# Easels, Lecterns & Display Stands: research notes

Date: 2026-10-25. Scope: freestanding or tabletop frames that hold something up at an angle or height rather than flat on a table: artist's easels (tripod/A-frame, H-frame studio, field/plein-air, pochade box, table-top, basic display), presentation and flip-chart easels and stands (bare stand only), lecterns and podiums (reading/speaking stands, portable and fixed), and small display stands (poster and sign holders, pedestal and telescopic stands, literature/pamphlet holders, book, plate and photo stands). Taxonomy: `supplies > crafts > easels`, add-on `data/schema-parts/supplies/crafts/add-ons/easel.jsonc`. This leaf bundles three functionally distinct product families under one label ("Easels, Lecterns & Display Stands"); per the brief, this stays one register and the `primary_use` split does the dividing, rather than three separate registers.

Sibling boundary: `supplies > stationery > boards` (`stationery-boards`, already published as of this run) explicitly routes "bare easels, tripods and lecterns" to this leaf, and states "a mobile whiteboard sold without its own board is an easel." The boundary this register observes, stated from this side: a stand or easel sold *with* its own writing or display surface (a corkboard, a whiteboard panel, a chalkboard) is `stationery-boards`'s product; a bare stand, tripod or frame that *holds* someone else's pad, board, canvas, sign or book is this leaf's product, even when it is marketed alongside a pad (a "flip-chart easel" is this leaf; the flip-chart pad itself is paper/`paper`'s domain, not checked here). No other sibling in `supplies > crafts` claims lecterns, podiums, sign holders or literature stands, and no other vertical in the catalog claims plate, book or photo display stands as of this run, so small tabletop display stands are kept in scope here by default rather than left unhomed.

Quote labels: **verbatim** = the sentence in quotation marks was in text I read myself (the fetch tool returned it in quotation marks, or I ran curl plus a Python regex strip and read the result, with pdftotext for PDFs); **fetched excerpt** = the fetch tool summarised or lightly reworded the page; **search excerpt** = the claim came from a search-engine summary because the page itself would not load. A search-excerpt-only number does not go in category.json. WebSearch reported its session budget already used (200 of 200) on the first two queries of this run, before any easels-specific search returned results, so this entire register was built from direct fetches (WebFetch, curl) and pages linked from those, with no search-engine results at all; every source below was reached by guessing or following a real URL, not by searching for it. Could-not-reach sources are listed at the end of this section.

## Seed

Seed script output (verbatim, from `seed.py --category easels`):

```
=== Example entry (from an unrelated category) ===
category: olive-oil
what this shows: a facet that experts rank first, that almost no listing carries, and that is expensive to backfill. The honest answer is to list it, mark it researched, and limit backfill to the top of the catalog. Also shows a derived metric hanging off a researched facet.
<example: harvest_date facet, definition/why_experts_care/backfill as in schema; months_since_harvest derived metric>

=== Persona ===
A barista trainer, who thinks in the variable most beginners get wrong

Ask yourself two things and write the answers under 'Seed' in notes.md:
  1. What is the easels equivalent of the example above?
  2. What would this persona notice first about easels that a typical buyer would not?
Keep what survives the data check. Drop the rest.
```

