I read a lot of files of very raw textual data and if the parsed data tastes good, it gets stored in a DB for analysis, if not, kicked out as error.
Its going to be fun watching people who do used the POSIX functions try alternatives. /[A-Za-z]/, I'd like you to meet umlat-u. The POSIX depended on the locale of the installed software which, assuming that's a feature and not a bug (usually its a bug) it'll be interesting to watch people reimplement that "functionality".
This is a fun 13 year old place to start WRT "simple" data verification, not because it exhaustively covers everything, but its watching a noob start from the beginning and hit about ten different roadblocks. There are plenty more of course.