Okay

welcome to SMI.

It looks like you did rather well on the quiz.

I'll have to see what's happened here

but that went well.

So

we talked about grammars

tree grammars actually

mostly.

And the idea is that we can very often

almost always in practice

in symbolic AI

divide

grammars and grammatical string processing into two phases.

One is kind of thinking of the objects as trees.

That's what we do with the abstract grammars.

And thinking about their communication form

you may want to say that

basically

what

is the serialization of the trees into something I can store in a file.

And so that gives us this distinction between the abstract grammar and the concrete grammar.

The abstract grammar is typically what we're thinking about in terms

what we're thinking

about in class

where we're thinking about the principles behind things.

And then

of course

there's the engineering stuff

basically

what goes into the file

how can I manage interoperability and all of those kind of things.

Those are more

they're less conceptual

more engineering type questions.

Both of them are important.

In courses

we typically concentrate on the more conceptual concerns.

I would like to kind of carry that a little bit further.

When you look at math

mathematical expressions

and you think of them in terms of an abstract

grammar

then you'll see that there's actually only five kinds of things we need to construct

a particular language for math.

Right?

We have literals

like numbers

right?

555 or something like this we had in our arithmetic examples.