from simple to complex to ( again) simple data model

[NathalieDelattre]

Speaking about the requirement to supply or not the INSPIRE/GML together with the alternative encodings.
If the complexity has been artificially implemented from original data sources , I don't see a real need to do so.
the reasons are:

1. risk of loosing the integrity of UIDs and lifecycle info:
   unique id and lifecycle info are usually maintained at the source level and migrating into complex data model might create artifical or automatically generated UIDs like for the property types. This might not be retrieved when transforming them again in the alternative simplified data model.
2. risk of artificial growing size of the dataset:
   Do we need to look at other solutions for reducing the size by for example, erasing the void information?

In my opinion,
We might consider the INSPIRE data specifications as a reference allowing different ways to pragmatically modelise the data according to use perspective but not necessarily to modelise it as such.

[thorsten-reitz]

@NathalieDelattre Do you currently see any unmet requirements given this issue?