1. **The easels equivalent of harvest date is maximum loading capacity.** It is the number a serious painter actually wants (will this easel hold my largest wet canvas without sagging or racking), it needs a judgement attached (it is the maker's own rating, not a certified test), and it is almost never on a retail listing — only on the maker's own spec page, and even there inconsistently (Mabef states it for every studio easel but calls one of its heavier-weight models "Big" while giving it a *lower* capacity than two lighter-weight models, which only a spec-page reader would catch). Survived the data check: four Mabef product pages (M/02, M/04, M/06, M/10) each stated it directly, so the facet is real and the backfill-cost-high, source-researched treatment mirrors `harvest_date` exactly. Kept as `max_loading_capacity_kg`, researched, backfill limited to top N in the painting branch.
2. **The barista-trainer persona (thinks in the variable beginners get wrong) notices that beginners judge an easel, a lectern or a stand by its overall size, when the mechanism that actually does the work is hidden one level down.** A beginner picks the tallest easel assuming it's the sturdiest, but Mabef's own range shows its "Big" model (M/06, 18.5 kg) has a *lower* rated capacity (30 kg) than two of its mid-sized models (M/02 and M/04, 32–42 kg, both rated to 40 kg) — the mast count and the crank-versus-ratchet mechanism decide capacity, not overall height. The same mistake shows up on a lectern (a beginner books the most imposing-looking podium, not one that opens underneath for a seated speaker) and on a display stand (a beginner checks the stand's height, not the slot's maximum insert thickness). Survived the data check for easels (Mabef mast/mechanism data) and lecterns (ADA knee-clearance standard); kept as `mast_type`, `canvas_holder_mechanism`, and `knee_clearance_accessible`. Dropped for display stands: I could not reach a brand spec page stating insert-thickness in numbers (see "Could not reach" below), so that specific beginner trap is recorded here but not asserted as a sourced facet; `holds_item_type` (what kind of item, not what thickness) is the sourced substitute.

## Sources and what each supports

### Primers (Wikipedia)

**Wikipedia, Easel**
https://en.wikipedia.org/wiki/Easel
- fetched excerpt (quoted terms): three design types — "A-Frame", "based on three legs"; "H-Frame", "based on right angles"; "Multiple Purpose", combining both with "finite rotational, horizontal and vertical adjustment".
- fetched excerpt: easel categories — "Artists' easels", "Pochade boxes" (mounted on camera tripods), "Display easels" (for finished artwork), "Facilitation easels" (for audience participation with large paper pads), "Field/Plein air easels" (portable, telescopic or collapsible legs), "French box easels" (compartments for supplies), "Studio easels" ("supporting panels weighing over 200 pounds and measuring over 7 feet in height" — this is the panel/canvas the easel supports, not the easel's own weight).
- fetched excerpt: materials — "wood, aluminum or steel."
- Supports: `primary_use` family text, `base_type` values (A-frame/tripod, H-frame), `material` values, `height_adjustable` why_experts_care context, `folded_or_stored_size` why_experts_care (field easels fold). Confidence high for the design-type taxonomy (consistent with Mabef's own six-category catalog, below); medium for the 200 lb/7 ft figure, which is reported by the tool's summary rather than a page I fetched and read in full HTML — flagged, not used as a numeric claim in any facet, only in the reference's own supports line.

**Wikipedia, Lectern**
https://en.wikipedia.org/wiki/Lectern
- fetched excerpt (quoted): "A lectern is a standing reading desk" with "a slanted top", versus "a podium is an elevated platform upon which a lectern is placed"; "Lectern vs. Pulpit": lecterns hold scripture readings, pulpits hold sermons, "churches often position them on opposite sides."
- fetched excerpt: "Sloped reading surface with a lip to prevent materials from sliding"; "Modern versions may feature motorized height adjustment."
- fetched excerpt: Judaism's amud (synagogue reading table) and shtender; Islam's rehal (folding Qur'an stand).
- Supports: the `lectern`/`pulpit`/`podium` vocabulary entries, `reading_surface_tilt` definition (the lip), `height_adjustable` why_experts_care (motorized height adjustment exists), `sound_system_builtin` context. Confidence high; this is a tool summary of the page rather than text I stripped from raw HTML myself, so labelled fetched excerpt throughout, not verbatim.

**Wikipedia, Flip chart**
https://en.wikipedia.org/wiki/Flip_chart
- fetched excerpt: standard pad sizes "25 x 30 inches and 20 x 23 inches" (plus metric A1/B1); three stand types — "metallic tripod stand... the commonest support", a table-top "isosceles triangle box" design, and a "wheeled mount."
- Supports: `pad_size` values and definition, `wheels_or_casters` why_experts_care, `base_type` context. Confidence medium: single tertiary source, no manufacturer spec sheet cross-check reached (see judgement 3 and "Could not reach").

