Repository navigation
Conversation
|
I have been doing a lot of Is there a backward-compatibility option so the above won't break? Or will that introduce circular import between new For myself, this backward compatibility is not critical. But other developers who build on |
|
@eteq - I think you may have misunderstood my comment: I had indeed been considering whether and in the few times, I do use it directly, I do So, overall, I'd say keep |
|
👍 to all of @mhvk's comment. I think it's mainly a matter of increasing prominence of quantity within units. Quantity isn't really anything separate from units, it's just a (value, unit) container. |
|
Ah, I didn't see #1591 until you mentioned it here @mhvk. So this is an alternative to that. The advantage of a separate subpackage is that it then provides a possible place for special quantity subclasses (e.g. |
|
whoops, yeah, sorry - just edited it now. They're only one digit off, so it's easy to get confused :) |
|
I also think we should not separate them - for one, most users don't even use |
|
I actually use |
|
At 2013-10-18 hangout, it was agreed we would follow the approach in #1591 instead of this. Closing. |
This was @astrofrog and @mhvk's comments in #1577.
Quantityhas grown into an extremely convinient and powerful tool for science. But right now it lives inunits, which both leaves it rather hidden in the documentation and makes it more difficult to work with. So this is a suggestion to move quantity into its own separate sub-package.Note that if we are going to do this, we should do it now (v0.3), because
Quantityis useful enough that a fair number of users already have started with it between 0.3 and the next version.This PR is not done as there are various parts of astropy that need to be adjusted for the new location (and docs moved), but it gives a flavor for this idea. It shouldn't be too hard to finish it up quickly if we do decide to do this.
cc @mhvk @astrofrog @pllim @adrn @mdboom