I've been working on some code to parse the shapefile format, which specifies some fields in big-endian and some in little (I have no idea why), but the binary package has made it really easy to deal with.
Agreed that the web doesn't have any "magical properties", but it has the "Worse is Better" survival characteristic; or at least it did in its early days. I think the web's continued dominance depends on the extent to which it retains those characteristics.
I think the snake-oil nature of (much of) SEO stems from the fact that it's difficult or impossible to measure in an accurate, unambiguous way. Since A/B testing includes measurement by its very nature, I suspect it will be more resistant to snake oil.
Wouldn't simply excluding any http request whose user agent contains "googlebot" from the displayed count make it dramatically more accurate without needing a switch to JS-based analytics?
This is my November Startup Sprint[0] project, inspired by a StackExchange question[1] from this HN thread[2]. The idea is that the app creates a unique email address for you, and any emails of the form "Lunch with Bill Tuesday at 12:30" that you send to that address will be added to your Google Calendar. I'd love to hear what HN thinks of it. Much thanks to HN user riskish for his design help.
It appears to be down at the moment, but there's a site called Wrong Tomorrow (http://wrongtomorrow.com/) that seems to me like a good start in this direction. Here's the blog post announcing it:
http://www.physics.nyu.edu/sokal/lingua_franca_v4/lingua_fra... http://www.fudco.com/chip/deconstr.html