Old Content:

BigMatrix, The Data Storage of Google PageRank, in Java

February 18, 2009 by Kyle Brady
Tags: Code, Code Samples, Full Code Release, Java
As required by the FTC, a Full Disclosure is available - this piece adheres to the Code of Ethics

---

This is Part 1 of a 2-part assignment:  replicating the original Google PageRank engine in Java, based on their published doctoral work.  This is, essentially, the matrix-based storage engine that they use/used:  large matrices implemented using linked lists for rows, allowing significantly faster access to data.

It also includes a "multiply" method, as this is important in the PR algorithm, which accepts a vector (column matrix), with its size ("height") being the number of rows in the data matrix.  It should be usable as a stand-alone class, provided you follow the parameters.

It was due at midnight tonight, and I was done earlier, so this is a scheduled post.

All code is now available via SVN, and a web browsing utility.

Note:  I am not advocating using this code in place of doing your own homework, or in any situation other than pure interest.  However, in the event that you do, I am not responsible for any resultant issues (legal, moral, or ethical) that may arise.  Finally, I am not presenting these as perfect cases, the best way to approach the problems, nor the fastest.  They are, simply, viable solutions.

Old Content posts are leftovers from a less structured, less civilzed era that are kept for posterity.
Kyle can be found on Twitter and MySpace, or reached via email.