if you want sh to put a # on the end of PS1 if its running as root,
why not modify sh itself? it'd be a lot faster than running id!! 
even then thats a system call and lots of extra junk not directly related to
the job of being a shell.. 

On Sun, 11 Apr 1999, Charles M. Hannum wrote:

> 
> >     ENV=$HOME/.shrc; export ENV
> ...
> >     PS1="$(hostname -s)"
> >     if [ -x /usr/bin/id -a $(/usr/bin/id -u) -eq 0 ]; then
> >             PS1="$PS1 # "
> >     else
> >             PS1="$PS1 $ "
> >     fi
> 
> Ignoring the obvious errors, there are two problem with this:
> 
> 1) It causes 2 other programs to be started up *every* time sh(1) is
>    run.  Consider the effect of this on a `make build' on a VAX.  The
>    performance loss can be measured in hours, maybe even days.
> 
> 2) There is absolutely no reason to do something that complicated.
>    Just save the host name in a variable and concatenate the default
>    PS1 (in the same way as my code does) in `.shrc'.
> 
> Why make the problem harder than it actually is??
> 
>