>It's not, really, because it - or any mechanical reformatting, for that
>matter - destroys violations of the rules that improve readbility.

totally!

>And yes, such things exist.  Consider as a simple example
>
>    else if (!strcmp(*av,"-debug"))
>     { WANTARG();
>            if (!strcasecmp(av[skip],"none"))  plevel = P_NONE;
>       else if (!strcasecmp(av[skip],"warn"))  plevel = P_WARN;
>       else if (!strcasecmp(av[skip],"note"))  plevel = P_NOTE;

actually...i'd make that

       if      (!strcasecmp(av[skip],"none"))  plevel = P_NONE;
       else if (!strcasecmp(av[skip],"warn"))  plevel = P_WARN;
       else if (!strcasecmp(av[skip],"note"))  plevel = P_NOTE;

and then my editor would be much happier.  it also looks better to me.
each line begins in the same place.  i just moved the whitespace.

>versus what you get if you run that inner if construct through anything
>mechanical.  At the very least it will destroy the alignment of the
>then-clauses (which in this case is perhaps not a big deal, but I have
>written cases where making everything line up makes a big difference; I
>just couldn't find one in thirty seconds or so).  It'll also pull the
>first if over to the left, thereby obscuring the parallel structure
>even more.

it could be worse.  you could insist on

	if (!strcasecmp(av[skip],"none"))
		plevel = P_NONE;
	else
		if (!strcasecmp(av[skip],"warn"))
			plevel = P_WARN;
		else
			if (!strcasecmp(av[skip],"note"))
				plevel = P_NOTE;

so that the if (clause) and the else have nothing else on their lines.
theyn you're screwed.  indenting by 8 then reaches critical mass very
quickly.  this is certainly a quick rudimentary style, but very
defendable based on the simplicity of the rules.

-- 
|-----< "CODE WARRIOR" >-----|
codewarrior@daemon.org             * "ah!  i see you have the internet
twofsonet@graffiti.com (Andrew Brown)                that goes *ping*!"
andrew@crossbar.com       * "information is power -- share the wealth."