On Jul 1, 2021, at 11:30 PM, George Michaelson <ggm@algebras.org> wrote:


Note that this could easily have been made a table instead of a bunch of one-line functions.

a table.. hmm. so like.. we could write .. engines to "read" the table
and do things in some kind of (maybe.. finite) way?

I know, lets write a DSL to MAKE THE TABLE...

Is all software "wheel of life" ?

It would be just an operator precedence table and two functions. One to
parse a factor and one for binary expressions. The table might be something
like an array of {singles, doubles, associativity}, where the Nth entry
has precedence N. The binary expr parser uses the table to essentially
group sub expressions based on precedence and associativity. This is an old
idea. I think at least 4-5 decades old.