Gif support

[dholms]

They're fun & right now we only support jpeg & png

[bnewbold]

I think the common consensus these days is to convert to video though. GIF animations are often huge and would get mangled if resized on upload to fit in, eg, 1MB limit. It is basically video, in terms of bandwidth, UX considerations, etc.

While we are at it, could also consider webp and APNG animations. I don't have any market share numbers about how prevalent those actually are, so maybe not a priority.

But, totally agree fun!

[TodePond]

+1 to this. They're needed by the '#BuildInPublic' users. ie: Developers/artists/creators post short work-in-progress gifs of what they're working on.

I think it would be worth prioritising this, because without it - those users can't migrate their posting over.
eg: https://twitter.com/steveruizok
eg: https://twitter.com/lorenschmidt

[nazgum]

im an indie game developer, twitter has (or at least had) a really popular community of indie game devs who share gifs of their progress on their games.

would be nice if bluesky/at supported them so this type of content can be posted there too

[Scrxtchy]

While we are at it, could also consider webp and APNG animations. I don't have any market share numbers about how prevalent those actually are, so maybe not a priority.

webp has rather great usage support across browers given that it is also supported on ios versions 14 and above and android 4 and above. apng has greater support for older devices/browsers, but it's compression methods aren't as favourable to webp.
Another project I have ties to considered apng for animated user avatars, but it didn't have much favours over the output file sizes that were given when compared to webp

The avif image format has a growing but not comparably high usage/compatibility in browsers, it's based on the AV1 video format and compresses better than webp in some cases, but getting data on how well these are support across non-browser experiences didn't bear any fruit

I think other data like it's encoding and decoding performance would be important for both client and server, but I don't have good data on these