A couple of things power users enjoy in OSX:

- Services
- The automator
- AppleScript
- Using these to design workflows involving different applications

If you are creative and willing to learn a bit, you are getting pretty damn enabled in OSX by these extremely well thought through tool sets.

And I actually even like the Finder. Miller Columns mode together with a persistent info window, and of course activated services linked to my custimized scripts, is working fantastic.

As someone who knows a little bit of mathematical modeling and statistics, I have to point out that they did not predict the percentages, since they already were known! The correct term would be retrodict.

I upgraded to 2.0.0 on my old PowerPC G4 iMac, which I like to use as a movie player "for the design". Warning for that! No sound, red stripes all over the frame... The upside is that it's really easy to downgrade, just move the old app bundle back from the trash can to the applications folder.

Wrong... parent is far to quick to judge. Indeed, the sample size is to small to calculate reliable quantitative estimates of the differences. Nevertheless, one may use non-parametric tests to check for whether there is a *qualitative* difference between the groups, and this is really what's claimed here.