On Thu, 2007-08-16 at 17:17 +0100, ext Andrew Flegg wrote: > On 8/16/07, Igor Stoppa <igor.stoppa at nokia.com> wrote: > > On Thu, 2007-08-16 at 16:51 +0100, ext Andrew Flegg wrote: > > > > > Given the importance of this to battery life, it would be cool if apps > > > were sent SIGSUSP when they were minimized or some the user stopped > > > interacting with them. If an application (e.g. media player) wanted to > > > override this behaviour, they could do so in their .desktop file, > > > .service file or osso_initialize() call. > > > > No. > > > > Apps that do useless stuff are buggy and the bugs must be fixed. > > Battery life eating is a particularly tricky bug to nail down to a > single application (or home applet). That's why users shouldn't try a bunch of new apps in one go. Maybe some way to easily share information about battery life would be useful on maemo.org ... Quim: what do you think about extending the rating of applications so that they could cover several aspects, including but not only battery usage? > > You are proposing a shortcut that is encouraging crappy code to be > > written, since the system will always take care of saying: "psst, > > pretend to be a properly written piece of code". > > Yes, I am. On the basis that this is a) easier and b) more likely than > getting end-user tools to properly diagnose which applications aren't > well behaved. That's not true. Proper tools don't need that. As Eero has already described, it's possible to use existing stats from the system. Powertop is not using such hacks and it works. Users have started complaining with dfevelopers and developers themselves have taken powerto in use. Eventually imo we have to come up with an evaluation suite that developers in first place and users later can easily refer to. > > Would you be happy if your foregrounded app would run and drain the > > battery while it's doing nothing and waiting for you to press a button? > > No, which is why I uninstalled Maemochron PDQ. And hopefully you have provided feedback to the developer. Developers should also be encouraged to use thir code on a freshly flashed image and verify that it doesn't alter battery life in unexpected ways. > > Have you ever wondered how come your typical GHz PC can have performance > > sometimes comparable with an internet tablet? > > No, I haven't wondered. I'm a software engineer: I understand the > issues, I just live in the real world of half-arsed Maemo ports and > believe it should be as easy as possible to do these properly. Notice that most ofthese bugs affect power consumption on a pc in the same way, so it's not really a maemo issue. "Ports" could be easily debugged in native x86 environment. -- Cheers, Igor Igor Stoppa <igor.stoppa at nokia.com> (Nokia Multimedia - CP - OSSO / Helsinki, Finland)