### Brand spec sheets (easels)

**M.A.B.E.F. (mabef.com), six easel catalog pages plus four model pages read directly**

The main site (`mabef.it` redirects to `mabef.com`) lists six top-level product categories: studio easels, lyre easels, easels sketch boxes, field easels, table easels, different articles (fetched excerpt). I fetched four individual studio/basic-easel model pages directly (curl, then stripped tags with a Python regex and read the result as plain text — labelled verbatim below because the quoted lines are exactly what the stripped text contained):

- **M/02** (http://www.mabef.com/en/products/studio-easels/m02) — verbatim: "DOUBLE MAST WITH CRANK", "Constructed of oiled, stain-resistant beech wood", "Also available in the electric version (motor 110 V or 220 V)", "Ideal for very large canvas", "Canvas holder height is adjusted by a crank mechanism", "Canvas holder inclination is adjusted manually", "Double extensible sliding mast". Technical information: "Weight: 42 Kg - 92 lbs", "Maximum loading capacity: 40 Kg - 88 lbs".
- **M/04** (…/m04) — verbatim: "WITH CRANK", "Ideal for very large canvas", "Extensible sliding mast" (single, not double). "Weight: 32 Kg - 70 lbs", "Maximum loading capacity: 40 Kg - 88 lbs".
- **M/06** (…/m06) — verbatim: "BIG", "Adjustable working angle by lever handles" (tilt), "Canvas holder height is adjusted by ratchet" (height), "Extensible sliding mast" (single). "Weight: 18,5 Kg - 40¾ lbs", "Maximum loading capacity: 30 Kg - 66 lbs".
- **M/10** (…/m10) — verbatim: "BASIC", "Adjustable working angle by lever handles", "No-collapsible safety system", "Canvas holder height is adjusted by ratchet", tagged both "Oil-Acrylic easel" and "Display easel". "Weight: 10 Kg - 22 lbs", "Maximum loading capacity: 10 Kg - 22 lbs".

Supports: `max_loading_capacity_kg` (range and the four-fold spread), `weight_empty_kg` (range), `mast_type` (single vs double, verbatim model names), `canvas_holder_mechanism` (crank vs ratchet vs lever, verbatim), `material` (beech wood, verbatim), the `payload_ratio` derived metric (computed directly from these four weight/capacity pairs), judgement 1 (no certified test — these are plainly the maker's own stated numbers, with no standard or test method named anywhere on the four pages). Confidence high: each number was read from the page's own stripped text, not a tool summary. One brand only — not generalised into a catalog-wide range anywhere in the JSON; every mention of "10 to 40 kg" or "10 to 42 kg" is attributed to Mabef specifically, following the house convention from the lamps and power-bank registers of reporting brand-specific numbers as the brand's, not as a category universal.

I also downloaded Mabef's 2025 digital mini-catalog PDF (25 MB, 2 giant pages) and ran `pdftotext -layout` on it; it returned zero lines of text, meaning the catalog is built from flattened images rather than text layers. Not used; noted under "script/tool problems" below as it is a data source, not a skill script.

### Brand/retailer taxonomy (display and presentation stands)

**Testrite Visual (testrite.com)**
- Classic A-Frame Easels category page (https://www.testrite.com/products/easels-and-art-materials/convention-and-presentation/classic-aframe-easels/classic-aframe-easels/), read via curl + Python tag-strip (UTF-8 decode failed on one byte; re-read as latin-1): verbatim — "All units open to 72" high", "Welded aluminum easel construction", "1" square heavy duty legs", "Front Plate Choices: Aluminum - Strong yet lightweight / Steel - Strong and magnetic", "Spring-loaded pad holder", model names "BLK STEEL MAGNETIC A-FRAME EASEL" and "BLK ALUMINUM A-FRAME EASEL". Sub-categories listed in the sidebar: Convention and Hotel Easels, Economy Display Easels, Elegant Baroque/Lobby Easels, Flip Chart and Display Easels, Gallery & Exhibit Wall Stands, Portable Presentation Easels, Steel Easels Print Racks, Table Easels.
- Telescopic Acrylic Stands category page (https://www.testrite.com/products/aluminum-drop-in-frames/stands-drop-in/telescopic-acrylic-stands/telescopic-acrylic-stands/): verbatim model names — "Telescopic Acrylic Stand with 11" Round Base, VERTICAL", "...12" Square Base...", plus sidebar categories: Perfex Pedestal SignFrames, Economy Poster SignHolders, A Frame SignHolders, Convex Poster SignHolders, EasyOpen SnapFrame Pedestal Stands, Evo Literature Holders, Information Centers, SignHolder Grippers.
- Home page navigation (fetched excerpt via WebFetch): "Pedestal Stands," "Retail Display Stands," "Telescopic Acrylic Stands," "Blade Poster SignHolders," "Poster Frame Stands," "Stellar Pallet Stands," "Literature & Info Stands," "EVO Literature Holders."
- Supports: `holds_item_type` and `base_type` values and definitions (pedestal, telescopic/tripod, flat frame), `primary_use` (presenting_to_a_room and showing_or_displaying families), the display-stand half of the mechanism section. Confidence high for the category and model-name taxonomy (read directly); no numeric weight-capacity or insert-thickness figure was found on either page, which is why `max_item_size_raw`, `folded_or_stored_size` and an insert-thickness fact for display stands stayed researched/unsourced rather than invented (see "Could not reach").

