Benjamin Uekermann

Why not creating a mapping per patch? I think this would be simpler overall. The only prerequisite would be that we need to know how many patches there are at...

Having this information in the preCICE config is a safeguard and, therefore, redundant. It is consistent with how we treat meshes and data fields already. Basically, we can use it...

I agree on all concerns and the proposed solution. Thanks for opening this :+1: The most common problem is not even mentioned yet: Often, coupled simulation get manually killed by...

Sounds good in principle. > Export everything by default Should this include all iterations? This could be too much. Does this mean to remove the option `every-n-time-windows`?

Got it. I am sure this would be a nice and simple addition, but it does not solve this issue, correct? (not saying we need to solve the issue now)

There is a large overlap between this issue and #1655. I am replying here for both. Let's simplify our lives: - Only export data at the end of a time...

> An interesting note here is that the support for mixed time-window sizes https://github.com/precice/precice/issues/1713 makes the time window naming convention tricky to understand. True. But also a rare case IMO....

We should also remove the deprecated ones: ``` REGULAR_PRIOR REGULAR_POST ON_EXCHANGE_PRIOR ON_EXCHANGE_POST ``` They still exist on config level.

Good idea 👍 Within preECO, we could auto-generate the picture together with the Readme.

"Configuration" section (like you have it now) sounds good to me. > once we agree on what we want to have. What else do we still need to define?