Mistakes are Lessons in Disguise

Every once in a while you get stuck in a rut. For the past four days I've been plugging away at a number of different data-mining algorithms for the Netflix challenge, and I have nothing to show for it. To make matters worse, what I thought was an insightful idea for a term paper for a clustering course now appears to be a suboptimal performer. I'm an application person, and writing literature reviews is not my idea of fun. Alas, it was not meant to be, it looks like it will have to be a review after all. After pondering this depressing thought for a good part of my day, I turned to my bookmarks to find some inspiration.

A few words of wisdom

Fail, fail again, fail better - that one drove it home. I did not fail, I kept failing better. In retrospect, the amount of practical knowledge I gained is enormous. As they say, information about absence, is not the same as absence of information - now I know what will fail in practice. I built a lot of scaffolding, and learned a lot along the way. It all depends on you frame of reference, and once I changed mine, I quickly realized that some of my work could be transformed into a much simpler but a very powerful application - a Ruby driven recommendation engine. That is, once I complete my literature review! On that note, let me leave you with this:


Ilya Grigorik

Ilya Grigorik is a web performance engineer and developer advocate at Google, where his focus is on making the web fast and driving adoption of performance best practices at Google and beyond.