Closed tantek closed 7 years ago
You'll have to help me understand this a bit better. Are you saying that newer kinds of replies, for example edits, should also include the u-in-reply-to
class along with their own class (u-edit-of
in this case) so that the author of the post can at least show it as a reply (if reply becomes the 2nd to last step)?
@prtksxna precisely! All (especially newer) responses should include the u-in-reply-to to enable fallback behavior with p-summary with any consuming code that may or may not understand that particular type of response.
Discussed in https://www.w3.org/wiki/Socialwg/2017-07-25-minutes#Post_Type_Discovery today's Social Web WG telecon, with resolution of accepting editor proposal, and approving new WD with that change.
Great! Thanks for the logs.
Changes made in repo and published updated WD: https://www.w3.org/TR/2017/WD-post-type-discovery-20170801/
Based on implementer feedback from @aaronpk and @sebsel, it may make more sense to move the "reply" recognition step to 2nd to last in the Response Type Algorithm, that is, right before the "else it is a mention".
Specifically, anyone posting a "like", "repost", "RSVP", (or any other new response type!) has incentive to provide u-in-reply-to fallback plain text equivalent perhaps in p-summary that expresses the reaction/response as prose, e.g.:
This way even if the receiving site only supports receiving Webmention replies/comments, something sensible will still show up, and provide a good experience to the author of the response, the author of the post, and any readers of the post.
If we formalize this kind of fallback markup as a publishing guideline (perhaps a feature request for https://github.com/microformats/h-entry) it also provides a nice path forward for expanding response types over time that provide meaningful behavior even to existing systems without explicit support for the new response types.