Errrr, I didn't.

It is a matter of scale. If the amount of data is small and the database design is simple then why not?
Least ways you can spare yourself the great overhead (and often cost) of creating a SQL database and having to support it (I've seen so many sites where whole teams of DBA's were needed to support relatively simple databases)

For under 5000 records I'm not convinced that opening a database connection squirting your SQL across the TCP socket and getting a reply back is any quicker than scanning the file. Especially if your data set has no suitable keys. Having said all that I'm quite a fan of AnyDBM_File as well

Maybe I'm just an old disillusioned DBA :-)

"We are all prompted by the same motives, all deceived by the same fallacies, all animated by hope, obstructed by danger, entangled by desire, and seduced by pleasure." - Samuel Johnson