icon cache scriplet guideline update

[Bill Nottingham]

Rex Dieter (rdieter at math.unl.edu) said: 
> >But this doesn't 'solve' anything, or 'simiplify' anything - it just
> >moves a touch into the shell script, and renames the command, and makes
> >things slower.
> 
> I'd bet the performance difference is not even measurable, so I don't 
> think it's worth arguing that.

time to do touch && gtk-update-icon-cache, with cache warm, average,
five runs: .166

time to do xdg-icon-resource..., with cache warm, average, five runs:
.270

62% slower. It'd be even slower if I had KDE installed, from a quick
glance over the script (or if I was running on a slower machine.) After
all, it's a shell script that runs sed on the path to walk it itself;
that *can't* be efficient.

> >So, essentially, it boils down to an objection to the tool being named
> >'gtk2-...'?
> 
> Or, it seems to be, a reluctance to use or bias against 
> http://portland.freedesktop.org

I'm against silly standards that make the system less efficient for
no actual gain. 

Bill