>The tension/problem as I see it is this: we need distributed networks, where people run their own services, and are servers as well as clients, but at the same time a large proportion of people do not or cannot run their own services. Solving that problem seems key to me--have people act as servers without realizing it (maybe something like with torrents).
We already have a model for this: email.
You and I can use whoever we want as email providers, including hosting our own, and if you don't think I'm a trustworthy provider, you can drop my messages.
The system isn't without it's corner cases, but it's certainly proven to be a robust model over the last ~30 years.
The only thing it requires is standards. We already have one for facebook messenger: XMPP. Facebook and Google used to support it, but they both shut it down because they wanted to lock their users in harder. We have CalDav for events, but again, none of the big players want to know about it.
We could have a "social status" standard, allowing you to send a robustly versioned life update to other servers from whatever social service manager you'd prefer, but you and I know that no amount of user outcry could persuade any of the major players to adopt it. Network effects, peer pressure and social expectation (e.g. please submit a snapchat along with your job interview) power their meteoric growth.
What we need is a "Regulating the Gauge of Railways Act 1846" for the digital age. We need international bodies to review software that has become mandatory to the modern world, distill it down to a set of living specifications, and then ram it down the throats of software vendors.
Come to think of it, you could build a Facebook, Instagram, or Twitter clone on top of email pretty easily. All it would take is a slightly reimagined client. Just whitelist addresses of people you "follow" and display their emails in a separate, more visually rich feed (e.g. automatically downloading and displaying images). Make the UI/UX of "posting" an email to your feed feel more like composing a social media post (e.g. "post a photo" instead of "add an attachment"). You could have a separate inbox for emails from people you weren't following/connected to, which would be just a traditional email inbox.
I think Facebook and Twitter work presicely because they take the free form content mode of an email (often a rich text document) and severely restrict it to a handful of more readily digestable options.
On any of the platform, there's only so many ways you can format a post. Facebook for example lets you do plain text, dress it up a bit, or attach some piece of content with optional text (a photo, video, or site embed), and then presents your post in a standard way after that.
It seems restricting when I put it all down on paper, but that restriction is downright liberating when you want to get something out quick. Choice can be overwhelming, and for better or worse, the lack of freedom in Facebook's post editor makes it much easier for folks around the globe to pick up and use. Email won't ever have that, and that's okay! Different tools for different jobs.
> Choice can be overwhelming, and for better or worse, the lack of freedom in Facebook's post editor makes it much easier for folks around the globe to pick up and use. Email won't ever have that, and that's okay!
I agree that restrictions are often liberating, and likely played a role in the success of the big social networks. But I disagree about email not being able to have the same kind of restrictions. It absolutely can. All you need is an email client that controls what you can put in the email body when you're composing an email. Most email clients allow pretty free-form messages - from plain text to hundreds of lines of HTML. But they don't _have_ to.
By the same token, there's nothing in HTTP or any other internet protocol that restricts the content of a Facebook post. Those limits are imposed by Facebook's clients (and presumably APIs).
Email has a widely understood delivery protocol and identities. That's really all you need to let people publish and subscribe to each other. The rest is just UI and marketing. (Which, of course, is no small feat!)
There are a bunch of email headers you could backwards shoe-horn into serving the purpose of allowing clients to recognize the difference between a life update and a traditional mail, or you could create a new one. You could then make it standard behavior to handle emails with that header via a social client if one is available, and otherwise discard/hide them.
Another possibility to ease UX in a traditional client would be to use the same subject line for all activity. Most clients nowadays will collapse those messages to a single thread, making it friendlier.
my question is, how to keep a standard email from being completely overwhelmed by traffic of the social network, assuming it isn't aware of the headers? Or, you'd just have to tell users to set up a suitable filter?
I think we already see some of these things to some extent like Github pull requests having a lot of usability via email.
I'm assuming that people participating in this network are all using a new kind of email client designed to work with these feeds.
On the receiver or "follower" side of things the client would filter emails coming from addresses you follow into a separate social feed box. All your other emails (e.g. from other random people, or the notification that the shoes you just bought online have been shipped) would be displayed in your standard, traditional inbox. Your list of actual, non-social-update emails would never be polluted by all the noise from your social update feed.
On the sender or "poster" side of things when you post something it would just send emails to your list of followers. Presumably all these people are using clients that are designed to filter the emails into either the social feed or the standard inbox. So you'd never end up spamming people with photos of your lunch on a Greek island with people who haven't signed up to see it.
Of course, it could get tricky if, say, alice@example.com is following bob@example.com and bob wants to send alice a plain old email that isn't part of his social feed. Using headers, as suggested above, could be great. Or maybe the social feed would be set up as a mailing list, and posts would go through a reflector address. Or else, maybe people would just set up a new email address from which to send posts to their network?
Or, maybe I'm misunderstanding your question?
I dunno. I haven't thought through everything, but I'm pretty confident that existing email infrastructure has all the raw material you'd need to build a workable social network.
Agreed that the existing infrastructure is sufficient.
My main question is how to support both direct and indirect (via a "social" client (as if email wasn't social)) usage in a sustainable fashion. Otherwise I don't see it gaining traction.
