Re: [csswg-drafts] [css-gaps-1]: Change the initial value for insets property to be 0 (#13137)

Reopening this issue because the second half resolution sparked further discussion about the nature of the proposed keyword and what its introduction would entail. The second half of resolution proposes to define a keyword (say `join`) that would serve as an alias for using an inset value of -50% to join segments.

The concern is that introducing that keyword in this way implicitly commits us to a very specific interpretation of "joining" which may not reflect author intent in many cases. A -50% inset join caused each segment to extend its interior endpoints to the midpoint of the cross-gap. While this produces a form of joining, it is only one of many possible ways authors might expect segments to join.

Take the following example a `join` keyword would take us to the middle of the gap which visually produces a notch whereas authors might have expected a join to be something like joining at the start of cross-gap decoration.

<img width="361" height="361" alt="Image" src="https://github.com/user-attachments/assets/66c3979b-0800-467b-97b4-f971e3a545d0" />

There are also some quirks when joining segments using insets -50%. We can see authors using the keyword and getting stop-and-restart behavior at segment endpoints which might not be desirable when thinking of joining segments. This becomes particularly
noticeable when dashed or dotted decorations are used. See:

<img width="157" height="159" alt="Image" src="https://github.com/user-attachments/assets/826fd07e-aa1c-4cb3-88fa-6b03b782cf32" />

Broadly speaking, while there is definitely a clear appetite for a keyword that controls how insets join segments, there are many plausible joining behaviors a web author might want. At present, we do not have sufficient signal or concrete use cases to determine
what that right default join should be. Committing to one specific choice (i.e. -50%) into a keyword at this stage feels premature and risks locking in an unintuitive or limiting behavior.

For these reasons, we propose deferring the introduction of a `join` keyword to a future version of the spec, once we have better insight into real‑world author needs and expectations around segment joining. In the meantime, percentage‑based insets already give authors the flexibility to experiment with and achieve a wide range of joining behaviors on their own, without prematurely standardizing a single interpretation.


-- 
GitHub Notification of comment by oSamDavis
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/13137#issuecomment-4000657017 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 4 March 2026 22:18:32 UTC