please leave a comment!
the way i saw it was that regardless of the rom, the phone will drain a lot or a little depending on how much or how little you use your phone. but thats not a test on the ROM itself, which is exactly what i was curious about, but in fact a test on how often you use your phone. So, although at one point i tried testing the battery usage as a function of how long i did certain everyday tasks (e.g. Facebook, texting, Angry birds, etc..), i noticed that this did not do an accurate job of testing the battery use by the OS. And honestly, whatever little differences i saw in data between different roms for these types of tests could be considered a margin of error for the test itself. So theres no value in that data. Over time, i realized that the only way to accurately test a ROMs battery usage was allowing the OS (and only the OS) to be running, but making sure that no other battery drains were active, like the display being on, the radios changing towers, etc etc. So thats why i tested the way i did. It was just a test for those people who already have their own ways of increasing phone life (extended batteries, juice defender, undervolting, etc) but had no way of choosing the most battery efficient OS, because there was no way of finding out which ROM is quantitatively the most efficient.
As for kernels, i agree it would have been a better test to include kernels, but in retrospect, that would have taken way longer because the only true way to test for it would be to test each kernel with one rom, then test the best kernel with every rom. but then honestly, who knows if that would even be a fair test because who knows if a dev makes a rom specifically (or only) for a certain kernel. so that wouldnt be fair to the ROM itself. If you wanna test for kernels too, then go for it, but i aimed to only test for the ROMs that are presented to us by the devs. i didnt wanna do any more modding to the phone because i dont think most of us really care to go as far as to flash different kernels in addition to the roms. some people just wanna flash a rom and be done. and this test was done for those types of people.
As for stock rom, i really wanted to know how much better custom roms were on battery life, in general, compared to stock. But by the time i realized how to conduct this test the way i wanted to, i had already unlocked, rooted and flashed a rom. So, at the time, i either didnt know or didnt want to do the work to go back to stock, and i figured it wouldnt matter anyway, because nobody who cared about custom roms would have a stock phone anyway lol.
But if someone out there still has a factory set galaxy nexus LTE/CDMA, with no mods whatsoever, i would encourage them to carry out this test to as strict a setting as possible and report back the results with screenshots. Thanks.
Edited by Shayes, 09 March 2012 - 03:55 PM.