion paper. Suggested replacement wording: "Swell by Light (Ichihashi, Howell & Oh, TEI 2025) raises textures by shining light through a dark printed pattern onto a heat-expanding paste; the same paper uses thermal simulation to adjust the print so that heat spreading through the sheet does not leave thin lines shallow or over-swell solid areas." The treatment's own note that the collaboration is fiction is correct and is kept in archivePlan.

References I could not verify:

• ACM DOI page: dl.acm.org refused the fetcher. The paper's details are confirmed through the lab page, the LMC record and Semantic Scholar instead. It is marked verified:false in basis.
• Not used: the Vimeo video (I could only reach a verification screen), the Perkins Flickr tactile-map album (found no album or licence), the Library of Congress item pages (failed to load), the David Rumsey item pages (blocked by robots) and the American Printing House museum records (redirect to another site).

Rights to check before any plate is used:

• Archive footage: the only public-domain film I found is Prelinger "Work Pays America" (1937), whose shot list includes braille maps being prepared and read. "Instructing the Blind Worker" (1944) is also public domain but shows only drill-press training. The NFB stock shot 3493 (hands reading Braille, extreme close-up), BFI "Braille Lesson" (1962) and AFB "Communicating with Deaf-Blind People" (1964) are marked "unknown — not granted" and used as references only.
• Stills: the public-domain LOC/Bain "Map of Europe, for blind" and "Making Matrix", the LOC Red Cross 1918 "Learning to read and write by touch", and NLM and Florida Memory Flickr Commons photos.
• Objects: the Howe 1837 atlas is CC BY-NC-SA 3.0 on the Rumsey site. For Smithsonian NMAH 600011 (Poates 1919 relief map) the metadata is CC0 but image reuse is restricted.
• movingimagearchive.com: I found no tactile or braille footage there. Its search page would not load, and the homepage lists none.
• Swell paper: Wikipedia's "Tactile graphic" page and Perkins describe commercial swell paper as alcohol microcapsules or carbon ink taking heat from a fuser. Neither page mentions infrared, so I make no infrared claim.

For memory: the AFTERLIGHT F2 keyframe module exists at /home/claude/afterlight/storyboard/film2.js. Swell by Light is a Future Feelings lab (Georgia Tech) project by Ichihashi, Howell and Oh, TEI '25. Noura Howell is Watson's NordiCHI co-author.

I wrote film3.js. It passes node --check, and shoot.py renders 5 canvases with no errors. I made one fix pass after looking at the frames: arms were drawn too long, and the opposite ridge in keyframe 4 was too dark to read. All five frames read after the fix.

Film id: F3, "The Last Beacon" (role: Communication and responsibility)

Medium: cut-paper silhouette landscape in tinted night layers, with two ridges and one valley that are the same in every frame. Silhouettes keep each fire fixed on its own hill, so nothing can be read as travelling between fires.

Beats: I kept your five beats and the treatment's protections.

1. Dusk: the far fire appears and the keeper turns toward it.
2. Hesitation: an elder grips the keeper's forearm, a second villager points down to the village, and the storm wall comes in over the ridge.
3. Close-up of the keeper's own hands striking steel on flint into tinder under the stack. Sparks stay local.
4. Reverse angle from a house doorway across the valley: the observer calls the household awake.
5. Rain has reached the valley. The first beacon is out, with only smoke and embers left, and the keeper's fire is low. The household goes into the shelter; hold on the door.
6. No fire ever moves between stations. The flame texture is cut separately for each station.

Archive plan:

• PHOTOGRAPHED layers are storm and sky plates: Project Thunderstorm 1947 (from movingimagearchive.com, public domain), Clouds c.1920s and Thunderbolt Hunters 1941 (both Prelinger).
• Bodies are rotoscoped from Fire Lookout (1957) and Kentucky Pioneers (1941, movingimagearchive.com, public domain).
• Flame texture comes from Forest Fire...Naturally! (1976).
• Every frame carries the COMPOSED SCENE slip. The settlement is fictional.

