On 08/14/2011 05:48 AM, Joshua Megerman wrote:
> I have read your response from 6/28/11 to this patch, and I am very
> dissapointed in the hypocritical nature of your decisions.
>
> While I understand your reasoning in wanting to have a split between
> Power"top" and Power"tune", the latter currently doesn't exist, and as such we
> only have Powertop to work with. You argue that Powertop shouldn't give
> people the ability to script the changes and that it should be documented on a
> web page for them instead, however there are 3 problems with that under the
> current implementation.
>
> 1) When you run powertop, it gives you the ability to make changes to your
> system, however it doesn't tell you WHAT those changes actually are, or how to
> do them yourself. You argue that people shouldn't automatically be making
> changes without understanding them, yet you allow them to do so directly using
> Powertop by letting them toggle settings without telling them how to do it
> manually (and thus how to REVERSE them manually either).
I completely agree with you and that's basically why I wrote this patch.
At least showing the options in Tunables tab or generating them in HTML
output ("To turn this tunable on/off, execute: ...") would be great.
This is not part of my patch, but once you have proper commands stored
per tunable option, it's not hard to do it.
I still don't see how the script can break anything. If someone uses it,
he has root on that machine and probably knows, what he's doing. He also
probably already tried turning on/off Tunables in powertop UI without
knowing what it does.
> 2) There is currently NO webpage documenting the changes. If there was
> currently valid documentation, your arguments would carry much more weight.
> yet it has been almost 6 months since you first advocated doing so and I can't
> find any documentation of the power changes made by Powertop 2. As a result,
> people keep asking for ways to make the changes in scripts, regardless of the
> validity of your argument from a big-picture standpoint.
Personally I think even webpage is not good place to put informations
like that. Why do we have man-pages then, when they are already on webpages.
Another reason why it's better to keep the commands in source code
instead of website is that it motivates developers to change it once
they change particular tunable. With these informations on website,
*you* would have to update the website on every release separately and
keep differences between version (so somebody with powertop 2.X would be
able to see what particular tunable does in version 2.X even when 2.X+1
is stable).
> 3) If powertop is supposed to be a purely diagnostic tool, then you should
> never have given people the ability to change tunables in the first place. By
> doing so, you have already validated people desire for automatic tuning on
> boot, and any attempt to tell them NOT to do so falls flat on its face.
Yes, I think so far it's half-implemented. It allows people to change
something, but they can't see what changed and they can't reproduce the
change without powertop.
> Please either accept the patch to generate a script of all the changes, or
> post your fabled documentation and point everyone to it. But stop being
> stubborn with your userbase just because you don't like them using your tool
> to do something that you gave them the ability to do yourself...
>
> Josh
Regards,
Jan Kaluza
_______________________________________________
Power mailing list
[email protected]
https://bughost.org/mailman/listinfo/power
