Pixelfed – An alternative to centralized image sharing platforms
pixelfed.social
pixelfed.social
This feels like a huge barrier to adoption for them beyond the nerdiest of nerds.
I like Instagram because I get to see what my friends are up to. How will I get my friends on this?
So indeed, it is a bit of an awkward time for it to be on the front page here.
But yeah, this is more or less federated Instagram, but the federation is still a work-in-progress, AFAIK.
Though - as that page points out - it's still a work in progress. There should probably be huge barriers to adoption for everyone regardless of nerdiness. But at the moment I'd guess they're targeting existing fediverse/Mastodon users.
To be quite honest that's the same reaction I had when I first landed on Instagram's homepage. Then of course, everybody was talking about it...
I think the lack of a common name/URL is the biggest problem that is preventing federated open source social networks from gaining mass popularity. I remember when identi.ca came out as an alternative to Twitter. It seemed like it was starting to gain some traction, then they switched to pump.io and closed registrations. I haven't heard of anyone using either since. Look at Diaspora, that never took off either. You might convince someone to try it out, they'd go to the website and it'd be too complicated for them. "Which pod do I choose?" Then they would just go back to using Facebook and Twitter.
That is actually a great UX question. For federated services how do you think we best solve this? I mean everyone wants to just not pick one. But different loss suits people better in terms of location/interests/something else.
UX folks, what patterns would you suggest to help solve this?
For distributed services, the solution must be such that the question "Which pod do I choose?" makes no sense. (There may not be a solution, in which case this may be distributed solutions' Achilles' Heel.)
Email is a mature example of a federated system, and so if there is an answer to the general question, I'd expect we'll have figured it out in the context of email.
So... what's the "general strategy for picking an email provider?"
AOL Instant Messenger didn't come out until 1997, email was already in heavy use by then.
Pick Gmail unless you have a good reason not to. And if you don't know any reasons pick Gmail.
Federation sounds nice until you realize that most people don't have particularly unique needs and a 'best for most people' will eventually emerge and dominate the market.
You are also free to set up your own instance(s) and you can follow anyone.
Otherwise, stick to an instance with people like you (i.e. don't put your account on a gardening or knitting instance if you are primarily going to post and read about FPS gaming as it will mess up the instance timeline for everyone else and you'll have little use of the instance timeline.)
Nowawadys it’s either hotmail/office365/whatever other name it has or gmail. It used to be a federated system and still sort of is, but your experience is vastly diminished if you aren’t using one of the biggest providers because you’ll have deliverability problems.
I was working on it when I worked for ebay and PayPal. I was working with Yahoo (who was big at the time) on what became DKIM, which was created to make sure that Yahoo mail was delivered successfully and that ebay and PayPal mail would be delivered successfully, among lots of other big corps.
Sure, anyone can use those technologies, but only the big players run interoperability tests on a regular basis, and provide reporting to each other, and other big corps.
Scroll a bit down and you'll see that you can filter through the servers based on your language and interests, as well as see the amount of users each server has. https://joinmastodon.org/
Unless the auth database is also distributed and you get to log in as yourself from any pod, it does matter.
As for the present, I'm sorta baffled that the anarchic dream of decentralization turned into not-quite-that-cool federation where you're tied to an instance, instead of the identity working across the net. If my chosen server goes down, I'd love to be able to just move to another one with my keys.
The one your employer provisions your account on. Your account will be @user@company.domain
Just silly ramblings that hopefully open up a discussion because god knows I don't have the answers, but I do have ideas.
1. How do I get my friends to use them - Solution? Maybe we now live in a world where privacy being part of the public discourse would enable something wonderful like a rich person who cares about this to spend for an ad campaign altruistically (I'm thinking in the vein of Brian Acton giving money to Signal foundation). That would be an effective seed to start a social network. Really play on the failures of modern social; Facebook (privacy), Twitter (sloppy moderation & bots) Google+ (...)
2. For the most part, no one cares it's federated. The upside of federation is "unstoppable" apps (to steal from Ethereum). Okay fine it's federated, just make the UX of finding and jumping into those fediverse's easy. Pinterest model of discovery would lend itself to fediverse discovery, i.e I type a search term - show me related fediverse.