References I could not verify:

• Currier lithograph, "Signal fires on the Slievenamon Mountains – Ireland, 1848" (LoC 2001705841): commons.wikimedia.org and loc.gov would not load. I confirmed the content and its PD-1923 tag on a Commons mirror (download.osmand.net) only, so it is verified:false.
• Perseus Hopper (perseus.tufts.edu/hopper): blocked by robots. I used and checked the Perseus atlas text instead (Smyth translation, 1926).
• Internet Archive search: archive.org cannot be reached from the shell, so I searched it page by page with WebFetch. Most Prelinger item pages list no licence, so those films are marked "unknown — not granted".
• No archive shot of flint and steel was found. "Survival Stresses" is listed as a candidate only. The hands in keyframe 3 need a live shoot.
• Clouds and Above the Clouds in Rainier: the item pages have no shot descriptions, so the storm and dusk-ridge content still has to be checked by watching them.
• The Han beacon tower photo (Bristol, William Cooper collection, c.1912–19): rights are not stated on the page, so it is "unknown — not granted".
• No dedicated fire-lookout or door footage turned up on movingimagearchive.com. Its search is not reachable by fetch, so I only browsed collection pages.
• Armada 1588: the only sources I could check are thin. Wikipedia has one sentence about the beacon system. H. T. White, "The Beacon System in Kent" (Archaeologia Cantiana, 1934) was checked as a bibliographic record only; I did not read the article. The aubreyresearch.com pages are unsourced and I did not use them.

Treatment claims: none are wrong. The research supports the film's core point:

• Polybius 10.43 says simple beacons could only carry messages agreed in advance, so each keeper has to interpret.
• The Han Juyan slips set station-level rules for how many fires to light.
• Aeschylus's "courier-flame" ("Beacon passed beacon on to us by courier-flame", Agamemnon 281–316) is the travelling-fire image the film refuses; it is cited as the counter-model.
• A loose historical parallel for the village's distrust is a 1545 false alarm, from a secondary web page (romneymarshhistory.co.uk). I did not check it independently.
• The King You beacon story is legend and should not be presented as fact.

Files are in /home/claude/afterlight/storyboard:

• film3.js
• film3_k1.png
• film3_k2.png
• film3_k3.png
• film3_k4.png
• film3_k5.png

F4 · The Same Water is written to /home/claude/afterlight/storyboard/film4.js. It passes node --check, and shoot.py renders 5 canvases with no errors. I checked the PNGs (/home/claude/afterlight/storyboard/film4_k1.png through film4_k5.png) and made one fix pass: submerged hands had washed out to grey, so I put skin tone back under the water veil.

Medium. Cut-paper figures and a stone basin over a drawn water-caustic plate. K1, K2 and K5 are locked overheads, with the two people seated at opposite rims as equals. K3 and K4 are three-quarter shots for the lamp reveal and the adjustment. There are no red/blue overlays. Sensation is shown only through hands: one rests flat, the other lifts out with curled fingers and dripping water. The light pool is the only visible cause, and the light is pale and cool white rather than orange, because the paper found shorter wavelengths felt warmer. K1 plants a faint brighter patch under the hand that settles. K5 repeats K1's framing, with two unequal pools and both people settled.

The beats as built:

1. One person settles and the other withdraws.
2. "Here, try here": they swap seats and their responses swap too.
3. Both look up at a ceiling bar with two lamps, one of them on.
4. Together they turn the second lamp on at lower brightness, aimed at the other person's place.
5. The same water, two unequal pools, both people comfortable.

Research basis (confirmed): Sosuke Ichihashi, Masahiko Inami, Hsin-Ni Ho and Noura Howell, "Hydroptical Thermal Feedback: Spatial Thermal Feedback Using Visible Lights and Water," UIST '24, Article 104, pp. 1–19, ACM, October 2024. DOI 10.1145/3654777.3676453 (https://doi.org/10.1145/3654777.3676453). Ichihashi is first author, from Howell's Future Feelings Lab at Georgia Tech. I confirmed this through the GT LMC publication record, the lab's project page and the UIST video. The ACM page itself refused automated fetch.

