stdoutSileSiesta.read_charge returns dq and e? about sisl HOT 15 CLOSED

I always use the dq column, so not changing it to de would be more convenient to me 😅

However, I'm not sure it makes sense to use q, since it will always be negative.

from sisl.

Well, I think we need to discern this.

Currently, my understanding of the output is actually that dq is de, i.e. it is in units of electrons. Hence, for correctness one should actually return -dq. So we are now left with a double problem ;)

from sisl.

Currently, my understanding of the output is actually that dq is de, i.e. it is in units of electrons.

Are you sure about this? The first few lines of the parsed Hirshfeld charges from one of my calculations is,

 	dq              e 	          S 	          Sx 	          Sy 	   Sz
atom 						
0 	0.46978 	13.53022 	2.96865 	2.96865 	-0.0 	-0.00000
1 	0.46978 	13.53022 	2.96865 	2.96865 	-0.0 	-0.00000
2 	-0.15659 	7.15659 	0.01046 	0.01046 	-0.0 	0.00007

From this it seems like dq assigns a negative charge to excess electrons, so then it would actually be -de right?

from sisl.

Yes, for Hirshfeld and Voronoi, these things got cleaned up. But I think for TS and Mulliken, this is not the case. I might be forgetting here.. But it would be nice to not get into problems here :)

from sisl.

If we had to decide to choose either dq + q or de + e then I would prefer to talk about occupations rather than charge, such that excess electrons would be positive and deficit of electrons would be negative (so de + e). This does go against @pfebrer 's preferences though. And if we're changing things I would actually prefer to use n and dn to talk about occupations/populations because I feel e could also be confused with the electron charge constant.

from sisl.

What about having both for some time?

from sisl.

Would also be fine for me.

from sisl.

If we had to decide to choose either dq + q or de + e then I would prefer to talk about occupations rather than charge, such that excess electrons would be positive and deficit of electrons would be negative (so de + e). This does go against @pfebrer 's preferences though. And if we're changing things I would actually prefer to use n and dn to talk about occupations/populations because I feel e could also be confused with the electron charge constant.

Isn't it electrons that we are talking about here? Occupancy could be occupancy of anything 😅 I think "e" is clearer than "n".

What about having both for some time?

I think this would be a bad idea unless one of them is lazily computed. But for that we would need to subclass DataFrame. I think it is better to just choose one and stick with it, whichever it is. There are many parts of sisl where the API is not stable yet for good reason, but I think this one is not a good reason to purposely break future compatibility 😅

from sisl.

Isn't it electrons that we are talking about here? Occupancy could be occupancy of anything 😅 I think "e" is clearer than "n".

Fair point. Still I feel like using e could be prone to confusion. I think the least confusing single letter header would be q, but then you would always have negative charges of atoms.

from sisl.

fair points all. :)

from sisl.

Ok, so we stick with dq meaning the charge on atoms.
A deficit of electrons is a positive dq.
A surplus of electrons is a negative dq.

Lets stick with that. :)

from sisl.

And what about e? I mean do we also keep it?

from sisl.

Yes, lets keep it. :)

from sisl.

Ok, so this issue is the identity operator 😄

from sisl.

ha :) true :)

from sisl.