Hans Moritz Günther

Sherpa has a number binning functions, e.g. by width, bin number, or minimum number of counts per bin, see https://cxc.cfa.harvard.edu/sherpa/threads/pha_regroup/ But the "optimal binning " from the Kastra paper is...

I'm in favor of removing C++ code when can (i.e. where the C doe is not needed for speed). My main motivation is convenience for developers. I can read `a...

CIAO uses multiple WCS at the same time. That's part of the WCS specs and can be dealt with in reading it in - you have to ensure that you...

I looked at some of them (in particular #2324) and if I get to finish this, I promise it will mess up all of them! (I have more commits locally...

@DougBurke : It would be great if you could take a quick look at my changes in `data.py` (both in sherpa/data and sherpa/astro/data) and `regard.py`. The tests still fail, but...

Oh, and I've also used Python 3.11 typing syntax, that makes or 3.10 tests fail. Again, easy to fix if we want to move forward with this approach in general.

I remove the "DRAFT", as this no passes locally. It's still going to fail on CI, because we still run CI on 3.10. We already has #2356 open to change...

I'm looking at this. I thought it would a good one to take before 4.18, but I don't want to do any more changes for 4.18. Yet, sine I started...

Another source of complication: `load_table_model` has parameters for `ncol` and `dstype`, but if reading with the arguments that are set fails, it will always try the `ncol=1, dstype=Data1D` version (through...

But then, we are creating a `TableModel` and table models are always 1D and not integrated. They only take two arrays on load (x and y), not `x_lo`, `x_hi`, or...