Peshastin · Colour · Archive

The archive.

The background to everything else here — the earlier passes at this problem, kept live at their own addresses. They are worth a look for one specific reason: the colour combinations in them are often lovely. What they got wrong was the logic underneath — there was no proper model of what physically sits together or what can actually be carried where, so some of the pairings are beautiful and impossible. That is the thing the current pages fixed. The colour instincts largely survived.

Start with the pairings if you only open one — it holds 249 combinations with both objects photographed, and its layout is the one carried forward into the current work.

4 earlier documents
249 combinations in the pairings alone
0 deleted
All still reachable
Archive

The earlier rounds, most recent first.

Ten Palettes From Your Shortlist

1 August
archive/palettes-founder-5823e0f936e8.html

Ten colour directions built from your 53 hex picks, each with five colours in roles, a written argument, an honest note on what is wrong with it, and a full contrast table.

Worth keeping
This is the round that got the FORM right, and the current directions document is built on its shape rather than replacing it. The decode of your picks — the amber as a lever, the green as a purchasable product, blue as accent-only, and the observation that your blues all came off rack gear and your reds off woven cloth — all survives and is carried forward.
Replaced by
Superseded by /families, which uses the same form on the deduplicated colour set and adds what to buy for each direction and how each behaves on camera.

Colour Pairings

1 August
archive/pairings.html

249 object-to-object combinations, each anchor paired with two or three matches, with photographs of both objects and the harmony reasoning stated.

Worth keeping
The photograph-led layout came from your feedback and it worked — seeing the two objects rather than two swatches is what made it usable. The round-robin fix that stopped the same anchor appearing on three consecutive cards is a lesson that went straight into the current document.
Replaced by
Its layout is the one carried forward into the current combinations page. What it lacked was not a model of what can move — it had one, and a list of thirty things it refused to move — but the finer question of which objects physically sit together in a room, which is what the staging page answers now.

Ten Palettes — No Input

1 August
archive/palettes-10.html

Ten directions built before your shortlist existed, labelled on the cover as missing the most important variable.

Worth keeping
Three of its ideas were genuinely good and survive in the current round in sharper form: the pop-art piece read as a complete pre-resolved palette by itself, the near-achromatic direction that lets the photographs carry all the colour, and the rack-gear direction that could belong to no other kind of property.
Replaced by
Superseded. It was explicitly a placeholder, made before any colour direction had been given.

Five Colour Directions

31 July
archive/palettes-7a28ebd8ed73.html

The first palette round, built straight off three deep-research reports and the OKLCh harmony analysis.

Worth keeping
This is where the console-blue finding first appeared — the observation that the clearest gap in the category is warm earth combined with a cool console note, and that the gap was already an object in the building. That finding is still load-bearing and is direction 06 in the current round.
Replaced by
Superseded twice over, kept for the record.

The current work is at the library, staging, combinations, the palette directions, colours not on site yet, the alignment page and the palette studio.

These archived pages were built on the colour data as it stood before deduplication and before the item taxonomy existed, so some of them show the same reading twice and some propose moves that are not physically possible. That is why they were replaced — but the ideas in them are still good, and that is why they are still here.