What the paper claims:

• Visible light cast on skin under water is perceived as a thermal sensation.
• Shorter wavelengths gave stronger warmth.
• Brighter light felt warmer; people reacted in about 1.5 s.
• Switching lights in sequence made warmth and coolness appear to move along the skin.
• Lighting all the submerged skin raised perceived water temperature by about 13 °C in 25 °C water, which is the "water itself feels warmer" illusion.
• One body of water can give different people different perceived temperatures. The authors list "shared water experiences" as a proposed application.

Treatment check: The treatment's summary of the paper is accurate. Two cautions to keep:

• The paper reports perception. It does not claim the water is heated, and the module says so.
• The two people negotiating together is the film's fiction, not a finding. The treatment already says this.

Other corrections:

• The brief says "Weber's three-bowl experiment". I found no source that credits E. H. Weber with it. The Franklin Institute and Scientific American versions name no originator. The documented early source is Locke, Essay, Book II ch. VIII: "the same water, at the same time, may produce the idea of cold by one hand and of heat by the other." The module cites Locke and the Franklin Institute activity instead.
• I confirmed Locke's wording through a concordance and a secondary quotation. I could not read the Gutenberg text as far as Book II, so the section number §21 is unconfirmed.

References I could not fully verify:

• LOC photochrom "[Roman Baths and Abbey, Circular Bath, Bath, England]" (https://www.loc.gov/item/2002696371/). It appears in LOC search results, but the item page returned errors, so it is verified:false with rights "unknown — not granted".
• The Hydroptical project page and the YouTube video are verified, but rights are not granted. They are lighting references and citations only, not footage to use.
• Home Movie: Hot Springs 1943 (Prelinger) is verified, but the item page states no rights.

Archive found:

• Public-domain footage (all verified):
  • Children of Japan (ERPI, 1941, https://archive.org/details/Children1941): shows a household bath heated by fire and a soak.
  • Health: Your Cleanliness (Centron, 1953): covers hand washing and bathing.
  • Master Hands, shot 10, on movingimagearchive.com (LOC set): hands handling metal parts, used as a rotoscope reference for the lamp adjustment.
• Public-domain stills (verified):
  • Torii Kiyonaga, Onna yu (bathhouse women), 1780s, at LOC (https://www.loc.gov/item/2008660821/).
  • Utagawa Toyokuni I, Women and an Infant Boy in a Public Bath House, ca. 1799, Met Open Access.
• Gaps:
  • No footage of anyone adjusting a lamp over water exists that I could find; K4 says so.
  • movingimagearchive.com's search pages would not fetch, so the only usable clip from that site came from its homepage listing.
  • Wikimedia Commons file pages for the Kiyonaga print would not fetch, so I used the LOC record instead.

I wrote film 5 as /home/claude/afterlight/storyboard/film5.js. It is registered as F5 · Past the Red with the role "The scientific foundation". It passed node --check and shoot.py with 5 canvases and 0 errors. I looked at the five PNGs (/home/claude/afterlight/storyboard/film5_k1.png … film5_k5.png) and all of them read, so I didn't need a fix pass. The figure work is plain: flat engraved shapes, not detailed portraiture.

Medium: copperplate engraving and etching, done as sepia line hatching on plate paper with a plate-mark border. The spectrum is the only colour. Past the red the paper is ruled and plain, with no glow and no false colour.

Keyframes:

1. 0 s: The prism sits in a curtained window and throws a spectrum onto the ruled stand. Herschel is seated, with three thermometers on small inclined planes in the shade.
2. 14 s: He moves No. 1 through the colours while his other hand waits on the notebook. The two standards lie in shade.
3. 28 s: He sets No. 2 on its ramp on plain paper just past the red and withdraws his hand.
4. 40 s: He holds No. 2 beside the shaded standard. Its column is higher. He then puts it back in the same patch.
5. 54 s: The column climbs on paper that shows nothing, and the standard behind it holds still.

