This site presents a method for inverting a matrix that may be singular or non-square. Included on this site is a demonstration page where you may enter a matrix and have the script calculate the inverse. The information herein has been around for decades, and nothing I make is ever new or unique, but it was good enough for a successful dissertation. I've worded this presentation in a way that skirts discussions about engenvectors, eigenvalues, or other topics that may scare off anyone trying to learn an easy way to solve a linear algebra problem.

I misplaced the original matrix inversion pages, so I've had to recreate these from scratch. I refused to pay US$65 semi-annually just to have this site hosted on a Java server so my servlet could perform calculations and return the results to the page.

The servlet is now replaced with JavaScript and runs entirely on your computer. (See the demo page.) This has the advantage of making the entire site run from static pages; a cheap hosting option. It also puts the burden of calculation on your computer. :-)

The instructions begin on this page. If you haven't been through the instructions, I'd recommend doing so before jumping off to the demo page, or grabbing the source and just blindly using the code. Note also that the sources and script do some additional shuffling of the data to avoid numerical instability (I choose to never divide a large number by a small one, for instance.)

I've written rather a lot of JavaScript for this site. If there are any problems with this site, they're entirely my fault (design, layout, mathematics, xhtml, css, JavaScript, etc.) I've tested much of this on Firefox, Opera, and Safari, Konqueror (various versions), but not on IE (I don't really care if it works there or not.)