Career Programming

Programmer interview challenge 2, part 4: Using “watchers” to play FizzBuzz “properly”

The Marvel Comics character known as “The Watcher” - “I am known as The WATCHER. My sworn task is to observe and chronicle great events within this sector of the universe. My curse is to always witness and never participate. I must be true to this duty, even to the brink of Armageddon!”
Marvel Comics’ Watcher. His people are sworn to observe, but not interfere in, everything that happens in the universe.

After reading the previous article on FizzBuzz solutions, Reginald “raganwald” Braithwaite, whom I know from my days as part of the Toronto tech scene, tweeted this:

JenniferPlusPlus agrees:

They both make a good point. If you’re playing the FizzBuzz game as the original children’s game and not as an exercise to prove that you can actually write a program, you’d do it like this:

  • Players sit in a circle, not unlike “the circle” in That ’70s Show.(Players don’t have to be high, unless it helps.)
  • The player designated to go first says the number 1, and each player afterwards counts one number in turn. The next player in the circle says 2, and so on.
  • However, for every third number, instead of calling out the number, the player whose turn it is should say “Fizz”.
  • …and for every fifth number, instead of calling out the number, the player whose turn it is should say “Buzz”.
  • The “Fizz” and “Buzz” rules, as the kids would say, stack. In other words, for every number that is both the third and fifth, the player needs to say “Fizz” followed by “Buzz”, or “FizzBuzz”.

So in the spirit of the original game, I’ve put together a FizzBuzz solution that uses “watchers” to keep track of  “every xth number”, with one watcher to keep track of when it’s time to say “Fizz”, and another for when it’s time to say “Buzz”. When it’s time to say “FizzBuzz”, they’ll work in tandem.

I created a class called WordWatcher, which can be summarized as shown below:

Here’s its code:

Some notes about this code:

    • For those of you who aren’t familiar with Python’s approach to class methods, the first parameter for every method in a class is self. It’s the one parameter you don’t fill when calling a method, because Python calls it implicitly (seemingly in violation of Python’s general guideline that explicit is better than implicit). There’s a reason behind this, and it’s explained in this article: Understanding self in Python.
    • Also note that instance variables are declared and defined in the initializer method, __init__(), and any reference to them is always preceded by self.
    • The observe_next_turn() method is meant to be called as the fizzBuzz method proceeds to each new number. It updates the watcher’s internal counter and sets the time_for_word flag accordingly.
    • The speak() method outputs the watcher’s word if it’s time to say the word, or an empty string otherwise.

For FizzBuzz, we’ll need to create two watchers:

  1. One to keep watch for every third turn, at which point it should say “Fizz”, and
  2. one to keep watch for every third turn, at which point it should say “Buzz”.

With the WordWatcher class defined, we can create these two watchers like so:

It will become handy to have these two watchers in the same place. Since the “ha ha only serious” joke about Python is that everything is a list, let’s put them into a list:

Let’s define a fizzBuzz() function that makes use of this list of word watchers:

If you’ve been following the FizzBuzz series of articles, most of this code will be familiar. The part that might need explaining is the part with the comment “This part might need some explaining”.

Explaining the part that needs explaining

Let’s look at the first of the three lines of code in that part:

  • The _ on the left side of the = sign is a throwaway variable. It says “I don’t care about what you do on the other side of the = sign; only that you do something on the other side of the = sign”.
  • On the right side of the= sign is a list comprehension, which is Python’s “show, don’t tell” way of building lists. This list comprehension simply says “call the observe_next_turn() method of every object in the list”.

Let’s look at the next line:

  • This line creates a map that converts the watchers in the list into the words they should say for this turn. If the current turn means that it’s time for any one of them to speak, the watcher will be mapped to the word it’s supposed to say. Otherwise, it will be mapped to an empty string.

And now, the final line:

  • For some reason, map() comes built into Python, but you have to import the functools library in order to use map()’s partner in crime, reduce(). Remember reduce() is a functional programming thingy that takes a collection of items, performs some kind of calculation on that collection, and returns a single value (which you might call a reduction of the collection).
  • The first argument that I’ve provided to reduce() is a lambda — a small function that isn’t given a name — that simply takes the current item in the list and adds it to the previous collected items. Applied over the entire list, it builds a “total”, which in this case is all the words output by the watchers’ speak() methods concatenated together.
  • The second argument is the words map converted into a list. This is the list that the reduce() method will operate on.

At the end of those three lines, current_result will contain one of the following:

  • The empty string
  • Fizz
  • Buzz
  • FizzBuzz

If current_result is still empty at this point, it means that it’s not time for any of the watchers’ words. If this is the case, the string version of the current number is concatenated to current_result:

Here’s the code in its entirety:

And for completeness’ sake, here’s the test file:

You can download and here (2KB, zipped folder with 2 Python files).

That’s a lot of fuss for Fizzbuzz. Why did you do all that?

  1. Reginald asked me to, and I’ve known and respected him for ages, and JenniferPlusPlus seconded the request.
  2. Wait until you see what customers ask you to do.

Did any of this stuff fly over your head?

  1. Don’t feel bad. I had the same trouble when I first learned functional programming, and that was back in 1991, when the computers that ran functional language interpreters were in labs. I spent a lot of time in Queen’s University’s DEClab, which was full of machines that were cutting edge at the time made by a vendor that no longer exists. Computer time, as well as info on any kind of programming, never mind functional programming, was a lot harder to come by. (In case you were wondering, the language we learned was Miranda.)
  2. If you’ve never worked in Python, some of it can be quite weird. It does eventually make sense.
  3. Let me know, either via email or in the comments, if there’s anything you’d like me to cover in greater detail.

Do you have an alternate solution?

I’ve love to hear about it and present it here! Again, let me know via email or in the comments.

What’s next

An elegant JavaScript implementation.

Previously, in the “Programmer interview challenge” series

2 replies on “Programmer interview challenge 2, part 4: Using “watchers” to play FizzBuzz “properly””

Leave a Reply

Your email address will not be published. Required fields are marked *