calculator vs REPL

[kirby urner]

https://mail.google.com/mail/u/1/#inbox/1516abb8ede5a1af

http://mybizmo.blogspot.com/2015/08/high-school-stem.html

===

I got to teach over audio / shared screen for four hours again last night.  Adult learners.  More informational than with O'Reilly School (OST) i.e. I don't evaluate individual student work. 

The certificate or credit is for showing up, not for running an obstacle course.

OST is asynchronous and based on written communication.  This other gig is more like call in radio, with live chat message board (text) and the ability to share a microphone, more like the real time technology Maria uses.

As you can see, I'm into dramatizing convergence to a limit by taking such as e = (1 + 1/n) ** n where n -> huge ("infinity" not a number in memory), and proving equality to say 300 decimal places.  This is doable thanks to an extended precision number type that most calculators do not have.

I used the same tools (and testing framework) to converge to Phi, Pi and e.  For Pi, I had one of Ramanujan's crazy sums [1] and another thing I don't understand.  Like, I can see that they're working but I have no idea where these geniuses got such algorithms. [2]

I'm not saying brute force convergence to constants like e, phi and pi is necessarily some holy grail, but it is a way to appreciate how computers take the drudgery out of what look to be tedious beyond endurance calculations.

We also get the benefit of seeing that convergence "literally" i.e. as a result of running code.  Textbooks will show these truths in notation, but there's something satisfying about squeezing actual numbers out of electronic circuitry that's doing one's bidding.

Kirby

[1]  https://mail.python.org/pipermail/edu-sig/2015-December/011345.html

(see link to Standford website for more traditional math notation)

[2]   https://mail.python.org/pipermail/edu-sig/2015-September/011317.html

(code that loops endlessly each time spitting out the next digit of pi)