For as long as I can account, one of the biggest sources of traffic to my intentionally un-promoted site is some embarrassingly unfinished work I did trying to suggest a uniform syntax for URI path parameters. This was something I knocked out as a precursor to some considerably more important work I was doing around a functional REST interface. The funny thing is, it's currently #1 on the GOOG for various and surprisingly distant permutations of its subject matter. It has been for some time, and it's not the only one of mine like it.

I should note that I think search engine optimization, otherwise known as SEO, is largely a fraud. More specifically, I have reason to believe that nobody has any idea about what cocktail of behaviour will consistently yield even close to the highest-ranking search result. I posit this holds even for the employees of the respective companies who work directly on the stuff. What makes it a fraud is when somebody claims otherwise.

From my perspective, SEO breaks down into two categories: active and passive. Active entails hiring an army of people, or more smarmily, robots, to plug links into various places around the Web. Passive is about setting up the content — or ultimately the data — of a given site to lend itself to easy, or more specifically, inexpensive consumption by search crawlers.

To me, active SEO is always at best tantamount to spam. Passive can go two routes. It can either attempt to subvert the search engines' algorithms, or it can simply present itself in a manner that is conducive to their consumption, and computationally cheap to process. After all, CPU time is money.

In my experience, successful passive SEO is not something you can tack onto a Web property after it's done, but rather weave it in as a part of the development process of the site. I believe this because my decade in software and the Web has taught me that the most innocuous or trivial cut corner, early enough on, can irrevocably hobble an expected outcome. Likewise, it can cripple a genuine attempt to present, as I mentioned, a cohesive picture of what the content is about in a way that is cheaply accessible to dumb machines that really don't have a lot of spare time.

It occurs to me then that much of an SEO consultant's job can be replaced by software, and the remainder can be supplied by information architects and content strategists, whose principal objective is to organize knowledge and communicate with people, rather than game the SERP.

This all said, there may still be some value in active SEO, which ideally demands human attention. The question here is how much attention, and from whom. I once visited a company who boasted that they had an entire floor in their building, in the order of about 30,000 square feet, dedicated to search marketing. I couldn't help but wonder at the time how wise an expenditure that was. Whether the money is spent in Bucharest, Bangalore or Los Angeles, you can still spot a dodge a mile away, and therefore eventually so will the crawlers. Likewise I wonder if the people involved couldn't have been doing something more rewarding, or at least more effective, with their time.

I can't help but imagine that a site with useful content that is clear of clutter and conducive to linking by regular people will have the bulk of its active SEO done in kind for the value it provides. The passive stuff, if you can successfully command the representation of the data from the URIs, HTTP headers and HTML markup on upward, should likewise take care of itself.

As for myself, I really wasn't trying. Go figure.