Saturday, April 13, 2013

Hammer Proficiency

Getting Fluent in New Programming Languages

When I first started to learn Python, I wasn't particularly impressed with it. The problem was that searching for an introduction to Python often led to pages that were an introduction to programming using Python as an introductory language, whereas what I really wanted was an introduction to Python for experienced programmers that told me the things that made it different from, say, C. 

So since I couldn't find the introduction I wanted, I referred to the reference docs. The problem with this approach is that it doesn't really get you fluent in a new language, rather it just allows you to write in a language you already know with a different syntax. I'm going to call this 'hammer proficiency', as in if all you have is a hammer, everything looks like a nail. Basically you've got a new tool, and you've learnt enough about it to hit nails with, but not much else.

Here's an example of being 'hammer proficient' in Python. I'm going to write a function that takes a list of strings and returns another list containing only the strings containing the text 'cat', but doing so from a procedural, C-style viewpoint. I know I'm going to return a list, so I'll look up list creation, and I know I'm going to have to iterate over a list, so I'll look up how to iterate over a range of numbers and to find the length of an array. I'm also going to have to look up how to search for strings in a string and add elements to a list. So I would probably come up with something like:

def filter_cat(strings):
    ret = list()
    for i in range(len(strings)):
        if strings[i].find('cat') != -1:
    return ret

There's nothing fundamentally wrong with this. The problem is that any modern language is likely to be flexible enough that you can write code like this, translating almost line-by-line from a more traditional C-style, procedural approach into the new language. But doing so means you don't get to learn the unique and hopefully advantageous features of the program that you're trying to learn.

If I was to write the same function as above now, after having learnt some more Python, I'd probably do the following, using a lambda expression, a list generator and the 'in' operator:

filter_cat2 = lambda strings: [s for s in strings if 'cat' in s]

Which is now using the features included with Python. Only once you start to know not only the syntax, but also the best approaches to take for a given problem for that language, can you really claim to be fluent.

Getting to Fluency

So, how do you get from hammer proficiency to fluency? Experience (and maybe hammer time?) is obviously one way to get there. I didn't find a wealth of information online answering this problem. Like I said, a lot of the introductory information I found started out from too low a level.

StackOverflow has an (unfortunately closed) list of cool features of Python. There's also a really good list of Python tips and tricks here. Finally, I've been doing a few of the Udacity courses (after taking the Stanford 'Introduction to AI' online class). They vary a bit, but the Peter Norvig classes and lectures are all excellent, and his 'Design of Computer Programs' class is too. And because he's using Python, it's serves as an excellent example of how to use the tool, and not like a hammer.

No comments:

Post a Comment