NWChemEx / Chemist

Library for describing chemical systems
https://nwchemex.github.io/Chemist/
Apache License 2.0
0 stars 0 forks source link

Decide on unit conventions #287

Open ryanmrichard opened 2 years ago

ryanmrichard commented 2 years ago

As I understand unit literals, each quantity is actually stored in some base unit. The base unit is privileged in that retrieving the value of the quantity, in that unit, does not require conversions.

Issue #286 focuses on the design of introducing units to Chemist and establishes the ability to choose the base unit as a requirement. Assuming this is the case, as part of the design we will need to actually choose the base units. This issue pertains to choosing the base units used throughout chemist. This issue is realistically a sub issue of #286, that has been split off to encapsulate the conversation.

Before getting too carried away it's worth noting that there are only so many linearly independent units. Conceivably it could be the case that the library we use (assuming we do indeed choose to go with an existing one) only allows us to set say 5 base units and then defines everything else in terms of those units for consistency. The consequence being that it may not be possible to have all quantities stored in the desired units and some conversions will occur. My suggestion is to settle on a library before tackling this issue.

That said, for sake of completeness, we note that the original proposal, #113, advocated for storing everything in atomic units. Since then @wadejong has advocated for storing mass in amu.

wadejong commented 2 years ago

amu we can handle through overload. As long as we are consistent on the internals.