Permissions

Friendica Red and Permissions Some more forward-looking documentation about the Friendica Red project...

For those of you that have been following the previous discussions, the zot protocol will give every communication entity in the decentralised network a unique ID with the ability to communicate securely.

Red, unlike its predecessor Friendica, will only use one protocol stack to communicate - zot.

This opens up some possibilities for interactions which are impossible in Friendica, in particular some very fine-tuned permissions.

The first place you'll see this come into play is with the "make friends" dialogues. In fact, unless we provide a "simple mode" it won't look like a social network "make friends" dialogue at all. It's just a permission system. You will be granting a person some number of permissions - any of which can be removed or revoked.

These will probably include:

[ ] Permission to receive stuff from me (means I will send them stuff) [ ] Permission to write stuff into my stream [ ] Permission to write/post/tag stuff to my wall [ ] Permission to send me private mail

[etc...]

This is actually a superset of all of the "follow" and "share" async friendship models. In order to model the "mutual friends" aspect of traditional social nets an additional option is called for, something like:

[ ] Revoke any of these permissions which are not reciprocated

---

In the Private mail subsystem there can also be a range of privacy settings which there is no possibility of implementing in Friendica due to its polymorphic nature

Accept private mail from:

[ ] Those I specifically allow [ ] People on this website [ ] Friends of Friends (e.g. if the mutual friend allows them to send mail, so will I) [ ] Anybody on the network

---

I will need help from some of you folks that are UI oriented to find ways of presenting this information in simple dialogues that make sense to anybody. (And then of course making it all work.)

There is a privacy aspect to this... in order to make some of these models work, the permissions assigned to a particular person need to be discoverable and therefore your friends will know precisely where you stand with them. We still will provide a "hidden" contact where this information is not made public, but consequently cannot be used to decide if a person is a friend of a friend.

This may seem to be a bit of a liability, but you will also be able to assign a "closeness" value to anybody you have given permissions to - and this is private to you. So you can give somebody full permissions, but put them out in the fringes and rarely see them, even if they're the "noisy" type. Nobody will know where they stand in terms of "closeness" unless you tell them.

It may also be possible to extend the private mail settings dialogue and re-use it for allowing wall postings and also stream permissions. We'll have to see how this evolves.

In summary, Red is going to be an experiment in redefining the decentralised social network in some very fundamental ways. Many of these features are possible in centralised networks, but providing them in a decentralised network has not (to my knowledge) been done.

Your input is important - especially at this stage when none of this has actually been committed to code. We want to get it right and let you do stuff that can't be done with any social tool today.