I love this paragraph from Doc Searls' post today:

Doc Searls Weblog · Context is King: "Phil Windley (disclosure: I’ve done work for Phil) gives a talk in which he provides a brief history of e-commerce. It goes, ‘1995: Invention of the cookie. The End.’ Thanks to the cookie, we have contexts — but only inside each company’s silo. We can’t provide our own contexts except to the degree that each company’s website allows it. And they’re all different. This too is a bug, not a feature. (Just like carrying around a pile of loyalty cards and key tabs is a bug. Hey, I know more about who and what I’m loyal to than any company does — and I’d like my own ways of expressing that.)"

He is frustrated that while in France google.com automatically defaults to google.fr (the French Google) despite any easy-to-access attempts to get to the US English version. Cookies are a fascinating topic both in terms of advertising and (especially now with the evolving "social" web) social sites. I wrote about a similar problem I saw in the way online advertisers were using cookies back in 2004 (wow, time flies):

Digital Moses Confidential: How Fresh is Your Cookie?: "It would seem that cookies would be the ideal and have enough inherit flexibility and durability to reign as the online direct response’s choice for tracking well into the future. The cookie quickly gained its pre-eminence in the online world because of its simplicity, its relative obscurity and its long-term reach. However, recent market forces are converging to form the perfect storm that threatens to make the simple cookie… well, crumble."

I was wrong about the gathering storm then, but I think I was right about the implications for sticking with the way cookies have evolved as the keycards (in some cases debit cards) of our online/social experience. It's a model that's bound to break eventually (hopefully) because as Doc points out later in his post, relying on cookies to inform corporate 3rd parties about our contexts (location, social, status, etc) is not optimal because we ourselves are made dependent variables. Instead, we should control our data and our context and be independent. That's what I'm trying to do with this whole harrelson.fm experiment and I hope it works. I'm having fun with it so far and it's been a blast to really have a reason to think hard about where my data flows and goes as it makes its way through all these tubes.