Open cortner opened 3 months ago
it seems reviewing this terminology will be part of AB0.4 so shouuld wait for this.
Now that AB0.4 is ready, I suggest that we replace get_id
with species
, or at a minimum to use species
as a fall-back for get_id
.
I am not a big fan of using integers as categorical variables since it is then very easy to get mixed up with indices or even do arithmetic on them when one really shouldn't. Big course of bugs ...
In general allowing user to define what to use would be good. But for a default species
is better than atomic_number
(current default). We can discuss tomorrow how to do this.
The way I see it, if I want to use an integer in my package, or a different "particle id", I can always internally do a conversion e.g. atomic_number(species)
.
My concern is that get_id
seems fundamentally the same function as species
. In other words, having both looks like increased flexibility but it really is just defining the same function twice with different names.
Maybe think about a concrete use-case where my perspective doesn't apply and explain it later in our meeting.
AS DISCUSSED: remove get_id
and replace with species
this is too generic a name and should be replaced, e.g. with
get_particle_id
or similar.