### Regulator

**U.S. Department of Justice, 2010 ADA Standards for Accessible Design**
https://www.ada.gov/law-and-regs/design-standards/2010-stds/
- verbatim (read as stripped plain text from the live page, which contains the full standard, not just a table of contents, past the ~300,000th character): "902.3 Height. The tops of dining surfaces and work surfaces shall be 28 inches (710 mm) minimum and 34 inches (865 mm) maximum above the finish floor or ground."
- verbatim: Advisory 902.1 lists work-surface examples including "writing surfaces."
- verbatim: "306.3.1 General. Space under an element between 9 inches (230 mm) and 27 inches (685 mm) above the finish floor or ground shall be considered knee clearance... 306.3.3... 11 inches (280 mm) deep minimum at 9 inches... and 8 inches (205 mm) deep minimum at 27 inches... 306.3.5 Width. Knee clearance shall be 30 inches (760 mm) wide minimum."
- Supports: `knee_clearance_accessible` definition and why_experts_care, the 28–34 in and knee-clearance numbers in judgement 2, the ADA tradeoff. Confidence high for the text itself (this is the primary legal standard, read directly, not summarised); confidence only medium for applying it to lecterns specifically, because the standard's own text never says the word "lectern" or "podium" — it is my own reading of the Advisory 902.1 example list ("writing surfaces"), and I could not reach a second document (DOJ business brief, Access Board chapter-9 guide — no such guide exists in their chapter list) that names lecterns by name. Judgement 2 is worded to reflect this: "likely baseline," not a settled fact.

## Data points used for derived metrics

| Model | Weight (kg) | Max loading capacity (kg) | Mechanism | payload_ratio |
|---|---|---|---|---|
| M/10 (basic/display) | 10 | 10 | ratchet, single mast | 1.00 |
| M/06 (big) | 18.5 | 30 | ratchet, single mast | 1.62 |
| M/04 | 32 | 40 | crank, single mast | 1.25 |
| M/02 | 42 | 40 | crank, double mast | 0.95 |

`payload_ratio` is not monotonic with size (M/06 has the best ratio, not the biggest or smallest model), which is exactly the "beginners get this wrong" point from the seed answer: a bigger easel is not automatically a better ratio of capacity to carried weight.

## Against the existing add-on schema

Current `easel.jsonc` has exactly one field: `sizing.max_canvas_or_board_size_raw` (string | null). `crafts-standard.jsonc` adds nothing (empty object; crafts inherits only `supplies-standard.jsonc`'s `high_signal_markdown` and `sizing` block). Field-kind vocabulary from `zamana-expertise/AGENTS.md`: vertical anchor, quality currency, hard filter, sizing (bought-by-number, variant-level dimension), or high-signal markdown; visually-derived facts (shape, style, visible hardware) are explicitly excluded from the schema. Proposals only; nothing in `zamana-expertise` was edited.