What Herschel actually did, and where the treatment differs:

• First paper, "Investigation of the Powers of the prismatic Colours…", Phil. Trans. 90, pp. 255–283:
  • There were three thermometers on small inclined planes.
  • The balls were "blacked with Japan ink".
  • One thermometer was put into a colour through an opening in a pasteboard. The other two stayed "under the shade of the pasteboard" as standards; the paper notes "No. 3 never differed much from No. 2".
  • The prism was set in the upper part of an open window, and the spectrum fell on a table.
• Second paper, "Experiments on the Refrangibility of the Invisible Rays of the Sun", pp. 284–292:
  • The window was darkened with a thick dark green curtain.
  • The spectrum fell on a small stand covered with white paper ruled in lines.
  • He moved the thermometers past the visible red on purpose, kept No. 3 as a fixed standard, and swapped positions to confirm.
  • He placed the heating maximum at "probably not less than half an inch beyond the last visible" red.
• Mismatch 1: In the treatment, a second thermometer simply "rests just outside" the spectrum and turns out to matter. That is a dramatisation. In the sources, the standards sat in shade, and the readings past the red came from a deliberate second series of experiments. I kept the treatment's beat, but framed it as a deliberate placement with a separate shaded standard.
• Mismatch 2: The NASA Space Place PDF says the two controls were placed "beyond the spectrum". That doesn't match the first paper, so I didn't use it for staging.
• Numbers: The UCSB OCR garbles the readings, so the film uses no figures.
• Possibly a fourth thermometer: the first paper also mentions a thermometer belonging to Wilson, and which one was which is unclear. The film shows three.

References I could not verify (marked verified:false or left out):

• The Royal Society Publishing DOI page (rstl.1800.0015) gave a client error.
• JSTOR 107057 and 107056 wouldn't load because of a client challenge. I cited them through the jimclifford.ca OCR pages and the UCSB PDF scans instead.
• The Wikimedia Commons file page for the Abbott portrait returned not found. I verified it through en.wikipedia's File:William_Herschel01.jpg page instead. It is tagged public domain, but the NPG has made a third-party copyright claim.
• The movingimagearchive.com search wouldn't load, so no clips from that site are cited.
• Curl to archive.org is blocked here, so I checked the archive items with WebFetch instead.

Footage notes:

• I found no archive footage of a Herschel-type experiment or of a thermometer reading heat past the red, and the module says so.
• The prism matte comes from Bray's "Experiments with Light Rays". Its prism content is known only from a user review on the item page, and its rights are unknown, not granted.
• The hand and mercury-column rotoscope sources are "Twenty-Four Hours of Progress" (1950, marked public domain) and "[Research Work]" (rights unknown).
• "Solar and Terrestrial Radiation" (1950) is cited only as a cutaway. It contains no shot of the experiment.
• The Science Museum Group thermometer (1927-1708) and the Shorland crayon portrait are CC BY-NC-SA 4.0.

Treatment claims: the treatment's factual claim, that thermometers registered heating past the visible red in 1800, is supported. Calling it "infrared" is later vocabulary; Herschel wrote of "invisible rays".

F6 · "The Sun at the Door" is written to /home/claude/afterlight/storyboard/film6.js. It passes node --check, and shoot.py renders all 5 canvases with no errors. The PNGs are /home/claude/afterlight/storyboard/film6_k1.png through film6_k5.png. I did one fix pass on poses: Tajikarao's arms, the Futo-tama and Tajikarao overlap in frame 4, and Koyane in frame 5.

Medium: flat colour planes of torn and cut paper with real drop shadows, staged in layers like a stage set. This suits a story that survives as performance (kagura). It copies no print and no living artist. Every frame carries an "ADAPTED from the Kojiki and Nihon Shoki, after Chamberlain 1882 / Aston 1896 · one telling, not Japanese belief at large" slip. The deities are always INVENTED paper. PHOTOGRAPHED material is used only for crowd motion and background plates of the community. Frames 2 and 5 carry the COMPOSED SCENE slip. The living kagura footage is CC BY-NC-ND, so it is used as rhythm reference only and never composited.

