Dollar a day ???

A couple of weeks ago I was watching Hans Roslings' TED talk about the gapminder technology. A very interesting talk about global data; the work they are doing is awesome! After thinking about how they presented wealth distribution and poverty I have started to have my doubts. Hans showed an amazing dynamic graphic of financial distribution since 1970. In the graphic they included the dollar a day as a measure of poverty since 1970. As years passed everything in the graphic moved to the right showing a positive progress in the fight against poverty. One thing didn't move, that was the dollar a day line. The way I calculate it (based on Canadian inflation rates) a dollar in 1970 would be five dollars in 2003. Even though the presentation is impressive, I believe it is flawed. After some reading it would also seem the whole dollar a day calculation could be flawed...

How to not measure the poor.
Poor but pedicured.
How not to measure the poor - a reply.

Don't get me wrong, I think all the work in measuring global poverty is great. I'm just being critical of how progress is being claimed and how is is measured.