On Sun, 17 Oct 2010 09:05:29 +0200 Andreas Gohr <andi@xxxxxxxxxxxxxx> wrote: > Not neccessarily. If more strings where removed than added, it might > be that the percent is >100 with still some missing strings... How about not counting strings which are not present in en? Seems like you are currently doing something like (count($en) - count($lang)) / count($en) where count(array_diff($en, $lang)) / count($en) would be more appropriate. Regards, Adrian