"h" should be the abbreviation of hours

[eevleevs]

"hr" is used as an abbreviation for hours, but this is not according to standard (see https://www.bipm.org/en/CGPM/db/9/7/), and much less common than "h" worldwide.

Here "h" is used for the Planck constant, which I think has much less use than the hours symbol. Both "h" and "hr" should be used for hours, to avoid misunderstanding. Planck should be called something else.

[jondoesntgit]

Changing this now could break backwards compatibility. An alternate solution would be to define a different registry that is according to the CGPM standard, which would define Planck as something other than h, and define hours as h and hr.

[hgrecco]

I think we should put constants into a namespace in the long run. But now I agree with @jondoesntgit here.

[jondoesntgit]

How would namespaces mesh with Quanity? Say that I wrote a quantity:

Q_(1, 'h'), would the program assume that it's hour, instead of Planck constant? I can see this leading to unintuitive/unexpected behavior for new users.

[eevleevs]

Not sure I get this. I think it's much more likely for a new user to expect h for hours rather than Planck.

[eevleevs]

By the way I vote for having constants in another registry. They may work the same, but they are a different thing in concept.

[cpascual]

Q_(1, 'h'), would the program assume that it's hour, instead of Planck constant?

If we ignore the backwards-compatibility issues, I think that the actual intuitive behaviour is to read Q_(1, 'h') as "1 hour".

Also, I think that separating constants from units is a good idea. IMHO constants should be Quantities, not units.

[jondoesntgit]

Agreed that it seems uninutivite to define quantities based on units (1.72 Planck constants). I would, however, like to avoid a situation where Pint tacitly interprets u.h as Planck constant and Q_(1, ‘h’) as hour. The symbol should remain constant.

I think Scipy’s units package puts all its constants in a namespace.

[bilderbuchi]

IMHO constants should be Quantities, not units.

Agreed, preferably this or at least separate namespaces for constants would help here. I was also just tripped up by the unintuitive parsing of "1h".

[akaihola]

Is there any simple work-around to create a custom unit registry in which h stands for hour?

[larstiq]

Grepping for planck in the codebase finds #82 where the alias h for hour was removed (in the opinion of the thread participants, that would be the wrong way around).

That happened in commit ae189a1 where pint/default_en.txt dropped the = h. That file is loaded in https://github.com/hgrecco/pint/blob/master/pint/registry.py#L268 if BaseRegistry(filename="") so it seems at least at that level one should be able to override with a custom registry. I did not check how the inheriting classes are handling this.

[jules-ch]

@hgrecco

IMO we should make a breaking change for this case, since it is really disturbing for new users & can potentially introduce bugs when using pint.

I like the idea of a dedicated namespace, but it's difficult to implement since we rely only on strings in the file definition.

Following BIPM standard & ISO/IEC 80000 series Quantities and units & IUPAC Green Book, roman font is used for units & italic symbols for quantities to avoid any ambiguity.
This could help us to differentiate between constants, that are quantities.

[jules-ch]

At least for the hour vs planck constant case, we can use unicode for planck constant.

[Flix6x]

Is there any simple work-around to create a custom unit registry in which h stands for hour?

@akaihola Not sure if you'd qualify this as "simple", but here's a hacky workaround:

import importlib.resources as pkg_resources
import pint

# Edit constants template to stop using h to represent planck_constant
constants_template = pkg_resources.read_text(pint, 'constants_en.txt').replace("= h  ", "     ").replace(" h ", " planck_constant ")

# Edit units template to use h to represent hour instead of planck_constant
units_template = pkg_resources.read_text(pint, 'default_en.txt').replace("@import constants_en.txt", "").replace(" h ", " planck_constant ").replace("hour = 60 * minute = hr", "hour = 60 * minute = h = hr")

# Join templates as iterable object
full_template = constants_template.split("\n") + units_template.split("\n")

# Set up UnitRegistry with abbreviated scientific format
u = pint.UnitRegistry(full_template)
u.default_format = "~P"  # short pretty

Check:

>>> u.Quantity("h").dimensionality
<<< [time]
>>> u.Quantity("planck_constant").dimensionality
<<< [length] ** 2 * [mass] / [time]

[hgrecco]

@jules-ch
Checkout my comment here:
#1078 (comment)

What do you think of this approach for constants?