| Facet in this register | In add-on today | Proposed kind | Note |
|---|---|---|---|
| `max_canvas_or_board_size_raw` (existing) | yes | sizing, as-is | Confirmed real: Mabef and Testrite both describe capacity in free text ("Ideal for very large canvas", "holds canvases up to..."), not a clean number, so the existing `_raw` string design already matches how the fact actually appears. Kept unchanged. |
| `max_loading_capacity_kg` | no | quality currency, limited backfill | The harvest-date-style fact: experts rank it first, it is essentially never on a retail listing, only on brand pages, and even then it does not generalise across brands (judgement 1). Propose adding it but explicitly scoped to top-N backfill in the painting branch only, the same caveat `olive-oil`'s `harvest_date` example carries. Whether it belongs inside `sizing` is unclear to me: Mabef prices it per fixed model (M/02 is always 40 kg), not per a chosen size variant of one listing, so it may be variant-invariant and belong outside `sizing` under AGENTS.md rule 2 — flagging rather than deciding, since I do not know whether Zamana's own catalog sells a single easel listing across multiple capacity variants. |
| `mast_type`, `canvas_holder_mechanism` | no | high_signal_markdown, or hard filter if dense enough | Real and sourced (Mabef's own model names encode exactly these two axes), but backfill cost is medium-to-high (brand pages only) and I have only one brand's worth of naming convention to judge density from. Safer proposed as a line in `high_signal_markdown` per the "distill... build quality and durability signals" comment, promoted to a structured filter only if a second brand confirms the convention is universal enough to be worth augmenting broadly. |
| `material` | no (not in easel.jsonc; a generic concept, not currently structured for this leaf) | ambiguous: hard filter vs high_signal_markdown | `supplies-standard.jsonc`'s own comment names "build quality and materials" as exactly what belongs in `high_signal_markdown`, which argues against a structured `material` field here. But wood/aluminum/steel/acrylic is also usually stated in the title, which argues for a hard filter per AGENTS.md rule 4. I flag the tension rather than resolve it; this register's `material` facet should be read as a candidate for either slot, not a committed proposal for a new top-level field. |
| `base_type` | no | flagged as visually-derived, likely NOT a new field | AGENTS.md rule 1 excludes anything "better determined by looking at the product photo than by reading the page." Tripod legs vs pedestal vs flat easel-back is visually obvious in nearly every product photo. I kept it as a facet in this register because it is still a real thing an assistant should know about the product (same logic as the power-bank register's `integrated_cable`, flagged "a photo also shows it, so this is borderline"), but I am not proposing it as a new schema field; the image pipeline should already cover it. |
| `reading_surface_tilt`, `knee_clearance_accessible`, `sound_system_builtin` | no (lectern-specific; no lectern add-on exists in `supplies/crafts` today) | hard filter, with `knee_clearance_accessible` as the safety/accessibility facet this register's brief requires | None of the four Mabef/Testrite pages describe a lectern, so these three rest on the Wikipedia primer and the ADA standard rather than a brand spec sheet; backfill cost is marked `high`/`medium` accordingly, and `knee_clearance_accessible` in particular has no marking to scan for at all (noted in its `backfill_note`). |
| `holds_item_type`, `pad_size`, `wheels_or_casters` | no | hard filter | Sourced to Testrite's category taxonomy and the flip-chart primer; all three are plainly stated on listings/titles when present ("poster stand," "with casters"). |
| `weight_empty_kg`, `folded_or_stored_size` | no | sizing (weight) / high_signal_markdown or sizing (folded size) | `weight_empty_kg` matches `supplies-standard.jsonc`'s existing `sizing.weight` shape exactly and could simply populate that generic field rather than need a new easel-specific one. `folded_or_stored_size` is a genuine "bought-by-number" fact for a portable easel (AGENTS.md rule 2's own cable-length example), but backfill cost is high since almost nothing states it; proposed for `sizing` with a low-density caveat. |

Add-on fields this register found no use for: none — the existing add-on has only the one field, and this register keeps it. Nothing here is "noise" to cut; the gap runs the other way; the add-on is nearly empty relative to what the two sub-trades (painting and presentation furniture) actually spec.

## Branches

Each leg is traced to a specific sentence above or in the register.

