Live geek or die tryin'

Don't Comment Your Code, Write Good One

Dislaimer: I don’t claim to have the absolute truth and I’m open to your feedback and critics.

Whether you’ve learned programming alone or had classes about the subject, you’ve probably been taught to use comments.
The more I write code and read about good practices, the more I realize that advice is exactly the one you shouldn’t give. It only encourages laziness, writing dirty code, and filling the cracks in by commenting every piece of code.

I believe that most of the time you should not use comments, but instead think every time before you write one: why are you doing this? Is your method or function name not explicit enough? Fix it. Is it complex? Refactor it and move its main parts to new methods or functions. Is it hard to understand? You’re writing bad code.

When I say exception, I mean specific cases where you really need to explain what you are doing: you’re hacking to get the shit done, writing some complex algorithm, etc.
Also, if one needs to specify where you did copy some code from, comments might be a good place for this purpose.

/**
* Source: http://stackoverflow.com/a/9261887/604041
*/
function getPopoverPlacement(popover, element) {
}

The reason why I’m writing this post is that I’ve seen this piece of code

// The logo's copyright checksum
const CHECKSUM = "herpderp";

used to justify the use of comments.

Wrong. I didn’t need to think twice before recommending to name that constant LOGO_COPYRIGHT_CHECKSUM.
That’s quite an easy example to counter, but one could improve almost any code out there as effortlessly as shown above.

Another downside of using comments, as mentionned by my friend Nicolas, is that comments need to be maintained, just like the code.
Not only this means more things to take care of and more work to do, but in real life the code gets updated while comments don’t.

Let’s see some examples of how we can avoid comments by writing better code:

From Jeff Atwood’s Coding without comments blog post

A simple but efficient example:

// square root of n with Newton-Raphson approximation
r = n / 2;
while ( abs( r - (n/r) ) > t ) {
    r = 0.5 * ( r + (n/r) );
}

System.out.println( "r = " + r );

Could be refactored to:

private double SquareRootApproximation(n) {
    r = n / 2;
    while ( abs( r - (n/r) ) > t ) {
        r = 0.5 * ( r + (n/r) );
    }

    return r;
}

System.out.println( "r = " + SquareRootApproximation(r) );

I recommend reading his article where he too, calls to avoid comments.
After all, Jeff Atwood’s word’s more valuable than mine, right? ;)

From the awesome Objects Calisthenics manifesto

Here, it isn’t about comments, but a complex code that a lazy-ass would have used comments to explain:

String board() {
    StringBuffer buf = new StringBuffer();
    for (int i = 0; i < 10; i++) {
        for (int j = 0; j < 10; j++)
        buf.append(data[i][j]);
        buf.append(ā€œ\nā€);
    }

    return buf.toString();
}

Well written, it looks like:

String board() {
    StringBuffer buf = new StringBuffer();
    collectRows(buf);
    return buf.toString();
}

void collectRows(StringBuffer buf) {
    for (int i = 0; i < 10; i++)
      collectRow(buf, i);
}

void collectRow(StringBuffer buf, int row) {
    for (int i = 0; i < 10; i++)
        Buf.append(data[row][i]);
    buf.append(ā€œ\nā€);
}

Is that a lot of code? Oh, sorry. :( I didn’t know you pay a fee for each single LoC you write.

Objects Calisthenics is a wonderful book about improving the object oriented thinking and coding. Read it, and if you can’t follow all the rules, start with some and improve accordingly.

Some code of mine

Take a look at this code:

def find_user_repos(username)
  resp = @http_client.call("/users/#{username}/repos")
  raise UserNotFound.new(username) if resp['message'].to_s == 'Not Found' if resp.is_a? Hash

  repositories_array = Array.new
  resp.each do |repository_hash|
    repo = RepositoryConverter.fill_object_from_hash repository_hash
    repositories_array.push(repo)
  end
  repositories_array
end

Starting from line 5, you have to carefully read the code to understand it. One might think let’s just add a comment!. No.
Let’s clean that mess.

def find_user_repos(username)
  resp = @http_client.call("/users/#{username}/repos")
  raise UserNotFound.new(username) if resp['message'].to_s == 'Not Found' if resp.is_a? Hash

  create_repos_array(resp) do |hash|
    RepositoryConverter.fill_object_from_hash hash
  end
end

private
def create_repos_array(repositories)
  repositories_array = Array.new
  repositories.each do |repository_hash|
    repo = yield repository_hash
    repositories_array.push(repo)
  end
  repositories_array
end

Here, I simply moved the code that creates a repositories array in another method, this has many advantages:

  • One can read the code and understand what’s happening, without caring about the implementation.
  • This method can be used elsewhere.
  • It can be tested… If it wasn’t private.

Where to explain how the big parts of my app fit together?

Probably not in the code itself.
I’d suggest doing this in a separate or in the README file.

Comments