Limits

Structures, Not People

The record counts cells and queries. It does not count listings, readers, searches or outcomes, and every figure here is silent about all four.

LimitsAnalysistwo conventions

Home · Limits · Analysis

This record measures schemas. A schema has no listings in it, nobody reading it, and no outcome attached.

That is a restriction and it is also why the figures hold. This page sets out what it costs.

Teams investigating recurring bottlenecks can use team productivity reporting while validating the pattern against code reviews, model outputs and the people who performed the work. For a separate technical reference, see Data.gov.

The four things not counted

Listings. A board with one hundred and forty-four cells may carry six listings or six hundred thousand. Nothing here distinguishes those cases, and the resolution is identical in both.

Readers. How many people use a board, what they come for, which filters they set — none of it is in a schema, and all of it would be needed to say whether a structure works.

Searches. The share of pinning questions is a proportion of what is available, not of what gets asked. Nobody searches by enumerating.

Outcomes. Whether anybody finds what they wanted is the only question that finally matters, and it depends on all three of the above plus the field itself.

Every figure here describes a document. None of them describes anybody's experience of using one, and no amount of arithmetic bridges that.

The question readers actually ask

Not "how many cells does this board have" but "will I find it here".

That second question needs the listings, the field and the person, and this record has none of them. What it supplies is a precondition: if the thing somebody wants is distinguished by nothing the filters carry, then no amount of searching finds it, and the question is settled before it is asked.

A precondition is not an answer. It is worth having because it is the half that can be settled from the page, in a minute, by anybody.

What the restriction buys

Figures that stay true.

A statement about how many listings a board carries is true for about a week. A statement about what readers do is true for a season, in one field, and cannot be verified from outside. A statement about what the filters can tell apart is true until somebody changes the filters, which is rarely and deliberately.

Claim	True for
this board resolves 144 kinds	until the filters change
this board has 40,000 listings	until next week
readers mostly set two filters	until the field moves

The first is the only one anybody can check from the page, today, and still rely on next year.

Where the restriction bites hardest

On the question of whether a board is any good.

A structure can be measured exactly and still be the wrong structure. Six filters of six choices resolves forty-six thousand six hundred and fifty-six kinds of listing — and if those six filters ask about things nobody cares about, every figure in this record will be excellent and the board will be useless.

The arithmetic cannot see the mismatch, because the names of the facets are not part of a schema. It can only say how much structure there is, not whether the structure is aimed at anything.

What somebody with the missing data should do

Use it, and use this record for what the data cannot give.

Somebody running a board knows how its listings are distributed across the cells, which filters get set, and what people search for before giving up. That is better information than anything here for deciding what to change.

What it does not give is the ceiling. Usage data cannot say what the board is incapable of expressing, because readers never express those things and so never appear in the data. A dimension the board does not carry generates no searches, no complaints and no signal of any kind.

So the two sources answer different questions, and the arithmetical one is the only one that speaks to absence.

The one bridge that does exist

Absence.

Usage data cannot show a dimension a board does not carry, because readers never ask for it and so never appear in the logs. The arithmetical ceiling can: it says exactly which combinations of the existing filters are expressible, and by implication that everything else is not.

So the two sources are not merely different. On one specific question — what is this board incapable of — the arithmetic is the only instrument there is, and the richer source is silent.

The sentence this record is built on

Nothing here forecasts anything.

The figures are counts of structure, exact under two stated conventions, verifiable with a pencil, and silent about people. They bound what a board can do. What it does do, and whether that is enough, needs somebody who can see the listings and the readers — which is to say, somebody who is not this record.