**Split 1, `primary_use`** ("Painting, presenting to a room, reading a speech, or displaying something?")

- when `painting or crafting`: then_ask `max_item_size_raw`, `mast_type`, `canvas_holder_mechanism`, `max_loading_capacity_kg`; makes_irrelevant `reading_surface_tilt`, `knee_clearance_accessible`, `sound_system_builtin`, `holds_item_type`, `pad_size`, `base_type` (added 2026-10-25).
  - then_ask traces to the four Mabef pages, which state exactly these four facts for every model and nothing about reading surfaces, sound or pad sizes.
  - makes_irrelevant traces to the same four pages describing only canvas/board holders, never a reading surface, microphone, display slot or paper-pad clip; a canvas easel has none of these mechanisms (Wikipedia's easel-category list likewise never mentions them). `base_type` added after the blind check: a canvas easel's own tripod/H-frame legs are already covered by `mast_type` and the mechanism text, so a separate base-type question on this branch has nothing left to ask.
- when `presenting to a room`: then_ask `pad_size`, `base_type`; makes_irrelevant `knee_clearance_accessible`, `sound_system_builtin`, `mast_type`, `canvas_holder_mechanism`, `max_loading_capacity_kg`, `max_item_size_raw`, `reading_surface_tilt`, `holds_item_type` (three added 2026-10-25).
  - then_ask traces to the flip-chart primer (pad sizes) and Testrite's "Flip Chart and Display Easels" plus three stand-shape types.
  - makes_irrelevant: a flip-chart stand holds a paper pad, not a canvas needing a crank/ratchet/mast rated in kilograms (Mabef's mechanisms are absent from the Testrite A-frame presentation-easel page), and it is not a reading lectern. `max_item_size_raw`, `reading_surface_tilt` and `holds_item_type` added after the blind check: a flip-chart stand's size is the pad size already asked, it has no reading-lectern tilt, and it holds one fixed thing (a pad), not a variable "item type" the way a display stand does.
- when `reading or speaking`: then_ask `reading_surface_tilt`, `knee_clearance_accessible`, `sound_system_builtin`; makes_irrelevant `max_loading_capacity_kg`, `mast_type`, `canvas_holder_mechanism`, `pad_size`, `holds_item_type`, `max_item_size_raw`, `base_type` (two added 2026-10-25).
  - then_ask traces to the lectern primer (sloped surface, motorized adjustment) and the ADA standard (knee clearance).
  - makes_irrelevant: a lectern holds a book or notes, not a rated canvas weight or a paper pad; none of the Mabef or flip-chart facts apply to a reading desk. `max_item_size_raw` and `base_type` added after the blind check: a lectern's reading surface isn't sized by the "max item size" concept this register defines for canvases/signs, and its base (post or cabinet) is already implied by `reading_surface_tilt`/`knee_clearance_accessible`, not a separate buyer decision on this branch.
- when `showing or displaying`: then_ask `holds_item_type`, `max_item_size_raw`, `base_type`; makes_irrelevant `max_loading_capacity_kg`, `mast_type`, `canvas_holder_mechanism`, `reading_surface_tilt`, `knee_clearance_accessible`, `sound_system_builtin`, `pad_size`.
  - then_ask traces to Testrite's sign-holder/pedestal/literature-holder taxonomy.
  - makes_irrelevant: a sign or plate stand has no canvas mechanism, reading surface, built-in sound, or flip-chart pad clip. Unchanged by the 2026-10-25 check: this leg already covered all ten branch-relevant facets.

**Split 2, `portability`** ("Folds down to carry, or stays set up in place?")

- when `travels with you`: then_ask `folded_or_stored_size`, `weight_empty_kg`; makes_irrelevant `wheels_or_casters`.
  - then_ask traces to the four Mabef weight figures and to Wikipedia's "Field/Plein air easels: Portable with telescopic or collapsible legs."
  - makes_irrelevant: something folded and carried by hand is, by definition, not rolling on casters while stowed; Wikipedia's field-easel description is "telescopic or collapsible," never wheeled.
- when `stays in one place`: then_ask `wheels_or_casters`, `base_type`; makes_irrelevant `folded_or_stored_size`.
  - then_ask traces to the flip-chart primer's "wheeled mount" stand type and Testrite's pedestal-stand category.
  - makes_irrelevant: a stand that stays in one place has no reason to state a folded size, because it is never folded for transport.

