Re: [MIT-Scheme-devel] help with tracing

[Catonano]

2014-11-18 19:16 GMT+01:00 Matt Birkholz <address@hidden>:

> From: Catonano <address@hidden>
> Date: Tue, 18 Nov 2014 17:19:32 +0100
>
> [...]
> I just wanted to provide sensory proof of the substitution model.

Did you try Dr.Scheme / Racket?  I've heard they can generate some
very nice diagrams of Scheme machine states.

I tried but the first thing I noticed was that the square function in called "sqrt" rather than "square"

I don't wanna be in a field where teachers in the videos call functions and people have to elaborate to map those names on another library.

I could get confused myself.

But I didn't know that Racket can generate these smart diagrams. Note taken.

 

> And by the way, I feel it's important to instill mechanical
> automatism in them,

I did an Internet search on "mechanical automatism" but usage seems
all over the map.  ?

> as it could turn out to be useful, should they ever run into formal
> study of the lambda calculus. Or even into informal attempts at
> comprehension.

Is the lambda calculus mechanical?  I thought it was symbolic
translation (a castle of cards), like any algebra.

In order to teach arithmetic in primary school, some reflexes were instilled into me. To add, for example. For multiplying it was worse. I had to learn the multiplication table as a spell to repeat with no consciousness attendance.

In my view, you need sensory experiences as bases to build abstractions on. Not the other way around.

While I was demonstrating the substitution, one of them asked me "dude, your point is to calculate a square root, right ?" hinting to that I was making a storm in a glass of water

I can't talk these people about algebra, let alone algebraS. Beware of algebras ;-)

In order to be reached by awareness (if and when) they'll need solid bases laid out some time before

And even if they won't be reached by any awareness, they still need to be able to run the substitution in their minds fluently in order to use scheme proficiently.

So, I'm going to adhere to the recommendation of the professor in the video (I don't remember if it was Abelsson or Sussman) that for the next few lectures the substitution model has to be held religiously and when they won't understand something, they'll have to apply it "mechanically" (as if they were machines themselves)

The understanding, hopefully, will come later

 

I wouldn't worry that your elementals will run into formal study of
the lambda calculus.  30 years ago, when the lambda calculus was
considerably younger, I couldn't find a class at MIT devoted to it.
Turing machines: yes.  Lambda calculus: no.

It's strange. It was the same for me. I was bombed with the Turing machine and I was lied to about the lambda calc. Lisp was hidden as a shame.

I discussed a thesis about an exotic equivalence between regular languages and another martian formalism (in the multidimensional case, because they were exploring the generalization of regular languages stuff). In order to do that I had to depict a complete overview of all the known equivalences between regular languages and other stuff. All of them BUT the lambda calculus.

When I discovered (years later) the equivalence between the Turing machines and the lambda calc I was like WHAT !?!

I was lied to. I don't wanna lie to these people.

 

> It's just a matter of being honest to disciples !

By "honest" you mean "perspicuous"?

Yes, I mean perspicuous. When teaching, the attempt at perspicuosness is due. It's honesty.

 

You aim to "instill" an "-ism" in "disciples"?  You're creeping me
out, dude. :-}

And you still don't know that I was mumbling to keep the Y combinator as a smart diversion, just to be sure to instill the complete -ism with no sloppiness ;-)