Two Hundred And Three Distinct Answers

The record contains four hundred and five shapes and two hundred and three distinct values of intake.

So the four hundred and five shapes are, as filters, two hundred and three different things. Half the variation in how a list can be built is invisible in what it admits.

Teams seeking a clearer history of implementation effort can take a look at the platform tour, then compare it with reliability and stakeholder feedback. For an independent technical reference, see Kubernetes.

How the duplication is spread

Unevenly. Some values are unique to one shape; the common ones are shared by up to twenty-one.

The powers of two attract the crowds, because those are the values a list reaches by having a single passing route, and there are many ways to build one of those. The awkward fractions — one hundred and sixty-three over two hundred and fifty-six, twenty-one over thirty-two — are each produced by one shape only, because they come from a specific threshold over a specific number of optional items and nothing else lands there.

The practical reading

A list with a tidy intake is probably one of many shapes with that intake. A list with an untidy one is probably unique.

That is a small observation with a real use: if two documents both admit one in sixty-four, they tell you nothing about each other's structure, whereas if two both admit ninety-three in two hundred and fifty-six, they are almost certainly the same shape.

Round answers are crowded and odd answers are lonely. The tidiness of a fraction says more about how many ways there are to reach it than about the list that reached it.

What two hundred and three bounds

It bounds how finely a filter can be tuned within this record.

Two hundred and three values across a range from one to one in sixty-five thousand five hundred and thirty-six sounds like fine control, and it is not, because the values cluster. Near the top, between a half and one, there are relatively few. Near the bottom the powers of two dominate and the gaps between available values are multiplicative.

So somebody wanting a filter admitting "about one in a hundred" will find the nearest available values are some distance either side, and will have to take one of them. The dial has two hundred and three detents and they are not evenly spaced.

Where the clustering comes from

The compulsory items contribute only powers of two. All the variety comes from the thresholds, and a threshold over n items produces only n plus one distinct values.

With the optional part capped at eight, there are not many distinct threshold values to be had, and multiplying them by powers of two produces a set that is dense in some regions and sparse in others.

That is a property of the record's bounds, not of filters in general — a longer optional part would give a finer dial. It is noted here so that nobody reads the gaps as significant.

The two hundred and three, and what they are not

They are not a tidy set. One hundred and fifty-seven of them are reached by exactly one shape, and the remaining forty-six are shared by between two and twenty-one shapes each.

So the duplication is concentrated rather than spread: most values in the record are unique to a single shape, and the collisions pile up on a few dozen values that happen to be easy to reach by several routes.

That shape of distribution is why reading a list is unreliable rather than merely imprecise. A given intake usually corresponds to one way of building a list — but the most common intakes correspond to many, and those are the ones a reader is most likely to meet.

Why this is the figure the record rests on

Because it is the gap between how a list is built and what it does, stated as a single number.

Four hundred and five ways to build a list, two hundred and three outcomes. About half the structural variety available to an author makes no difference to the filter, and there is no way to tell from the document which half any particular list is exercising.

The same filter can be written four hundred and five different ways, and two hundred and three different filters exist. Reading the document tells you which way; only the arithmetic tells you which filter.

What to do with the figure

Mostly, stop looking for a shape that hits an exact target. Pick the nearest available value, note which it is, and spend the remaining effort on the parts of the document the arithmetic does not reach — which, as this whole section has argued, is where the shapes differ anyway.

Near this one

• Length Cannot Be Read As Rigour
  The conclusion this section has been building to, stated plainly, with the three specific errors it rules out.
• Comparing Two Lists Honestly
  The method, in four steps, and the one assumption it rests on. It works even for a reader who rejects the coin
• Twenty-One Lists, One Filter
  Twenty-one shapes in this record admit exactly one profile in two hundred and fifty-six. They run from eight i
• The Ladder Of Classes
  Forty-six values here are shared by more than one shape, and the largest classes form a descending ladder: twe