Powerful Python

Aaron Maxwell's Powerful Python course looks great. It's outstanding because, as far as I can tell, it explains abstract ideas clearly and with practical examples. If becoming exceptionally good at Python ever makes it to the top of my list of professional priorities then I'd certainly take the course. For now though I'm going to make some notes on the marketing emails he send:


Good mental models:

def f(x):
    return x+2

Function f is a nameless function object tied to a variable called "f". You can also think of it as a function called "f", but the first mental model is more useful. It lets you reason about function factories, passing functions as arguments to other functions, and writing decorators.

Also, apparently, Virginia Satir was such an incredible family therapist that people wrote books about how great she was. She once said "Most people think the will to survive is the strongest instinct in humans, but it isn't. The strongest instinct is to keep things familiar".

Staying motivated

Most people think that motivation is about willpower, but its not. It's about creating "wins". When you invest your time and effort, you get that "aha!" moment where you see the computer do something new, or you get that feeling of a little breakthrough. It makes you excited to take another step and go a little further. And get another little win.

95/5 rule

OOP is not just a tool for organizing your code. It's a tool for organizing how you think about code. Here is a roadmap to break into the top 1% of python programmers:

  • Create powerful abstractions that accelerate and amplify all the python you write.
  • Write automated tests, which lets you break through the ceiling of complexity, so that you can suddenly write truly complex software as easily as you used to write simple software.
  • Create robust high-performance scalability - writing software than gracefully handles increasing magnitudes of data without breaking a sweat.
  • Learn how to use the higher-order abstractions that the most important and prominent python libraries are built on (pandas, flask, django, pytest, requests).

Hidden meaning of "("

class Prefixer:
    def __init__(self, prefix):
        self.prefix = prefix
    def __call__(self, message):
        return self.prefix + message

Use it like:

>>> simonsays = Prefixer("Simon says: ")
>>> simonsays("Get up and dance!")
'simon says: Get up and daance!'

What is the type of simonsays? If you look at just the last line you would think its a function, but it is not. It is an instance of Prefixer.

There is a translation step. When you write "simonsays('foo')" it is translated (kind of) to "simonsays.call('foo')". In other words, simonsays( triggers a function call.


>>> def increment(x):
...     return x+1
>>> increment.__call__(2)

You can do it with regular functions too. There is a difference between the python that you write, and the python that the compiler perceives. The ( is replaced with .__call__(.

Current favourite python book

There are 4 levels of Python code:

  1. Syntax
  2. Idioms
  3. Patterns
  4. Architecture

Check out "Architecture Patterns with Python" by Harry Percival and Bob Gregory.

Secret Instinct

Something that the best programmers have in common:

They really hate repetitive code.