Algorithms: An outside view

When I began to see the application of algorithms in computer vision, among other fields, I was faced by what I had to acknowledge: That there isn’t really an “efficient algorithm”. The subject itself was akin to doing a double take on all these different approaches to performing various tasks. It was about re-visiting these abstract ideas and using technology’s gifts to bring them out of abstraction, into physical existence. By studying algorithms, I was really studying all these different abstract methods of performing tasks, the process of bringing them from abstraction to physical existence and comparing their effectivity in terms of performance, complexity and ease of deployment, among others, with respect to performing certain tasks.

To think of an algorithm as being not as good or better than an other seems primitive or simplistic. For each task, given the number of variables, different approaches and nuances could consume varying levels of time, etc. But I also realize the importance in generalization in algorithms just as with anything else in this world. It is not practical to break everything down to its constituent indivisible ideas (if that is even possible) and try to analyze and discuss it in said form.

It would help, perhaps, though, if a larger section of society, particularly the ones studying these subjects came to appreciate this fact and were aware of it every time they made a generalization. The O(n) for time and space may not always be defined using sufficient detail and thus, must be used with caution. A combination of dogmatic lines of thought and laziness would attempt to disregard this fact.