Liam Huber

> I assumed that this transformation has to be done by creating a prism and corresponding matrices based on the cell of the current step instead of the initial cell,...

> Shouldn't R be R(n)? Perhaps I misunderstood -- for interactive jobs, where the structure is being re-set, then I agree that the rotation matrix is step-dependent. Then I think...

> It is highly desirable for complex workflows that the POSCARs generated have the option to have their atomic ordering preserved, I'm not convinced that the VASP files themselves need...

> However, I note that issues such as https://github.com/pyiron/pyiron_atomistics/pull/700 https://github.com/pyiron/pyiron_atomistics/issues/718 make this somewhat extremely dangerous, especially if errors such as species assignment are still running around in the code, which...

After digging into it over in #492, I just want to put the answer here so it's linked together with the issue for posterity. > It is highly desirable for...

Renaming it is reasonable to me, but is there a reason we want the duplicating behaviour even for internal use?

I haven't gotten to checking but one of you might know offhand -- is the status also stored in hdf? If so then I guess the snippet needs to update...

> It is stored, but not sure where, might be just `job['status']` even. We probably want to update this in the stored version too then. IMO it would also be...

Yeah, if it's fully and clearly defined there then IMO it's enough to say "any adapter must match this spec: . Although that is contingent on pysqa spec having a...

> As it is an impact change and I do not know if it works with the GenericMaster and so on I would prefer to have a warning to make...