Well, I can see your point, but these are only really partial solutions to the problem. For me the issue is that these are relatively one off scripts, however they must be repeatable (months later I may need to go back and regenerate exactly the same data). Just as in OO we strongly bond our procedures with our data, I believe the same thing for data mungers applies.

If I went the naming route I would still end up with hundreds of <500byte files in my directory, with the possibility of the data being changed or the scripts going missing. When they are strongly bound I can assure myself that no-one else has changed them (or at least that they have indeed been changed), also that when I move them around they are an atomic unit, and that the only way they go missing is if I lose both the data and source.

Hmm, thinking about it, I suppose I could use alternative data streams to store the perl code, which would solve the binding issue, after all I am on a modern OS. (heh, its not often a MS user can say that...)

:-)

Yves / DeMerphq
--
This space for rent.