On Tue, Oct 6, 2020, at 1:23 PM, Dave Crocker wrote:
A coworker and I had -just- been discussing a nearly identical idea. I
mentioned it to Bron, who mentioned this draft to me.
I just had two things I thought might be worth chiming in on:
First: This one is a bit tedious, but since I was going to post here anyway:
the definition of emoji should refer to emoji characters *and sequences*, lest
it omit lots of non-single-character emoji. (Even the humble red heart is more
than one character.)
The other thought was about the question of what goes in the body.
Specifically, we discussed a means to distinguish "Look, I just want to send 👍🏽
and nothing else" from "I would like to write a treatise on the ways in which I
agree, but if you are not prepared to bask in the glory of my prose, I guess
you can summarize it as 👍🏽". This would allow (for example) a means to count
up reactions of a type, even when there's content, and would also allow user
agents to avoid drawing whole messages when the content was trivial. Something
like "Header-Contains-Whole-Message: yes" would sum that up, meaning "reaction
if provided, otherwise subject". It's out of scope for this draft as written,
but I think multiplies the draft's value.
ietf-822 mailing list