Keyframes:

1. Sealed door (0 s), shown in cross-section: Amaterasu sits inside in her own light, and the gathering stands outside in the dark with the roosters on a perch.
2. Uzume on the overturned tub (13 s): firelit, with the gods laughing.
3. The door opens a crack (26 s): Futo-tama holds up the sakaki (jewels on the upper branches, mirror in the middle, cloth offerings below), Koyane recites, and Tajikarao waits hidden against the rock.
4. The mirror shows her own face (39 s): Tajikarao takes her hand, and Futo-tama draws the straw rope behind her.
5. Daylight on the gathering (52 s): people's shadows appear, Uzume rests on the tub, the roped door is small in the distance, and only Amaterasu's sleeve is in frame. The frame ends on the community, not the sky.

References I could not fully verify:

• Toshimasa 1889 Commons file: I could not open the Commons file page itself; only a Commons mirror (czwiki), which lists it as public domain, and the ukiyo-e.org entry. I verified it at the MAK instead (KI 10404, sammlung.mak.at). The MAK page lists both PDM 1.0 and CC BY-NC-ND 4.0 without saying which applies, so the rights are recorded that way in the module.
• Kunisada 1857: verified as en.wikipedia File:Amaterasu_cave_edit2.jpg (hosted on Commons, public domain) and at the V&A, E.5594-1886. The V&A's image terms are non-commercial.
• Aston 1896: the Internet Archive item page is verified (NOT_IN_COPYRIGHT). Its full text was cut off before the Rock-cave passage, so I read the passage in the Maricopa Open OER reprint, which labels itself Aston 1896.
• Chamberlain Kojiki: read at sacred-texts.com/shi/kj/kj023.htm. The page prints no section heading; it is probably section XIV.
• movingimagearchive.com: none of its 24 collections is about Japan, and its search page did not load. I found no clip there for this film.
• Archive footage: there is no public-domain Iwato kagura footage that I could confirm. The moving-image references are:
  • Prelinger "Gion Festival 001": undated, no rights statement.
  • 1928 Nikko home movie: no log sheet, so what it shows is unconfirmed.
  • BFI "Rice Festival in Kyoto" (1911): rights unknown.
  • Ise Dai-kagura 2023: CC BY-NC-ND. All of these except the Ise clip are marked "unknown — not granted".
• Ise Jingū: I verified the About page, which calls the Holy Mirror "the symbol of Amaterasu-Omikami" and describes its move to Ise. It does not itself say this is the Iwato mirror. The direct Naiku page would not load.

Treatment and brief claims that are wrong or only half-supported by the texts:

• The tub: "overturned tub" is in Aston's Nihongi ("a tub bottom upwards"). Chamberlain's Kojiki has a "sounding-board", with a footnote saying a parallel text reads tub.
• Who handles the tree and rope: in the Kojiki, Futo-tama holds the uprooted sakaki, and he draws the rope. In the Nihongi main text, Koyane and Futo-dama dig up and hang the tree, and Nakatomi no Kami and Imibe no Kami draw the rope. The film follows the Kojiki.
• The mirror: only the Kojiki has it pushed toward her and has her gradually come forth looking at it. Even there, the text does not say she took the reflection for another deity; that reading comes from Uzume's line "a Deity more illustrious than Thine Augustness". In the Nihongi main text she peeps out and is led out by the hand. The mirror hangs on the tree, but no one shows it to her, and no laughter is recorded.
• Tajikarao: confirmed in both texts. He stands hidden beside the door and leads her out by the hand.
• "Shimenawa": this is the later name. The texts say "bottom-tied rope" (Kojiki; the Nihongi adds "also called a left-hand rope"), and the Kojiki gives the line "Thou must not go back further in than this!"
• Cleveland Museum caption: its description of Hokkei's pentaptych says a stone was "placed to trap her". That is not in either text, and I did not use it.
• Solar identity and the mirror: the treatment's claim that both are documented within the tradition is supported by Ise Jingū's own page.

Possible memory note: for AFTERLIGHT F6, the Kojiki (not the Nihongi) is the version used, because only it has the mirror shown to her. All deity imagery is INVENTED cut paper.

