View Single Post
Old 09-30-2006, 05:12 AM   #8
ali doesn't litterali doesn't litter
ali's Avatar
Posts: 302
Karma: 116
Join Date: May 2006
Device: Iliad, dude!
Originally Posted by rlauzon
When iRex says "we are done", then I'll compare what I have to what they originally promised. Then, and only then, will I have the right to complain that something doesn't work as promised.
I guess that's the problem. iRex will never say "we are done" - why should they? It's much cooler to talk about "constant improvement" than to say you're done and provoke complaints of those who showed patience.

The other thing is that "constant improvement" will not suffice for battery life, IMO. Being something like a programmer myself, and having programming students, I know a little bit how these things work. Essentially, you can spend some time and implement the main powersaving method (turning off things between updates), that will give you a big boost in battery life. But afterwards, everything you can do is minor improvements, tweak a parameter here and there, optimize some code somewhere and so on. But each of these minor steps is difficult, and gives just small battery life improvements.

And this is what happened - in 2.5 and 2.6, they got the big boost from implementing PM, resulting in a factor of over 2. And now there will be a few small steps ("5-10%"), barely noticable, and then it will stop there. So "10 hours" is the final answer, unless they throw away their software design and begin basically from scratch, with something better - which will likely not happen.

It's a stone old lesson in software design (resp. anything design) - if the basic structure is bad, you can't go too far. And the Iliads software design is made such that you can get first results really quickly, but then get stuck. It's a power management nightmare.
ali is offline   Reply With Quote