Hello !
First of all, token are great and I love the fact that you can reference one token to another and do calculations.
I tested the system on the DSFR design system (french government official design system) which is the perfect playground for this.
Here is some feedback from my journey (numbered for easy referencing):
- I tried to import tokens from a Figma style export and some manual fiddling. FYI, if there is an error message at import talking about invalid characters in token name even if there isn’t any, it’s because you have to add the token group in the “active sets” part of the JSON first.
- Nothing tells you that a token is used on an element, only the computed value is displayed, so it’s easy to replace it with an hard coded value and don’t see any difference (and ruining your design system in the process
). I believe the next updates address this. - No OKLCH (or LCH) support in tokens, it would be great for color computations and accessibility!
- Tokens are alphabetically sorted (I guess) but it’s lexicographical and not numerical so 10 is before 2 in the list.
- Pills presentation is compact but can be a little messy, a list display will be great to quickly find tokens.
- Is there a context/theme setting by artboard like in Figma? I couldn’t find it but it would be very useful when you don’t want to switch your complete file in light, dark or mobile, and when your file is dedicated to an interface or a product flow, with mobile and desktop pages sitting together in the same file. Or when the file is a display of components with different contexts (like login modal in light and dark).
- When you create a collection, it’s not obvious that you have to activate it so the tokens are visible/usable. I scratched my head 10 minutes on this, not understanding why the color won’t apply when the token was active and selected. Maybe start a collection activated?
- Can I link a color token to a color style (asset)?
- The token export JSON is not in the same order than in the UI, nor alphabetical. I used the “JSON > sort document” in VSCode so no big deal, but a sorted export or an export “as is” could be useful.
- It seems that you can’t move token from a group to another via the UI. It would be great to organize your token without opening a code tool and do some design ops directly in Penpot (we are designers after all ^^).
- No emoji or unicode in the name of tokens, I guess that’s because the W3C standard says so. Not a problem but I was surprised.
- Problems for accessing tokens in groups: the expand arrow is buggy, it expands random other groups or don’t expand at all. I’m guessing a problem on arrow/content IDs.
- No auto-completion in token aliases (when you start to type
{…in the value field). Would be very useful when you have a lot of tokens. I believe I saw auto-completion in the token presentation video, but it does not work on my files. - No search available in the token panel. (Update is coming I heard, if possible allow the search in the token panel and not just in the individual fields).
- The token description is not available on hover, you have to right-click/edit to know how to use the token. Maybe add the description when hovering and in the list presentation? (see above).
- When you delete a token sub-level group (right-click on it > delete) the parents groups are also deleted up to the first ancestor and no ctrl+Z can save you
- When you select a token with a font family and try to apply it on text, the text stays the same and the font field in the right panel displays “deleted font”.
- Tokens are not part of a library? That kills the principle a little, and kill your entire design system if you delete the file containing the tokens (I read in another post that it’s like this for the moment but will evolve soon).
I can’t wait to see the next update with tokens that can be applied in each individual fields, it will be an enormous leap forward for the quality of life! 
Keep up the good work and thank you for giving us an alternative to Figma!