3. Make it known to the people it matters to, in simple terms about how you don't track them. This is really really important. People care about being tracked, but unless you make it super simple and informative that what you do is different, how can I see that using something that is open source, federated and made by people who care about privacy is actually going to reduce tracking? Startpage and Mozilla do good work here about describing the state of the union, and how what they do is different.
The UX for boosting, replying and interacting with users on other instances is pretty straightforward, you merely fill your handle once when trying to boost/reply to a post found on another instance (if you don't follow or otherwise have a copy of that toot on your local instance) and it will OAuth (IIRC) to boost/like, or redirect you to your instance if you chose the reply option (with said toot your replying to pulled up). Overall pretty slick IMO.
One factor in this is there is generally not a wall between linking users to content on the fediverse, and them being able to view it. No Quora, Twitter or Facebook full screen popups that cut off reading a thread or toot, just here's the thread :P
Open linking/access to public toots unlike other "social" media is a big differentiator.
I also have my own email domain hosted at Fastmail.com (I don't want to manage an email server).
This way I'll manage MY own data. I'll share it with who I want. And it's a fun project. :)
Anyways. It's fun... I post silly inconsequential pictures because I know only a half-dozen people will ever see them. The public timeline is not very lively so I hope w/ better federation and more attention it can be a little more active.
I'm pretty sick of Instagram though, so Pixelfed and the fediverse in general seems fun.
The web app is fine and works fine on mobile. A little clunky, but that's how I likes my indieweb. :)
By not supporting following in the other direction, these specialized platforms can stay focused on the content shared there, instead of also supporting microblogging, video, photos, events, and whatever other types come up in the future. But of course you're not limited to just following in one direction -- comments go both ways. So someone can comment on your Pixelfed photo with a Mastodon account, and you can respond to that comment with your Pixelfed account, and they'll see your comment back on Mastodon.
This is another reason to prefer federation: your social graph is not owned by (at the mercy of) a business whose interests may not even be aligned with yours (bad advertising is only one of the many consequences of this).
> extreme
I'd not call "there's something better" an extreme approach. The maintenance burden is what makes federated services difficult to operate. That's the issue right now, but people are working to resolve that too.
> https://blog.datproject.org/2019/03/22/three-protocols-and-a... > https://news.ycombinator.com/item?id=19466929
The comments there show other problems too. Other problems that may never be solved. For example not being able to take your social graph with you if you change Mastodon instances. Having issues with the admin or an instance seems like it’ll happen often enough. But you’re essentially punished for that. That’s just one example.
It is good though that people are trying to solve the problems. Hopefully my cynicism will be proven wrong and I’ll only be happy about that.
I wonder if they are open to patches that would make it federated.
> Why isn't Tildes decentralized/distributed/federated?
> Decentralized communities are interesting and have a lot of potential, but that model also introduces its own problems and difficulties. Tildes is already attempting to do quite a few things differently to improve the quality of online communities, and I'm more interested in focusing on those goals without introducing the additional complexity of decentralization.
> However, since Tildes is open-source, someone else could certainly use it as a base for their own decentralized version.
I'm not interested in running a decentralized version myself, but I'd love to see someone else try to make it work.
I send out invites daily. It's not intended to be difficult to get an account, I just want to prevent it from getting out of control when it gets a huge burst of attention due to situations like a subreddit getting banned.
Thanks for all the feedback! The project is still pretty young and we look forward to shipping a stable release in the coming months!
A couple of questions:
- What is the business model? If millions of people are on this and there's no ads/monetization -- how is the service covering costs?
- If people had to pay, then it might limit adoption, how are you going to handle that?
- How are you guys thinking about support, monitoring content (removing illegal/copyrighted content)
On Mastodon, most use Patreon, some are pay to sign up and some don't really accept donations at all.
Regarding moderation; I'll keep my family and friends under control and bigger instances will have to delegate moderation rights to trusted members of that community – much like on any forum. It's a far more sustainable model compared to the nightmare Facebook and Twitter, which is relying heavily on algorithms and outsourcing.
