Magic Numbers

Magic is awesome. Take it as a fact or be a cynic, but it won’t change its awesomeness. So what makes magic so awesome? If I had to put it in a sentence or two then I’d say, it’s the intimacy of illusions, drastically diminished distance from the seemingly impossible and the ever green ability to make things appear and/ or happen out of nowhere. Magic does things that are hardly aligned with logic. The very essence of it is in the opaque interface that separates the miracle to the eyes from the marvel of the brain that made it all happen. The fundamental idea behind magic is nothing but an abstraction so absolute that the lines between reality and fantasy are blurred to the extent of near extinction. The magician desires nothing but an induction of belief in the audience’s mind that something can be done, without them having to go through the details of how or why. In fact I’d go as far as saying that the word magical can safely be used as an antonym for the compound self-explanatory. Magic can’t be understood, or else how would you expect it to amaze.

The train of thought I’d like you to carry from the previous paragraph is that magic is esoteric and that too intentionally. Magicians must hide how they did it otherwise the world is full of keen eyes and ambitious minds to figure things out. As I’d introduce in the subsequent paragraphs, such keen observers are sometime benevolent as well. Sometimes it’s a bad idea to keep them from knowing the secret.

A similar trait of esoteric abstraction can be seen in descriptive solution to a problem. An example of this would be a computer program. A computer program is based on an algorithm, the logic of which is meant to solve a problem. What this logic contains (commonly) is a combination of instructions to be followed and decisions to be made, by the machine. Often these solutions are dependent on very specific data. This dependency is so specific that the programmer does not even have to store this data in a variable. Why? Because the data is not going to vary. All that needs to be done is writing that data down, in your program and let it work in order to obtain the solution. For example, if I am writing a program to double a given number then somewhere in the body of my program I would have to write the following instruction:

Result = 2 * number;

In this statement, the number 2 is the kind of data that I am talking about. Now this example might not apply to the criticism to follow. It is only intended to get you acquainted with the notion of magic numbers. A rough definition could be:

A magic number is any constant written exactly, as part of an algorithm, without having any contextual representation.

What I mean by contextual representation is anything like a variable, or a macro assigned to such numbers that may specify what their purpose is. It is not always as obvious is in the above example as to why certain numbers are used in a solution. These can prove to be as much of a nightmare as magic in general is like a dream to the observer. If someone is willing to go through an algorithm and figure the logic behind it, then these magic numbers are the primary hurdles among other things. They are not abstracting, they are blatantly obscuring the straight truth: the real, intended logic.

Magic numbers are necessary in many cases, but what the author should avoid is using them directly. Bearing a little overhead and assigning some descriptive macro to the magic number would go a long way in enhancing the clarity of the solution and will ensure its legacy.

Avoid using magic numbers, abstraction should only be to the extent of ignoring details when there are bigger fish to fry. Hiding something (permanently) should never be the intent (speaking strictly in context to this post).