Algorithm – Programming Series (1 of 3)

After several years of programming in various languages for several different platforms there are several things I’ve come to realize about programming. I’ve broken it down into 3 separate topics, each with their own discussion. Each discussion will be added over the next few days.

The heart of any program is the algorithm that actually performs a task at hand. There are often plenty of auxiliary functions that typically must also occur in a program, but what makes each and every piece of software unique is how these functions are tied together to perform a task. Often the algorithm is the piece of code that requires the most debugging, the most attention to detail, and most importantly, a new and often creative idea. Auxiliary functions frequently are written by a third party, or in the case of many high level languages, may actually be features of the language itself. In PHP, as an example, there are functions available to query a database, write HTTP headers to a browser, read data from a web page or sort an array. All of these functions typically require no more than a line or two of code for them to perform a lot of work more or less automatically.

Let’s apply these functions to a basic model of the Google search engine. Things start off with a web crawler, an auxiliary function that gets a web page. The crawler then finds links on the page, opens them and continues on indefinitely.

The crawler feeds its data to Google’s algorithm called Page Rank. This is the core of their process. It counts up the number of links to a given page, and what context the page is linked to. Note that there are no pre-defined functions to handle this task, thus it is the main algorithm because it contains the new and creative idea mentioned above.

Next we move back to some auxiliary functions. The data that is generated by Page Rank is then stored in a database, another built in function in many programming languages. Then, when a user does a search, the database is queried and returns a response that contains pages that match the user’s input. Finally, this data is then sent to the user’s web browser, all of which can be handled by various functions that are already built into a programming language.

Please understand that this is a very simplified analysis of Google, and it was intentionally designed to show how an algorithm is a small part of an overall program, but is what makes every program unique.

Posted on March 7, 2009 at 12:25 am by Matt Walsh · Permalink
In: Blog