Hacker News Reader

You should adjust your belief to fit the available data, don't adjust the data to fit your belief.

I would normally agree, but in my opinion, surveys fall into the "Lies, damned lies, and statistics" category. They hold the same value as apache-bench numbers that are posted every time a new webstack appears.

There's good data and bad data.

I heard a few people at the next table talking at a restaurant about the new Thunderbolt. Try said it's "a droid" and it was a solid phone because Microsoft built it. I won't comment on if the percentage is valid or not, but a LOT of normal people out there are confused about the current mobile landscape.