It would be hard to give a true "apples to apples" comparison using a intel Igp as you are forced to use the foss video drivers in Linux.
Good point. How would one measure power consumption without the graphics driver, i.e. is it possible to "remove" the effects of the graphics driver from the measurement or measure with the graphics turned off? I would imagine that I could remotely log into ubuntu and force DPMS to turn the graphics card off, and then query the battery through the command line, but I have no idea how I would do the same in windows.
when reading the hader of the article it looks like a hoax because there are no major differences in power consumption but in hardware support.
The tests have been done very unprofessional without any os optimizations (e.g. simple kernel tuning, basic power settings) which would bring up better results.
Looking at timeline all tests show same time line for boot which is for me on every system I use 1:3 linux : windows.
My experience in power consumption on Ubuntu 11.04 is that except of missing hardware support for e.g. hybrid platform (intel/nvidia) the power consumption is almost the same without any big differences. In case of hybrid platform this varies in Windows 7 having 5 hours standby while Ubuntu 11.04 has only 3 hours because the power saving function of hybrid platform can not be used and nvidia graphics is always switched on. Strange how hardware vendors advertise supporting Ubuntu with official enhancements like Dell does but not publihing any official drivers to the public. If developement would go that way like was done with Ati cards with faster linux drivers the results would look more different.