Indeed. If you think about it, little/simple reduces the options for
variation - fewer dimensions of variation. Allowing scalable variation
by necessity adds complexity.
----
Enterprise systems and information exchange standards are typically non-trivial enough to need that variation.
The "too simple" data model has the same consequence as the "too simple" programming language - alot more code gets written.
If you take a system naturally suited to OO modelling and decide to write it in C instead of C++ because C is "simpler" it's soon becomes apparent that you will be writing alot more code.
People kid themselves (and others) that this doesn't apply in the data modelling scenario because they get hoodwinked by what seems to be an upfront economy that lets them drive straight into coding. The reality is it's a lot more coding.