Yeah, supporting both uses would be critical. I think you're right that making sure it could interop with existing email clients in as painless a way as possible would be critical.
I'm not sure how much of a chance it would have of succeeding (I'd like it, but would anyone else?). One thing that gives me a bit of hope is the rising popularity recently of email newsletters on specific topics or from specific writers/thought-leaders.
>I'm not sure how much of a chance it would have of succeeding (I'd like it, but would anyone else?)
It may, because it doesn't have to build it's own network effect, it can piggyback on emails existing userbase, which is basically everyone.
I was talking with my housemate, we came up with the idea of a combination push and pull following mechanism. I can offer you the option to follow me by sending you a mail with the X-SOCIAL header and a "follow offer" header content, plus a really simple mail body that says "hey, substance just asked you to join their social updates list. it seems you don't have a social client, <click here> to get one". If you've got a social client installed already, the body gets hidden and you get the follow offer in your social client.
However, you could also send me a mail with the X-SOCIAL "follow request" header, and a similar introductory body: "Hey, skewart wants to keep in touch with you via social mail. Download a social client <here> to connect with them". If I have a social client installed, I get a friend request instead.
The great thing is, your email inbox has a timestamped record of all these requests, so it's like a somewhat unreliable replayable social history, which means if I don't have a social client, I can read the introductory body, go grab a social client and come back, and it will re-parse my inbox and offer me the same follow request, no need for you to resend it.
I'm going to start talking to a few people that might have good input and have a swing at writing a comprehensive draft specification, as best I can. Might not be a bad idea to have a proof-of-concept/reference implementation being developed alongside. Wanna take this to a different medium? Skype, wickr, steam, email?
People using socially non-aware clients would have to set up a filter. I would also include friendly "I see you don't have a socially aware email client, get one here" bodies with all messages. The data transfer becomes the marketing tool.
Email is insecure and was overwhelmed by spam, driving people to use systems (such as Gmail) that are good at filtering spam.
Any distributed system will need to figure out how to deal with spam and abuse. That's hard to do without there being a dedicated team to deal with flagged messages.
It's hard to deal with spam and abuse even when there is such a team or effort. That's one reason Twitter is struggling, because unlike Facebook (where people can at least be semi sure only friends/family/connections view their content) Twitter is open to the public to view anyone's profile that isn't explicitly private.
Moderation ends up becoming an awkward balancing act between not completely restricting freedom of speech and stopping trolls/bullies flaming the hell out of anyone they disagree with while simultaneous trying to make it scale. Which is perhaps even harder than determining between what's free speech and what's a threat/harassment.
If someone can come up with a social network that doesn't discriminate between people from different backgrounds/with different political stances while filtering out actual trolls and which manages to do this on a large scale without a huge team of human moderators, then they'll likely pose a massive threat to the existing sites in the market.
It's very true, but this is a known and heavily researched problem. We have solutions to it, compromised solutions perhaps, but solutions none the less.
The current state of social networks vs email is night and day. I would never host my own email (despite self-hosting almost all my other online services) because the anti-spam techniques in wide usage are so punitive if you get your config wrong (your IP and domain basically get permanently sin-binned), but I use gandi.net as my email provider, I'm able to choose to pay them money and in exchange get my mail handled by someone I trust.
I'd gladly pay to use facebook, but run by someone I trust. However, social services are centralized to such an extreme even vs email that I actually cannot pay to have secure and respectful social networking, as facebook has successfully locked everyone else out of the game.
I digress. Email is not perfect, but it's a hell of a lot better than social.
Then you miss out on the social discovery aspect which is a key part of social networking. You're just avoiding the problem of where to go by cutting off your own legs.
Seems contextual whether that's good or bad: one theory of sleep is that we lose consciousness so that we don't go out and get ourselves killed in low visibility conditions.
Social discovery is obviously something people want or they wouldn't use social media services in the first place. Pointing to the existence of other contexts where that doesn't matter is neither relevant nor insightful.
It's like saying 'why worry about these problems, just don't don't use the the internet at all.' I still like paper books and other things and usually spend a day a week off the internet but that doesn't help anyone who's trying to solve an internet problem, does it?
How do you mean? Whatever EEE-type strategy they may be pursuing, you can still switch from GMail to another provider and keep corresponding with the same email users without any effort. This is not true for modern social media.
We already have a model for this: email.
You and I can use whoever we want as email providers, including hosting our own, and if you don't think I'm a trustworthy provider, you can drop my messages.
The system isn't without it's corner cases, but it's certainly proven to be a robust model over the last ~30 years.
The only thing it requires is standards. We already have one for facebook messenger: XMPP. Facebook and Google used to support it, but they both shut it down because they wanted to lock their users in harder. We have CalDav for events, but again, none of the big players want to know about it.
We could have a "social status" standard, allowing you to send a robustly versioned life update to other servers from whatever social service manager you'd prefer, but you and I know that no amount of user outcry could persuade any of the major players to adopt it. Network effects, peer pressure and social expectation (e.g. please submit a snapchat along with your job interview) power their meteoric growth.
What we need is a "Regulating the Gauge of Railways Act 1846" for the digital age. We need international bodies to review software that has become mandatory to the modern world, distill it down to a set of living specifications, and then ram it down the throats of software vendors.