Can anyone explain why the idea is raised so often when wakelocks are discussed? [LWN.net]

Can anyone explain why the idea is raised so often when wakelocks are discussed?

Posted Jun 3, 2010 10:18 UTC (Thu) by khim (subscriber, #9252)
Parent article: Quotes of the week

I don't understand why kernel developers so often offer "just use correctly written applications" as argument when wakelocks are discussed? Everywhere else kernel uses different policy: "support broken applications as long as it does not hurt the unbroken ones... much". The fact of the life: 90% applications are crap... and 9% are just plain broken and shouldn't work. The OS saves them where it can so the whole thing works somehow - and that means users will be using said OS. If the OS will only support 1% of "good applications" it'll be dead sooner or later (see Netware: it was much better then Windows, *nix or Linux - if you only used correctly written applications). Why the power management is so different?

to post comments

Can anyone explain why the idea is raised so often when wakelocks are discussed?

Posted Jun 10, 2010 22:40 UTC (Thu) by Wol (subscriber, #4433) [Link]

Because power management is critical to using your computer...

If you have a nasty app that won't sleep, your battery is flat in 2 hours when you thought it was going to last all day in idle (do some work on the morning commute, close/suspend your laptop for the day, open it at the start of the commute home only to find it's dead ...).

But the other way round (as I've been bitten, by linpus/Acer One) is that you do a network mv, then the laptop suspends and shuts down the network. Hang on a sec - I've just left my computer to get on with some work, and it shuts down! What's worse, because it was an mv, the shutdown unfortunately broke it with the result that a load of files disappeared in transit :-( !!!

Cheers,
Wol