F7, "The Man Who Stole the Month": done. The module is at /home/claude/afterlight/storyboard/film7.js. node --check passes, and shoot.py rendered 5 canvases with no errors (/home/claude/afterlight/storyboard/film7_k1.png … film7_k5.png). I did one fix pass after looking at the renders: the Moon now sits above the box instead of being hidden by it, the clerk's tablet sits on top of his figure, and the tablet in frame 4 no longer runs into the border. One small fault is left: in frame 5, Strepsiades' kneeling foot slightly overlaps the bottom meander.

Medium: Attic red-figure vase painting. Terracotta figures on black gloss, black relief lines, dilute-brown folds, added white for the Moon, stars and the old man's hair and beard, and meander bands top and bottom. No single vase is copied. The cast stays the same across all five frames: Strepsiades (bald, white beard), the Thessalian woman (peplos and sakkos), a young clerk, a bearded creditor with a staff, and two witnesses.

Beats (0, 13, 26, 40 and 52 s):

1. Strepsiades lies awake counting debts on his fingers, with the Moon in the window, his son asleep and his tablets on a stool.
2. The woman draws the Moon down into the open round crest-case.
3. The stone focuses sunlight onto the clerk's wax tablet as he writes the suit.
4. Strepsiades holds up the blank, dripping tablet while the creditor points and two witnesses swear and remember.
5. The box lies empty, the full Moon rises behind the hill, and the creditor holds out his palm.

Text verified:

• I checked the quotations in Hickie's 1853 public-domain translation (Project Gutenberg #2562): "purchase a Thessalian witch, and draw down the moon by night … shut it up, as if it were a mirror, in a round crest-case"; "the money is lent out by the month"; the "beautiful and transparent" stone from the druggists; "when the clerk was entering the suit … melt out the letters of my suit".
• Perseus' reader pages block automated fetching. I pinned the line numbers with Perseus' vocabulary tool on the Greek text instead:
  • 749–751: σελήνη, καθαιρέω, λοφεῖον
  • 752–758: Θεσσαλός, κάτοπτρον, ἀνατέλλω, δανείζω
  • 766–769: λίθος, φαρμακοπώλης, διαφανής
  • 772: ἐκτήκω, γράμμα, δίκη
• So the Moon scheme is about lines 749–756 and the burning-glass scheme about 766–772.

References I could not fully verify, or whose rights are not granted:

• Wikimedia "drawing down the moon" category: Commons would not load from here, so I could not confirm the category. The image usually given for it is the Hamilton/Tischbein plate of 1791 (NYPL, public domain). NYPL titles it as two women preparing to fight before Diana Taurica; the moon-drawing reading comes from D. E. Hill (1973). I cite it as contested and use it for pose and costume only.
• movingimagearchive.com: its search was not reachable, so there are no clip-level references from it. Archive plates come from Internet Archive / Prelinger and ESO instead:
  • Hamm's full-moon shots in "Classic Television Commercials" parts I and II (public domain, timecode 01:37:14:27).
  • ESO's crescent-Moon-setting time-lapse (CC BY 4.0).
• Rights not granted:
  • "Prove It With A Magnifying Glass" (1959) and the Calvin 1968 "full moon views" reel.
  • The ESO "A full moon rises" still (licence not stated on the page).
  • The British Museum Nimrud lens.
  • The British Library wax tablets.
  • The Douris cup.
  • The 1962 Epidaurus "Helen" newsreel (InC-EDU). I found no open-licence footage of a Greek theatre performance, so this newsreel is only a staging reference and is not used as a plate.
• I found no archive footage of an ancient-style burning glass. The 1959 magnifying-glass film is the nearest match.

