Re: Is there any valid reason to add an idiotic script to /etc/init.d by an default Debian install that only cause a PITA?
On Tue, 14 Jun 2011 16:16:38 +0200, Ralf Mardorf wrote:
> On Tue, 2011-06-14 at 13:53 +0000, Camaleón wrote:
>> But that's a known issue, right? Or at least is not different from
>> other problems that any user which requires a specific/fine grained
>> setup or tweaks may experience.
> Yes and yes.
>> But then you are enforcing a cpu frequency policy that can cause
>> problems to the majority of the users... are you suggesting that?
> What kind of problems?
With todays clock speeds you may care about overheating. While you can
expect servers to be usually well cooled you may be more cautious when it
comes to desktops and notebooks. And having the latter always running to
their highest clock speed can lead you to unforeseen restarts.
>> > Yes, but not everybody knows about CPU frequency scaling and those
>> > issues, so you'll read about mysterious audible glitches for the
>> > audio signal all the time on audio mailing lists.
>> Err, if that's the only annoyance, have you ever considered to write a
>> page in the Debian wiki and pointing users to there? :-)
> No, perhaps a good idea, anyway, I suspect that users won't read
> billions of manuals, Wikies etc..
If a user is instructed to read a documment but he/she refuses to do it
(because of laziness or whatever excuse they give) then it's up to them.
Users have to read, being linux or windows users, we cannot expect all of
our problems being automatically solved without moving a finger... err, I
mean, an eye :-)
>> It's not about "load" but "frequency speed" and my micro reduces its
>> cycles from 2.8 Ghz. (normal run) to 2.0 Ghz (idle+ondemad) which I
>> prefer. Why wasting cpu cycles and increasing the heat of the whole
>> system if I don't really need all that power all the time?
> There's no big difference for the load and the load will cause the heat.
> I'm using the GNOME applet anyway. Before users run into issues
> 'performance' as default for everybody would be the better choice.
Why "better"? :-)
A mobile user may prefer to have a more conservative scaling frequency
and neither "ondemand" nor "performace" governor is "better" for him.
> Is there any CPU where CPU frequency scaling really makes a difference
> for load?
Define "load" in your context. Choosing a CPU frequency is a user issue,
what is fine with me it can be bad for you so the default setting has to
be the more fair and that's "ondemand" ("powersave" and "performance" are
the more agressive profiles).
>> Well, not at all. The moral of the story was that most of the linux
>> users like to know what are they doing, they like digging into their
>> system's guts and enjoy learning while windows users are more
>> accostumed to things work "magically" :-)
> And there are billions of users who tried to switch from Windows to
> Linux just regarding to ethical reasons or better software regarding to
> their needs and they all need to use Windows, because the Linux
> community don't care about people who only wish to use a tool like
Saying that is a bit unfair >:-)
You said "hey, let's make users to compile their own kernel" but reject a
default policy that can be easily modified? Who is the one that now
doesn't "care about people who only wish to use a tool like Linux"?