The article was a little wordy but in general I understood what the author was getting across. For data integrity, each lookup table should be a separate entity with a foreign keys coming from each table that references (uses) the code. That said, I have also used (or been made to use) the MUCK approach, and yes, in production with some success. Of course, bad entries eventually made their way into the tables, but it was not very often.

-Vic