A slider had a floor on a parameter that has no optimum (textstock fill depth)

Standing Rule 8.

Sweep a parameter before you style it, and never floor one that has no valley. Before adding any control, name which of the three it is — valley, wall, or consequence — because a wall belongs to the physics and a consequence belongs to nobody, and neither is yours to cap.

**What was seen** TextStock's "fill depth" (how many words are packed into the shape) could not go below 1000. Every stock came out dense; a sparse, readable one was unreachable. Pawan: *"number of words to fill lower range is 1000 let it be 50 to 3000 let users select what they want."* **Root cause** The floor was protecting nothing. Fill depth has **no valley**: more words are monotonically denser, forever — there is no optimum in the parameter to find, so 1000 was simply one person's taste hard-coded as a limit. The real limit is not density but **legibility**: the scale where a word stops being a word and becomes texture. That is a *wall*, not an optimum, and walls belong to the reader's eye and output size, not to the slider. The frame comes from Pawan's own `form-under-pressure.html` (seven fractal plates, 2026-07-16), which separates three reasons a form is what it is, and they are not interchangeable: - **a valley** — two costs oppose, so an optimum exists (vessel radius: the cube law falls out); - **a wall** — no tradeoff at all, it just stops being worth it (the lung stops at generation ~23 because around 16 bulk flow ends and pushing air buys nothing — *"the lung does not stop where it is optimal, it stops where it is pointless"*); - **a consequence** — nothing is optimised, the shape is just what the physics did (DLA dendrites). Its sharpest admission is Plate I: the tree's *length ratio* is genuinely pinned by a valley, but the *branch angle* has almost no optimum — sweep it and the valley hardly moves — **and you cannot tell which is which by looking at the tree.** Two sliders, side by side, one paid for and one very nearly free. Our sliders are in exactly that position and we had never swept one. Plate V carries the matching warning, and it is aimed at us: *"look for the compromise on this plate and you will invent one."* Earlier the same day I did precisely that — built a tradeoff story about career skills before checking whether the data had a valley or a consequence. **The fix that shipped** `design/textstock/TextStock.html`: `#fillDepth` range 1000–4000 → **50–3000**, step 250 → 50 so the low end is actually selectable. No floor, because there is no valley to protect. (Same session: `.wg-s` weight edits and `#gridAdd` now set `_userW` so a hand-typed weight is not overwritten by the position ramp — a different bug, same theme: the tool silently overriding a number the user chose.)

More from the same record, pulled by shared topic — not written new for this page.

Standing Rule 8. Sweep a parameter before you style it, and never floor one that has no valley. Before adding any control, name which of the three it is — valley, wall, or consequence — because a wall belongs to the physics and a consequence belongs to nobody, and neither is yours to cap.