Social Media (Facebook)
Social media is a big focal point of decentralized open source development. Not only because it's probably the bigget use of webapps, but because it's hugely impactful in the wider world in a way that's appears to be increasingly negative.
Edsu isn't a social media platform, but one can be built using it. This fact is potentially a nice benefit in that it will allow for cheap experimentation with models; i.e. instead of every new idea requiring its own authentication and transport protocols, they can just use Edsu's. At that point the designers can focus on the social interaction mechanisms, which is exactly what needs sorting out.
One particular benefit of Edsu versus dedicated OSS social media initiatives is that it's valuable even if no one you know is using it. So there's a draw to have an Edsu account without Metcalfe's law's help. So if you were to create a new social media platform there's less of a barrier to get people to join if they're already Edsu users (just a permissions grant, rather than a creating a whole new account).
Furthermore, if the generic primitives of a social network are stored in a standardized format (direct messages, small posts (tweets), bigger posts (blog posts and updates), friends lists, etc.), then by simply going to a different Edsu app a user can use that new model without losing post history or connection with friends on other networks.
The simplest model is that of an aggregator: an app would store a list of people that a user follows, and would open an Edsu connection to each to see if there are any new posts, presenting them in a chronological list.
However, this shows up a scaling issue inherent in the simplistic use of Edsu - for each friend you have there needs to be a network connection. Say you had a friends list of 300 - web browsers will only open around 200 websockets, so they couldn't be concurrently updated, and in any event it would be a resource hog.
This may be a feature not a bug, in that perhaps it limits things to the Dunbar number, but in any case there's a variable degree of centralization that can be reintroduced.
For example, people could use their Edsu accounts as aggregation nodes. One way is for a user to open up permissions on an appendable name block to a group of friends, and when they post something on their own account they append to that "shared" block. So by subscribing to that name you can monitor the activity across an entire group, and only open up connections to accounts when they have something new.
You could in turn aggregate subscriptions to these appendable names to form a trees and other complex structures, spanning thousands or even millions of accounts.
There's also, of course, the option of caching the post content of these aggregation nodes. This way there isn't so many network connections being created. It's eased somewhat by the fact that caching is simpler with Edsu that a lot of other protocols, but still raises issue around intellectual property and the right to be forgotten, etc.
Having said all that, Edsu is focused on people as individuals, not as part of organizations, "audiences", or any of the other usual focuses of proprietary social media. So it's entirely possible that Edsu might not be a good fit for mass media applications - that's OK, Edsu is not meant to be everything to everyone.Next: Source Code Hosting (GitHub)