Remove extrapolation

[uekerman]

Using an extrapolation for implicit coupling has no guaranteed benefit. Sometimes it gives better results, sometimes not. I would not advise beginners to use it. That's why I suggest to remove it from all tutorials.

Please note This might break the system tests as the results now might differ.

@reviewers: What do you think? How big is the pain to update the reference results for all system tests?

[MakisH]

This was kind of intentionally left here, so that one can easily see what the differences between explicit and implicit are (and aitken vs IQN). But I see that this is inconsistent and dead code, which could eventually lead to trouble.

[MakisH]

I really like the idea of making the tutorials less scary.

Updating the system tests reference results would be easy, but I would also like @Eder-K's opinion on that.

The most tedious step would be to manually check all the results that changed if they still are physical and close to the (potentially validated) previous results. But it's ok if the tutorials are not completely accurate: the purpose is to learn how to use preCICE.

[Eder-K]

I agree with @MakisH in that it should be easy but a little tedious. I will check how the results are altered using this version.

[BenjaminRodenberg]

In case our systemtests now go crazy: this is good, because we want to get rid of the extrapolation and then we will know where action is needed.