How does this scenario work with pixelfed? I take it that the content is hosted on a specific server and that the pixelfed app is just a frontend that knows how to find that server. Is abuse reporting built in to the app somehow? How does the end user know that action is being taken? How can law enforcement close down pages?
Note that legal liability for hosting illegal content may be huge. There is one 18-year old in NZ who is being charged and threatened with 14 years jailtime for sharing the terrorist video from the mosque massacre. https://www.nzherald.co.nz/nz/news/article.cfm?c_id=1&object...
Users can report posts which then has to be looked at by the mods on that specific instance/server. Since most servers have a modest amount of users, and a relatively high moderator to user ratio, moderation becomes quite manageable.
- If an instance is going to shut down for any reason, how you can handle your account and your content? Do you lose all your follower base and/or content? Can you transfer your account to another instance?
Questions about the whole federated ecosystem that I never saw answered completely.
PixelFed has a possible awesome future ahead!
Supported Fediverse Projects
This is a partial list of well known supported projects
Anfora – Self-hosted photo gallery social network.
Pleroma – A federated microblogging alternative.
Mastodon – A federated microblogging alternative.
Misskey – A federated microblogging alternative.I can see why a for-profit app like Instagram would force everyone to sign up to browse, since that helps their bottom line, but I can't see why that decision would be made here.
https://www.instagram.com/explore/tags/ycombinator/
This seems to be the url to browse tags on pixelfed.social:
https://pixelfed.social/discover/tags/ycombinator
Unfortunately it wants you to sign in.
By that I mean, does Pixelfed integrate with Mastodon in such a way that a Mastodon server could also run a Pixelfed server instance, and use the same user database for both services?
For example, the Mastodon server I am on, vis.social [1], seems like a place whose user base might want to use Pixelfed. That Mastodon server describes itself as “a social space for anyone in data, visualization, creative coding, and related arts and research.” (Said description is what prompted me to choose this specific server when creating my Mastodon account.)
If you really want people to move away from social networks run by for-profit corporations then you need a single non-profit organization with a charismatic leadership, a captivating mission statement and lots of people to create public awareness.
Then you just moved the centralization to a different entity. Wikipedia has plenty of problems, power plays, politics, bullshit process, bias etc.
I'm trying to justify to mysef how this is useful in any way but it sure _feels_ fascinating!
Can someone explain this?
Keeping registrations open on what is often perceived as an "official" instance can lead to that instance getting most of the new users of the network, since it's usually the one you'll find if you search for the corresponding software. mastodon.social (previously open, now invite only) is a good example of that, IMO. I'm not sure if it's a good or a bad thing, but I do feel it subverts the idea of a federated network if most users are on a handful of known instances.
I don't think it makes sense to have non-active members just for faking the numbers if it's non profit anyway.
Is this compatible with existng solutions like Mastodon?
And why just images? Why not videos/documents/music or even events?
How long will the images be available? Who is paying for storage?
How does the technology compare to IPFS?
Pixelfed is fairly new, and probably should get it's minimum viable product working before worrying about supporting more formats. And there's a lot of ActivityPub-based software out there, it makes sense to focus on a specific experience people are looking for.
You host Pixelfed where you want to, or use someone else's Pixelfed server. As long as either you or someone else is paying to host it, it's there.
ActivityPub style federation isn't like IPFS. Data is largely still contained on the server it's posted on, but you can communicate with other servers so that people can be notified, interact, and comment on things posted on other servers. But it's a much more traditional server model than IPFS. When a Mastodon or Pixelfed server goes offline, it's content is gone.
2) It's purposeful specialization. PeerTube is a video hosting service, Funkwhale is a music hosting service, GetTogether is an event hosting service. No idea what services provide online document services - they likely aren't ActivityPub-compatible.
3) Ask the administrator of your instance. The Fediverse, of which PixelFed is part, is composed of a mesh of networked instances.
4) I don't have enough IPFS experience to answer.
Nextcloud is ActivityPub compatible and might suit that use case.
I think more things should be "web first". Today even Reddit keeps tirelessly pushing an app - for a web site!