On Tue, 13 Dec 2005, Curt Sampson wrote:
>> When you've written all the code, it's also much easier to say "screw it" 
>> and leave things to figure out for users w/o doing any documentation.
>
> I'm not saying that that's a good thing. But I think that it's a better
> state than not having the feature at all. I've seen stuff without docs
> and written docs for it where I wouldn't have done anything at all if
> the feature hadn't existed.

You mean like the vm.idlezero sysctl which apparently is in NetBSD for 
ages but was never mentioned in sysctl(3), sysctl(8) nor src/doc/CHANGES?


> Basically, refusing code due to lack of user documentation is saying
> having half a loaf is worse than having none. I'm not sure if that's
> what you were aiming at or not, but it felt a little like it might be.

The comment wasn't on the patch, but on the lack of documentation.
Only.


  - Hubert