Like the rest of the geek blogosphere, I watched a bit of the Google Press Day event and then checked out some of the new toys they launched, including Google Trends and Google Co-op (or www.google.com/coop — which seems to have more to do with chickens, but then hyphens don’t translate that well into URLs) and I was somewhat underwhelmed — if that’s a word. Particularly with Google Trends, which seems like mostly a buzz-meter a la Alexa, with dubious long-term value, although Steve Rubel has had some fun playing around with it.
But it was Google Co-op that really floored me. This thing makes the incredibly complicated and mind-bogglingly obtuse Google Base look like a Hallmark greeting card. I know that Google is all about algorithms — not just because it’s obvious, but because that’s what Eric Schmidt said during the press conference; he said that Google Co-op wasn’t about social networking, but about making search better, and that for Google pretty much everything was about algorithms. But damn — Google Co-op in its current form is enough to make a non-math geek hold his head in his hands and weep, or head for the Xanax.
It seems simple enough. Google wants you to “Help other users find information more easily by creating “subscribed links” for your services and labeling webpages around the topics you know best.” Great. I know how to label things with tags like at del.icio.us, and I know how to subscribe to things via RSS and whatnot — how hard could this be? Well, pretty hard actually. If you go looking for more detail, here’s what Google tells you:
“In order to use the API you need to define one or more ResultSpecs. A ResultSpec contains a Query and a Response. The Query gives a general trigger pattern of queries for which you want to trigger your result. The Response provides a template for the output you want to display when the trigger pattern is matched. The id attribute of the ResultSpec tag uniquely identifies the ResultSpec. Every ResultSpec must have an id attribute.”
And that’s just the introduction. It gets worse. You have to learn about structured queries and data objects and output methods and formats, not to mention the frighteningly named “extractors” and validators. Obviously, this is for people who like programming in the same way I like ice cream. And I know that like Google Base, this kind of setup is meant for companies and services to set up their own databases and then feed them into the Googleplex, just as Google Base does for used cars and real estate and whatnot.
But still. Couldn’t it be just a little more user friendly for us non-programmer types? As Paul Kedrosky notes, it breaks one of his primary rules, which is that people are lazy. Even Danny “SearchEngineWatch” Sullivan couldn’t make it past the introduction, and Cynthia over at IP Democracy said that Google needs to remember the KISS rule. But how can you keep it simple when you’re hiring 300 math geeks a month?