Treatment and brief claims that are wrong or overstated:

• Wax: Aristophanes says "melt out the letters" (ἐκτήκω). Wax is implied by the clerk's tablet but never named.
• Lawsuit: the suit is being entered by a clerk, and Strepsiades plans the melting in advance. It is not an existing record he goes back to destroy.
• The two schemes are separate: they are two answers in Socrates' quiz, and Socrates praises both ("Capital!", "Cleverly done, by the Graces!"). Putting them in one plot, adding the creditor's witnesses at the melted tablet, and ending on the Moon rising are our adaptations, and the film's caption slips say so. The creditor-with-witness comes from the later Pasias scene.
• Nimrud lens: the British Museum says it was probably an inlay and that its optics were likely accidental. I use it only as a look reference for the "transparent stone", not as evidence of ancient burning glasses.
• Burning glasses: Pliny, Natural History 37.10 (crystal in sunlight used as a cautery) is the verified ancient source for the practice.
• Date: 423 BCE is the first production. The text we have is the later revision of about 419–416 BCE.

I wrote film 8 to /home/claude/afterlight/storyboard/film8.js and registered it as F8, "A Letter from a Dead Star", role "The cosmic closing film". node --check and shoot.py both run with no errors. I looked at all five PNGs (film8_k1.png to film8_k5.png in the same folder) and made one fix pass: I made the observer's star stand out from the star field, broke up the too-regular petals of the explosion in frame 3, and made the pencil arcs and the planet's town lights easier to see in frame 4.

Medium: a stipple dot field on black paper with hand-pencilled spacetime arcs. The star is only ever a field of arriving points, so dots keep the appearance and its source in one material. The pencil arcs show the time between them. There is no text and no travelling particle; the light is drawn only as a widening front.

Keyframes:

1. At 0 s, the observer stands by a small lit settlement on a dark planet and points at one star.
2. At 12 s, the source, far back in time, is still a blue-white star inside a ring of gas it shed earlier (modelled on SN 1987A).
3. At 24 s, the core collapses, the outer layers are thrown off, the ring lights up in bright spots, and a small dense remnant stays at the centre.
4. At 38 s, the collapse light is still crossing empty space while older light keeps reaching the planet, whose towns carry on.
5. At 52 s, later, there is a new house, the lantern is on the ground, and a child points at the same star. Its light left before the collapse.

References I could not verify or clear:

• Journey Into Light (c. 1973, Prelinger, filmed at Kitt Peak): the item states no rights, so I marked it "unknown — not granted".
• Who's Out There? (1975) and Space Navigation (1967) on movingimagearchive.com: both are public domain NASA/NARA films and I confirmed the pages exist. Neither page describes individual shots, so I could not confirm which shots show stars or telescopes. Someone needs to log them.
• The ESO artist's impression on the Chandra animations page: I did not check ESO's terms.
• The Commons page for Tycho's 1573 Cassiopeia chart would not load. I verified it through en.wikipedia.org/wiki/File:Tycho_Cas_SN1572.jpg instead. It is tagged public domain, but the page warns that a US public-domain tag is missing.
• The fictional observer, settlement and planet have verified: false in basis, because they are invented.

Problems with the treatment and brief claims:

• SN 1987A distance: sources disagree. Wikipedia gives about 168,000 light-years, NASA JPL 160,000, and ESA/Hubble about 170,000. The neutrino burst was 25 neutrinos detected 2 to 3 hours before the light. That lead came from photons being slowed on their way out of the star, not from neutrinos travelling faster.
• Betelgeuse distance: published central values run from about 548 light-years (Joyce et al. 2020, 168 parsecs) to about 724 light-years (Harper et al. 2017, 222 parsecs), so the brief's "550–700 ly" is slightly too narrow.
• Is "dead star" accurate?
  • SN 1987A: partly. The blue supergiant Sk −69 202 has ended as a star, but the remnant is most likely a neutron star (JWST, 2024). That is strong evidence, not a direct detection.
  • Betelgeuse: no. It is a living red supergiant that has not exploded; its 2019–20 dimming was caused by a dust cloud.
  • Tycho's 1572 star: only loosely. It was a Type Ia supernova: a white dwarf, already a stellar remnant, blew apart and left no neutron star or black hole, only an expanding gas shell.
  • In the film, "dead" should mean that the source has changed, not that nothing is left.
• The NASA SVS SN 1987A film is a simulation. It is listed as a reference and must be labelled as a simulation, not as a recording of the event.