Future of the dialog element

[annevk]

This came up in #4918 but probably deserves its own issue. At the moment only Chrome implements this element and Firefox and Safari do not.

It would be good to find if we expect this to change somehow.

Things that would help here:

• Usage and demand from web developers
• Implementer positions

I think Firefox is somewhat reluctant at the moment, though not opposed to implementing the underlying primitives, but I'll double check on that.

cc @smaug---- @rniwa @tkent-google

[domenic]

With my HTML Editor hat on, I am supportive of removing this from the spec. It sits uneasily with our commitment to interoperable, multi-implementer features being in the HTML Standard. Although from what I recall, it has multi-implementer interest (with Apple being the second), in practice that has not progressed toward prototyping work.

In general, I expect all features will pass through a phase of single implementation + interest-but-no-prototyping from others, but if they stay in that phase too long, we should revisit them, like this thread proposes.

With my Chrome team member hat on, I believe at dialog's current level of usage, removing it is unrealistic, and I'm not sure we'd want to work to drive that down. It still is a good idea to have a dialog in the platform. (Although, the design of this particular one is not the best.) But moving it back to incubation, until such a time as it gains serious movement in becoming a more interoperable part of the web platform, seems like a reasonable thing to do. Concretely, this would look like splitting it off into its own WICG spec.

[mconca]

From a Mozilla point-of-view, I have not seen any arguments against implementing this, other than it is currently a low-priority and will take some amount of work.

Conversely, the primary reason for implementing the dialog element (and then promoting it) would be to ensure users see a familiar UX across websites that is also accessible. Today, websites either roll their own (or polyfill for Firefox) with implementations that range from nice to complete garbage. I suspect few, if any, of those sites test for accessability.

It appears the usage in Chrome is close to 6% of page loads. I found that to be higher than expected.

Maybe the spec needs some cleanup, but I don't see any reason to not implement this.

[myakura]

In [1] an Apple engineer added its skeleton and later in [2] they added a platform status entry noting "Under consideration". so there seems to be some interest in Apple on the element.

[1] https://trac.webkit.org/changeset/247527/webkit
[2] https://trac.webkit.org/changeset/249905/webkit

[domenic]

It's good to hear that there's interest from both Mozilla and Apple. However, I think my concern about keeping things in the spec too long (in this case ~6 years) with just "interest", and not "movement toward shipping", still stands.

[rniwa]

FWIW, Apple's WebKit team is continually interested in dialog element as a replacement for showModalDialog since we know some enterprise applications are relying on either dialog element or showModalDialog to exist.

[muan]

Usage and demand from web developers

👋 GitHub's front end team is still very much interested in <dialog>. It's just that from our perspective, <dialog> as the spec stands currently does not fulfill our needs completely (this is referring to the past/current issues around <dialog> in this repository). We are paying close attention, and will re-evaluate and/or adopt if changes were to happen.

[annevk]

Closing this as Firefox is actively implementing.