The Separation Model – Part 14 – Data Classification
How is it you know a student from a vendor and a vendor from a teacher? Well, it's all context right? You visit the school to see the teacher and you visit the industrial complex to see the vendor.
As the saying goes, "If life was only that simple." Of course you can find a student at a school OR an industrial complex - the former is every day life and the later is a field trip - Yeah!
Since we can't count on context, we must use something else to identify what's what and who's who. Let's call that thing Data Classification.
Classifying things falls under that wonderful term named taxonomy. It's where you stuff anything you have with the terms which most clearly identifies it
All jokes aside, this isn't a concept which comes easily to the disorganized database developer. There are many ways to classify your data and knowing which one to use, and when, can be the confusing part. Staunch friends of E.F. Codd will tell you to use as many tables as possible and the finish-it-quick demands of an off schedule project will scream "just add another field".
Without having bias towards either of the extremes, I hope this video will provide a good deal of insight into how to approach your own Data Classification needs.
Comments
One Table Many Value Lists, No Relationships
This is how I like to do multiple value lists in one table.
http://dl.dropbox.com/u/2273592/value%20lists.fp7
global SETCLASSIFICATION
Hello,
Edited: my original question was regarding where the global SETCLASSIFICATION gets set. For some reason it wasn't getting a value when I ran the script as adapted to my application. But lo & behold, after restarting FMP 12, it is getting a value (although still a mystery to me where it does get assigned a value).
Thanks!
Tom
Mutually exclusive "classifications"
Hi Matt, and thanks for this very helpful video. Do you have any tips as to how to handle mutually exclusive classifications? I'd like to use this setup for all attributes of People, including gender, ethnicity, etc. Setting a classification will mean unsetting others (e.g. if we set somebody's gender to be Male, we need to check that they weren't previously tagged as Female), I guess via toggling the alternative classifications within the set. But with something like ethnicity, which has many possible values, that could be quite a task. Maybe I'm missing something that would make it easier?
Regards,
Dave.