**Split 3, `height_adjustable`**: no branches. Its answer narrows the set on its own (fixed-height versus adjustable products) but does not change which other facets matter on any of the four `primary_use` branches — a fixed-height easel still has a mast type and mechanism; an adjustable one still has the same. Per the skill's own instruction, "questions that are independent of each other need no branches," so none are written, rather than inventing an untraceable leg.

## Trimmed detail

- `mechanism.plain`: "the same idea reduced to one job" (display stands) originally continued "— a pedestal or a flat frame, nothing that bears a painter's weight or a reader's notes" before trimming to the 150-word cap; the dropped clause is covered by the facet-level `base_type` and `holds_item_type` text instead.
- `canvas_holder_mechanism.why_experts_care`: a clause naming the specific models ("Mabef's M/02 and M/04 use a crank; M/06 and M/10 use a ratchet or lever") was cut for the word cap; the model-to-mechanism mapping is kept in the sources section and the data-points table above.
- `knee_clearance_accessible.why_experts_care`: the sentence originally also named the knee-clearance depth figures (11 in at 9 in height, tapering to 8 in at 27 in); kept in the facet's own `backfill_note` is not quite right either — it is kept in these notes' ADA source entry and in reference `ada_2010_standards`'s supports line instead.
- Seed answer 2's display-stand "insert thickness" trap is kept here in full (see Seed) rather than compressed into the `holds_item_type` facet, since it did not survive the data check as a sourced fact.

## Could not reach

- Blick (dickblick.com) and Jerry's Artarama easel buying guides: 404 (guide path) and 403 (buying-guide path) respectively; Blick's own product and search pages render client-side with no product links in the fetched HTML.
- Testrite product-detail pages beyond the two category pages read (no numeric weight-capacity or insert-thickness figure found on either category page read; individual product pages were not linked in a form curl could follow without a product ID).
- displays2go.com (a dedicated display-stand retailer/spec source): 403 on both category pages tried (Cloudflare-style block), with and without a browser user agent.
- AmpliVox (the US lectern/sound-system brand): the domain `amplivox.com` resolves to an unrelated UK hearing-diagnostics company of the same name; no working domain for the lectern brand was found among several guesses.
- Oklahoma Sound / National Public Seating lecterns: `oklahomasound.com` redirects to `nationalpublicseating.com`, which returned a bot-check redirect (HTTP 307) to both `curl` and `WebFetch`; the Wayback Machine snapshot of the old Oklahoma Sound homepage returned only an empty shell (archive.org's own JS-rendered page), and WebFetch cannot reach `web.archive.org` at all in this environment.
- Jack Richeson, Best/Beka, Jullian (French box easel maker): none of `jackricheson.com`, `bekaeasel.com`, `jullianusa.com` resolved; `jullian.fr` redirects to an unrelated packaging company, suggesting the brand has changed hands or domains since.
- WetCanvas (an art-community forum this register would have liked to read for first-hand "what to look for in an easel" advice): the specific forum URL tried returned 404; no working alternative was found without a search engine.
- CPSC and ASTM were not fetched directly for an easel- or lectern-specific load or tip-over standard; judgement 1's "no standard found" rests on its absence from every spec sheet, primer and regulator page actually read, not on a direct search of those two bodies' own databases.
- A second source (beyond the ADA standard itself) explicitly naming lecterns or podiums in an accessibility context was sought via the Access Board's own guide series (no chapter-9 guide exists in their list, confirmed by fetching the guide index) and via ADA.gov business-brief URLs (404). Judgement 2 is worded as "likely," not settled, because of this.

## Script/tool problems

- Mabef's official 2025 "mini catalog" PDF (25 MB) produced zero characters from `pdftotext -layout`; `pdfinfo` shows it as a 2-page, very wide (poster-format) InDesign export, almost certainly image-only despite the PDF having a text layer flag. Not a skill-script bug (this is `pdftotext`, a system tool, not one of the skill's own scripts); recorded here only because the brief asked for any tool misbehaviour to be logged. The skill's own `validate.py`, `render.py` and `viewer.py` ran clean with no errors on this register.

