Take None Of It

An important concept is that all data has shape, or length along dimensions. A vector has a length (number of items in the list). A matrix has two dimensions (each with a length) for number of rows and number of columns. What what about the "lengths" themselves? The NumberOfRows - what shape should that be?

Well it is a Number. It doesn't have any shape, it is a "scalar".

But if we ask what is the "shape" of a data object, the answer will always be a vector, as there could be as many as approximately 150 "dimensions", each with a length.

So if we want to use any one of those dimensions as just a plain "number" we will need to "scalarise" it, or "un-vector" it.

There are a few ways of scalarising a single element vector - but I like the method of "taking none of it".  

So the method I use is (iota zero rho) rho NumberThatIsReallyAsingleItemVector.