Hans Moritz Günther

On the plus side, I use Tablemodel to calibrate the gain in Chandra/ACIS and it works great for how I use it, e.g. https://github.com/hamogu/dPHA/blob/master/CasA.ipynb I'm not discounting all the problems...

That usage confused me initially. It would help to split into two models: `TableModelInterpolatedOnX` and `TableModelExactBins` (names open to bike-shedding). The old `TableModel` could remain in place as a wrapper...

Some development has happened since this issues was opened, and I learned some more from digging through code, but it's all in different places, so I want to just key...

Sherpa is definitely fully numpy 2.0 compatible and fully Python 3.13 compatible since the 4.18.0 release in October 2025. The code has been running on numpy 2.0 for a year...

I just checked "conda create -n test_conda_sherpa -c https://cxc.cfa.harvard.edu/conda/sherpa -c conda-forge sherpa", which suggests to give me numpy 1.26 and python 3.12. I suspect that DS/CM set it up that...