You’re right. Not all want them. For myself I never had a GIF-avatar and think it’s silly. But I still think that this is a reasonable feature for a lot of users. My timeline in Diaspora is full of animated GIFs, although I regularly unfollow people who are posting too many too large GIFs.
I’m the same. That’s why I think it makes sense to enable animated GIFs but give users the option to disable them in user preferences.
Everyone, there seems to be consensus on this, shall we have a proposal and a quick vote?
Proposal: Officially support animated GIFs, with opt-out
As seems to be the majority wish in the discussion so far:
- provide proper support for using animated GIFs in posts, avatars etc.
- provide users with the option to disable all animated images in their user preferences.
This seems to be the best solution which pleases both those who like animated GIFs and those who don’t.
I’d suggest voting on this proposal and implementing what is voted for as soon as possible, and then deciding about other formats at a later date.
Outcome: There seems to be (qualified) support for the principle of supporting GIF animations in avatars and posts. Next step: look into how this might be accomplished, and whether it’s technically feasible to do it in a way which works well for the network.
Votes:
- Yes: 12
- Abstain: 1
- No: 0
- Block: 0
Note: This proposal was imported from Loomio. Vote details, some comments and metadata were not imported. Click here to view the proposal with all details on Loomio.
Why specifically gifs by the way? Gif is an inferior format for all purposes. If you want animation as vector - better use SVG, that’s the right way to do it. If want animation as raster - use video as others proposed. WebM with no sound should be a good candidate.
I.e. video can be even GPU accelerated. I doubt animated GIFs specifically are. So performance wise small video file of the same size will be preferable.
Because GIFs seem to be the most commonly posted animated image format, and what was ‘brought to the table’ in this discussion by Sean (and what was being discussed in Github as well). I’m certainly not meaning to rule other formats out, simply saying ‘let’s deal with this one issue now and look at others later’.
As I said below, I hate GIFs (and animated images in general), but it’s what people seem to be using at the moment.
Commonly used doesn’t mean it should be proliferated and burden D* users with it (if its performance is inferior). If user wants to go out of the way and create an animated avatar (creating animation requires some skill), user can figure out how to create a webm video. If not, someone can put a tutorioal. I don’t see a problem with offering a good option which isn’t common, and not offering a common bad option.
I think such setup can work out:
- Give an option to add WebM avatar (limit resolution and file size, somehow limit it to no sound - needs to be researched how to automate that).
- If browser can’t show video - fall back to showing static image for avatar, otherwise show that WebM in a video tag.
Can we have a longer voting time for such a controversial topic please?
So what technologies are out there to accomplish the proposal? I think we should investigate that first.
@shmerl support video is something completely different. Here we simply have a tag and users used it to put gif. We did not officially say that we support animated avatars because it can be too heavy for small servers / connections / mobiles. So if we want to officially support animated avatars, we have to put an option to podmins / users to display them or not.
We are far away than a feature which would request a tag. Too many problem : it’s heavier than animated images, so it’s worse for bad connections / servers, only the recent browsers support it, it’s not the same codec for Chrome / Firefox (webm) than for Safari / IE (H264)… We have many things more important to do than implement video for avatars.
It’s not different conceptually. Forget about tags for a moment. The concept is simple - animated avatars. First think whether we even want it or not. Let’s say we do. How to implement it in efficient way and with sensible fallbacks?
it’s heavier than animated images
This is very questionable. If you want to research it seriously, create a test case with 50 animated gifs and 50 WebM videos of the same resolution and duration. Test the CPU utilization in both cases (when all are present on the page). Then you can have a conclusive understanding. My expectation - WebM will perform better CPU wise, but nothing is of course guaranteed. I wouldn’t worry about crippled browsers like IE and Safari. They are still catching up for many things, and WebM support is one of them. They can have a static fallback if they can’t play the video.
We have many things more important to do than implement video for avatars.
If so, you shouldn’t be busy with animated gifs now as well, since it’s the same thing, but with wrong tools.
So my view on this - either do things right (research, test, implement the video avatar with static fallback), or don’t even bother with it until you have time.
However if this is just about “keep no limitations on img which potentially allows using SVG and GIF” - that’s a different scope of the discussion.
However if this is just about “keep no limitations on img which potentially allows using SVG and GIF” - that’s a different scope of the discussion.
The beginning of the discussion was that.
About research of what we should do to correctly deal with that, I’d really prefer to focus on these things for the moment.
Hi Jonne, I thought it was worth creating a proposal as the discussion had seemed to stall some time ago but there seemed a reasonable consensus for this one proposition. So I thought I’d create the proposal to focus our minds: is this actually a good proposal and should we vote for it. I’m quite happy for it to be shot down.
If you don’t think that enabling animated GIFs is a good idea even as a short-term solution, please argue against it. I’m quite happy for my proposal to be shot down, and if I’m convinced it is not a good proposal, I’ll change my vote.
I’m not trying to force us down any route; simply to provide a means to speed up the decision-making.
This is another “do this!” proposal without any technical background. So I repeat my question:
What technologies are out there to accomplish the proposal?
I still don’t like and that’s my opinion and I’m going to vote on that. No further arguing required than I’ve already done. I just wanted to point out that we’re voting on something while we have no idea how we can implement it.
And the second point is too directly what I said, 9 days isn’t enough time IMO. Nothing meant to subversively to boycott the proposal. I mean what I say.
Well, one way we could do it is programmatically with a bit of JavaScript. A quick search of StackOverflow brings this up. Making it work cross-browser is my only real concern here.
We could just call a variable from user settings asking whether the user enabled for gifs avatars to animate, if true, gif avatars are animated, if false, they simply stay as the first frame of the gif animation.
Sean, javascript in the browser is a nice approach. If this is not possible to make this happen to all browsers OR if this is not performant to the user, we can also create a dynamic URL that flattens a given GIF via ImageMagick into the first frame and delivers this image to the user, who does not want to see animated GIFs. This would have been my first guess.
Personally I think javascript should be avoided here and the user choice respected when offering information from the server. What about when a user uploads a profile pic and if it is an animated gif - another version is created which only contains the first frame. Then depending on user settings the correct version is offered.
How would this suit the way we store profile pics and federate them?
Also the approach by @rasmusfuhse sounds good to me as well.