Should `$hu-opacity-types` and `$hu-scale-types` use percentages instead of scales?

[stowball]

Currently, $hu-opacity-types and $hu-scale-types use scale values like 50: 0.5, but should they use percentages like 50%: 0.5 like width etc? Would that be clearer?

[joeldbirch]

I think I prefer the unitless scales, which in this case (and other cases) map to something that does not use % in CSS. Width values are different. Not only do they actually use % for percentages, but they can also take other units, making the use of % meaningful.

[stowball]

Width values are different. Not only do they actually use % for percentages, but they can also take other units, making the use of % meaningful.

🤔 That's an interesting point.

However, as a counter-point, do the unitless scales more represent something that isn't sequential, or doesn't have a near 1:1 mapping?

When I think of opacity, I always think in percentages ("can you make that 50% opacity?"), and often get annoyed that CSS represents it in 0 - 1.

The same is true for scaling. "Can you make that 50% smaller?" instead of 0.5.

[joeldbirch]

I’m not yet convinced on this either way, but to add another thing: would it be right to punish people for “correctly” remembering that opacity/scale values are not percentage-based?

Also, would an entirely accurate opacity:0.5 not be a valid and possibly stronger option?

[stowball]

would an entirely accurate opacity:0.5 not be a valid and possibly stronger option?

Definitely more valid, but I don't like it 🤣

[joeldbirch]

Fair enough. But “going rogue” and making up stuff runs the risk of straying from what makes Hucssley so refreshingly predictable.

[stowball]

making up stuff runs the risk of straying from what makes Hucssley so refreshingly predictable.

😭

Perhaps these types of values are more likely to be customised to suit your design system, so using a scale may help to reiterate that?

Just looking at Material Design, and it's totally plausible that you would change the scales to only the allowed values in your system

[joeldbirch]

Good point, that's a valid use of a scale. It may get a bit confusing that the current scale looks like "it's just a regular opacity value without the ".", which—if you change the underlying values of the scale—becomes misleading. But if the solution to that was to change to a different scale completely, eg 400, 500, etc the user misses out on having insight into what the underlying value is (when they haven't been changed). Hmm.

[stowball]

It's a tough one, that's for sure