The docs for our expression arguments are totally unclear, and only usable with examples. The problems are:

1. The difference between number and scalar is confusing. It should be obvious (somehow) that scalars are just normal numbers, and a "number" is the result of an expression that is a series. The tagged vs untagged thing here is hard to explain.
2. Arguments are sometimes named and sometimes typed. That is, sometimes we describe the type of the argument, and sometimes we put a human-readable name as the argument. We need to do both.
3. Return types are also poorly documented.

We need to come up with a standard format to describe these. Here's a proposal:

## func_name(var_name var_type[, name2 type2]) return_type

Example:

## forecastlr(s series, y_val scalar) number

Returns the number of seconds a linear regression predicts s will take to reach y_val.