YAMAMOTO Takashi wrote:
>> On Sat, Nov 18, 2006 at 07:49:23PM +0900, YAMAMOTO Takashi wrote:
>>>> We should have a documentation for sysctl MIBs that is inluded in man.tgz
>>>> and uses MIB names, not the constants. (in a new sysctl.7 ?)
>>> i think it's a good idea.  do you want to do it? :)
>> What exactly would it mean? My idea was to completely remove the
>> symbolic constants (CTL_* etc.) and the header files where they can be
>> found from sysctl.3, in the descriptions replace the constants by MIB
>> names (e.g. CTL_HW -> hw, HW_MACHINE -> machine) and move the
>> descriptions to sysctl.7, leaving only the documentation of the
>> sysctl*() functions and some examples in sysctl.3 . Do you agree with
>> that, or should the symbolic constants remain documented somewhere?

I think the consensus was that we should for now keep the constants in
the *headers*; I don't see any point in keeping them documented.

> i agree that the following
> 	
> 	HW_MACHINE
> 		The machine class.
> 
> should be
> 
> 	machine
> 		The machine class.

that works.

> i agree it's better to move them from sysctl.3 to a dedicated page, too.

yes. when adding a new sysctl variable, there should be only one
man-page to update, sysctl.7.

-e.

-- 
Elad Efrat