reactions

forking from %p9WE+Ez..., because i'm interested in splitting vote messages into "reactions" and "flags".

back when Patchwork was called Pheonix, over two years ago meow, i think before our current network (we had a "hard fork" back in the day), we used to have free-form text reactions. since then, Slack has popularized free-form emoji reactions, which has then been copied (with less choice of reactions) on Facebook, GitHub, Loomio, and others.

here is a proposal to change vote messages into react messages.

why?

like or not to like

message schema

draft

{
  type: 'react',
  expression: 'agree',
  link: msgId
  mentions: [
    {
      emoji: true,
      name: 'agree',
      link: emojiBlobId,
      size: sizeInBytes
    }
  ]
}

user experience

rather than a "Like" button, we have a "React" button. clicking this will bring up an emoji palette (could be code shared with the existing emoji autocomplete), which allows you to choose any emoji that accurately represents your feelings, including any custom emojis

open questions

• what have we learned from the vote message schema, what do we want to keep or change?
• is it worth creating a new message schema, why not re-use vote?
• do we allow for free-form text reactions?
• what is an "expression"? is my expression the same as yours with the same words, or is that not supporting subjectivity, is this something like custom emojis where each expression is a unique message, or is using custom emojis enough?
• is using mentions here good or bad?
• how do you take back (remove) your reaction?

i don't have the answers to anything, i'm only starting a conversation. pile on!

To be a frist class thread citizen, this schema needs further thread-properties:

  root: MessageId,
  branch: [ MessageId, ... ]

https://twitter.com/pfrazee/status/1131227762637836289

Unicode seems more efficient but being able to add your own blobs also seems like a worthy idea just a question of efficiency i suppose. I’m also not entirely sure what happens when i do “: grinning :” in patchwork - does it upload a blob or just make reference to it?

Either way I think it’d be important for clients to have UIs where commonly used emojis are in the ‘default’ list - that way groups can form their own language of emoji. Perhaps just adding up whatever is in the cache - and sorting by count would be enough. It also might help with not uploading new blobs every time (though tbh not entirely sure how that would work and whether you’d risk using an emojiBlob not in the follow graph of one of your followers) ???

@utunga3 ❀: agree on the subjective language of emoji! at the moment, i think all clients (except for maybe Patchfoo) don't do anything special with regards to emoji's, what you see in the text (:grinning) is what is published.

Matt had a great idea some time ago to change this, so every emoji reference was replaced with a blob id (in the same format as custom emojis), while also publishing the blob for that emoji using Patchwork's local file as the blob content. this works well because blobs are de-duplicated by design due to their id being the hash of the content. also it means we would avoid the unfortunate situation where different clients have different emoji palletes, so i might have meant to say one thing but what's rendered on your client is subtly different.

@gwil: you might like "like or not to like"

my personal opinion is %djiMk7U...