## Corrections after the blind check (2026-10-25)

1. **CONTRADICTED, fixed.** `ada_2010_standards`'s supports line said "8 to 17 in deep"; re-fetched the live standard directly (curl, stripped tags) and confirmed §306.3.3 gives 11 in deep minimum at 9 in height and 8 in deep minimum at 27 in height, with no "17 in" figure anywhere in §902 or §306. Changed the supports line to "knee clearance 27 in high, 30 in wide, 8 to 11 in deep minimum" and fixed the same wrong number in this file's "Trimmed detail" section, where the error originated.
2. **PARTIAL, fixed by reframing.** `pad_size`'s definition and evidence line claimed "two common sizes"; `wikipedia_flipchart` actually names four (A1, B1, 25x30in, 20x23in) on a continuum by country. Could not reach a manufacturer spec sheet stating two fixed SKU widths (3M and Quartet pages 403'd; WebSearch budget was already exhausted for this register, confirmed still exhausted on this run). Reworded the facet's `definition`, `why_experts_care` and the `cold_start_question` to describe a larger-or-smaller choice across a size range instead of a clean binary, and reworded the reference's supports line to name the actual four sizes and the country-dependence.
3. **Unsupported inference, fixed by labelling it.** `mast_type`'s "single mast" value for M/06 and M/10 is never stated on either page; re-fetched M/06 directly and confirmed its own text ("Extensible sliding mast") never uses the word "single" either. Added a line to the facet's `backfill_note` stating plainly that "double mast" is verbatim (M/02 only) and "single mast" is inferred from its absence elsewhere, not quoted. Also broadened `evidence` from just `mabef_m02`/`mabef_m04` to all four Mabef refs, since the inferred claim covers all four models, not just the two already listed.
4. **Branch completeness, fixed.** Added the facets the checker listed as unaddressed to each leg's `makes_irrelevant`: `base_type` to `painting or crafting`; `max_item_size_raw`, `reading_surface_tilt`, `holds_item_type` to `presenting to a room`; `max_item_size_raw`, `base_type` to `reading or speaking`. `showing or displaying` was already complete, left unchanged. See the updated "Branches" section above for why each addition holds.
5. **Backfill-cost inconsistency, fixed.** Raised `mast_type` and `canvas_holder_mechanism` from `backfill_cost: "medium"` to `"high"`, matching `max_loading_capacity_kg`'s `"high"` rating, since all three need the identical per-model brand-page lookup and the first two are, if anything, less obviously named than the capacity figure.
6. **`base_type` source-honesty gap, fixed.** Added a `backfill_note` to the facet itself (previously this caveat lived only in this notes file) stating it is mostly read from the product photo, not listing text, and that the title should be treated as a secondary check against the photo rather than the primary source.
7. **"Assume fixed" default bias, fixed.** Added a `backfill_note` to both `height_adjustable` and `reading_surface_tilt` stating that treating a listing's silence as "fixed" will systematically undercount adjustable products whose listings don't mention the feature, rather than being random noise.
8. **Generic-question traps, fixed.** Reworded the `portability` split/facet question from "Will it travel with you, or stay in one place?" to "Folds down to carry, or stays set up in place?" (borrows the facet's own `definition` wording: "folds down to carry"). Reworded the `height_adjustable` split/facet question from "Does the height need to change for different people?" to "One height for you, or shared by different heights?" (anchored to the facet's own `why_experts_care`: a studio easel or lectern "shared" by people of different heights). Neither question names a part unique to this category, since the split itself (portable vs. stationary; fixed vs. adjustable height) is genuinely shared with other furniture categories; the fix ties the wording to language already sourced elsewhere in this register rather than inventing a false category-specific hook.

9. **Consistency follow-on to fix 2 (not separately numbered by the checker).** The `judgements` entry "Flip-chart stands assume one of two paper-size conventions..." made the same two-sizes overstatement as the old `pad_size` facet text. Reworded it to "sized to one point on a range of paper sizes" to match the fixed facet, so the two don't contradict each other; left `confidence: "low"` and the single-source note unchanged, since that caveat was already honest.

No findings were rejected; all eight required fixes were applied as the checker specified, or as the checker's own stated alternative (reframing `pad_size` rather than re-sourcing it, since no second manufacturer source was reachable).