Hi Markus, hi Graeme, thanks Markus for detailed email and willingness to help! It's great if you find an alternative testing option useful, use it consistently within specific module and all contributors to that module are happy with that! Speaking of myself.. I'm just looking at all the amount of tests that I wrote since 2003 and struggle to image what needs to happen to a) motivate me to learn 'something else', b) to get used to it, and c) re-write all tests in CCTBX and Phenix that I've accumulated over these years! Especially that I'm totally happy with the way things work currently. I'm sure "a)" and "b)" are 'trivial', but.. I simply have other things to do! And yes, "c" is less trivial.. Clearly there are better technologies! But see above. As many of us, I'm trying to balance between doing science, writing code and supporting users. This makes me think trice before diving into new adventures unless I absolutely must. I appreciate that learning "approx_equal" is something new to newcomers. But it is as new as "xray.structure", "flex" or "miller.array", and I imagine this is perhaps the smallest concern compared to the rest of the learning curve! All the best, Pavel