Remove REFCASE keyword about ert

[markusdregi]

@joakim-hove should correct me if I'm wrong, but here is my take on it:

There is a concept of time in libres, which is currently very closely connected to the time unit in Eclipse. Furthermore, as you know, we have the concept of observations and simulated data that is both mapped onto your time axis above.

The purpose of your reference case is to aid you with the above concepts, giving you a set of points in time, listing the keywords describing the simulated data you could expect to find in a forward model, as well as supporting extraction of observed data directly from your reference case.

Some of the problems users experience with the refcase is due to the underlying concepts and not the refcase itself. And while the users think removing the refcase would solve all their problems, they would currently just be on even shakier ground. It might very well be that the refcase should not be a first class citizen in an ERT configuration as it is today, but it could equally well be that it should be the basis for a tool that helps you set up your initial observations and similar. Either way, we should revise the foundations first, before me make decisions regarding the tooling on top ;)

[markusdregi]

Needs #193

[markusdregi]

The refcase enables early checking of observation keys and so forth. Could use OPM-common for this validation instead directly on the .DATA-files. All cases would probably not be supported..

[andreabrambilla]

I don't really know much about these tasks, but I was wondering, why would we need to rework how time is treated (#193) before addressing this?

[markusdregi]

Refcase will not be removed in ert2 for backwards compatibility. Will never be introduced in ert3. We don't need an issue to remember that.