The Lab Now Opens With What We Turned Down

The Lab is where we publish our working notes: the products we tested and turned down, the products that made it, the research behind our reasoning, and the corrections we make to our own published work.

Until today it was one long list, newest first, ten notes to a page.

What that meant if you actually used it

The list held every note we have ever published. It showed ten of them, and the rest sat behind a Next button with no page numbers, because a styling rule had hidden them. Reaching the far end meant clicking Next seventy-three times.

The order did the rest. Newest first sounds neutral, and it is not. Our most recent work has mostly been notes about our own system, so those filled the opening pages. The first product verdict appeared on page seven. The first piece of research appeared on page eleven. The opening page contained no rejections at all, on a page whose own title promises to show you what we rejected.

Nothing was hidden and nothing was missing. It was simply arranged so that the thing most people came for was the thing furthest away.

What it does now

The Lab is an entry page rather than a list. It opens with two doors.

The first is the products that did not pass, and it is the larger of the two. It carries the count, a plain description, and the most recent rejection with the sentence that note gives for itself. The second is the products that made the cut, deliberately quieter, with the count and three of the newest names.

The two doors are different sizes on purpose, and the reason is in the writing rather than in the design. When we turn a product down, the note has to say what stopped it, and no two products are stopped by the same thing. When a product passes, the verdict line says it passed. Reading every note we have published, the rejections hold almost two hundred distinct explanations and the passes hold three. Building the two doors as twins would have promised you something the second one cannot deliver. The full argument for a product that passed lives on its own page, which is where it belongs.

Below the doors, the essay explaining why we publish any of this is unchanged and now sits where it reads as an argument rather than a preamble. Under that, six research investigations, each showing the study it rests on: the field it belongs to, the author, and the year, on one quiet line. Then the three records we keep about our own work, each with its count and its newest entry named. Then the five most recent notes of any kind.

Nothing was taken away

The old chronological list still exists, complete, in the same order. Everything, newest first at the foot of the page opens it, and every page-numbered address that worked before still works and still lands where it always did. What that list did not do was print the date it was ordered by: every card on it carried the revision date instead, which is now corrected on every lab list.

Two things this does not fix

The doors lead to a page holding all five hundred and eighty five product verdicts, and that page still shows ten at a time. We have moved the crowding rather than removed it, and that page is the next one we are rebuilding.

Both halves of that page have since been rebuilt, and the rejected half now opens by showing how near each rejection came to our bar.

The research investigations shown here all rest on cited studies, and while building this we noticed that every one of those citations was added during a single stretch in January. Nothing published since carries one. That is a gap in our own work, it is not fixed by this change, and we would rather say so here than let a tidy page imply otherwise.