In response to my Knitting Practices post, on Facebook my father commented “The word “practice” is apt. Is there an influence from yoga?”

The answer is obviously “yes,” though the route is somewhat indirect and travels through a story about programming. Stick with it for a little while.

Generators and Python Memory Efficiency

I was talking with a dancing friend about memory efficiency in Python programs, particularly with regards to loops and the range() function. Say you want to do something 1000 times, in Python the easiest way to do this is:

for it in range(1000):
   do_thing(it)

This says, “make a list of numbers from 1 to 1000,” and then call the do_thing() operation on that number (i.e. assigned to the variable it). range(1000) evaluates to a list, which Python stores in memory and the code above loops over.

The way to get this effect in other languages (JavaScript below) is to do something like this:

for (var it=0; it<= 1000; it++) {
   do_thing(it)
}

The for loop has three statements: an action to perform running the content of the loop (create a variable), a condition that will terminate the loop by returning false (variable is less than or equal to 1000), and an operation to run after each loop. The effect of the JavaScript is the same as the Python, except that the Python has to build this (potentially large) list for grins.

There’s a couple of quick answers to this specific question:

1. (In Python 2) use xrange() which is a special iterable type (i.e. it works with loops), that doesn’t need to build a list in memory, it just spits out values incrementally as needed.

   Let’s imagine that it does this by having a function that returns values starting at a certain point (i.e. 0) pausing after each value, and then returning the next value the next time it runs. Thes ea re called “generators” in Python.

2. Wait for the Python 3 switch to complete: In Python 3, range() is a generator of sorts, and it’s efficient in this way.

3. Not care about memory use so much. In most cases this will not be the bottle neck in your application. Really long lists of integers may take up megabytes of memory, but that’s not a huge deal.

Generators are great, and they’re worth using in most cases, but no one will laugh at you if you don’t use them in this situation.

The coolest thing, really is that you can really easily write generator functions. Here’s a silly example: :

def animals("input_animal"=None):
   for animal in [ 'cat', 'dog', 'cow']:
      if input_animal != animal:
          yield animal

This function returns an animal, as long as the input_animal isn’t in the list.

I explained how generators work to my friend and he said something like “Nifty, I guess they’re not really part of my Python practice yet.”

The phrase sort of stuck with me.

Programming is a Practice

There’s a bunch of theory to programming that’s grounded in how computers work, and a lot of things that are actually useful for programmers to know but in truth programming is a practice. Being able to look at a design or a program and understand how to write code to achieve some goal or connect two pieces of functionality, is really about the practice

Beyond understanding what makes good software good and how the machines work, most of programming is knowing the tools really wall and being able to identify the right tool for the job. Everything else is just typing and flow.

While the fundamental knowledge and knowledge of the tools is always important, at some point programming becomes more about being to figure out what the right solution is for any given obstacle or situation.

Knitting as Practice

Knitting is kind of the same. There are some fundamental skills (stitches, operations) and there are some basic fundamental modalities/patterns¹ that you use and combine to make some common objects.

Knitting itself is always about repetition, stitch after stitch, but it’s also about repetition on a higher level. Knitting the second sock, the second sleeve. Making pair of sock after pair of sock, or sweater after sweater.

At some point being a knitter stops being about the skills and modalities/patterns and starts being about actually making things and figuring out how to apply what you know about knitting to a new situation: the thing you want to make, the yarn your using, and the tools you’re using.

It’s all practice. All the way down.