What is the big deal about Big-O notation in computer science?

Question

How would Big-O notation help in my day-to-day C# programming? Is it just an academic exercise?

Answer 1

There's no "big deal".

It all depends on the kind of work you're doing. If you're working on front end, you could spend months mired in all kinds of interesting and potentially complex things that don't have anything to do with Big O.

If you're working in an organization suffering with scaling problems, you might find yourself an out of the box solution that suits all your needs, and you only need the ability to understand what Big O means, to properly understand the kind of performance you're going to get when you call X function. Or, it may be important for you in your work occasionally when you have to tie pieces together, or write a new algorithm that is a composite of others....

The final case (<1%) is that you're working in academia, where of course the discovery of a new algorithm that is an improvement in order is a huge deal potentially, and it's going to be very important to your daily work. No one is probably going to have to tell you this, since it will be impossible to proceed down this path without recognizing that importance.

When it comes to the interview process, it's a different matter altogether. I'm afraid it's a bit of a hazing process amongst us engineers. We do it to each other, but really we all let some of our knowledge that isn't as useful to our daily work degrade over time. Like most engineers, you're going to brush up when it becomes useful, so this isn't really a concern, other than the fact that when you quit / get fired you're going to have to interview again, so... best to simply chalk it up as one of those annoying things that humans do, and simply sacrifice the time required in preparation for the interview process. I like to think of it as honor-based. The courtesy of studying up on my algorithms shows honor to my next potential employer. Of course, they may see it differently, but that's not my place to say :)

Answer 2

Big O notation allows you to analyze algorithms in terms of overall efficiency and scaleability. It abstracts away constant order differences in efficiency which can vary from platform, language, OS to focus on the inherent efficiency of the algorithm and how it varies according to the size of the input.

Answer 3

Naw, I was wondering that too, but now I find myself thinking about big-O just about every time I use a library.

Big-O lets you know the asymptotic running time of any function, that way you can decide whether data structure A is faster than data structure B for your purposes.

For example, you might be tempted to use something like an ArrayList when what you really need is a Queue. When you try to add an element to an ArrayList, if you can see that the running time is O(n) (because it needs to create a new array and copy all the elements over... sometimes) but in a Queue it's O(1) then you can easily see that the queue would be faster. This is actually kind of a poor example as there are many other differences between these two structures, but you get the idea ;)

Answer 4

Big-O is a means of measuring or meaningfully ball-parking the performance of an algorithm in terms of time. So if any optimization needs to be done in that respect, big-o is a valuable tool. It is a foundation chapter in algorithms and data structures classes. I agree with other replies mentioning that you might not use it directly in your day to day programming work, but even that day to day code has a performance that can be measured if required.

Answer 5

Writing good software is largely about understanding and making informed decisions about trade-offs in your design. For example, sometimes you can tolerate a larger memory footprint for faster execution time, sometimes you can sacrifice execution time for a smaller memory footprint and so on.

Big-O notation is a formalization of these trade-offs so that software engineers can speak a common language about them. You may never have to formally prove the Big-O characteristics of an algorithm you design, but if you don't understand the concept on an abstract level, then chances are you won't be making good trade-offs in the software you develop.

Answer 6

Think of efficiency, my friend!

The difference can be seen if your boss is yelling at you to find the address of clients by their name and you are given a huge pile of unsorted papers and an address book indexed by name!

In big-O notation, this is O(n) - running through your huge pile of unsorted paper, and O(1) - looking up the index by name.