I have a dream since many years, and I think I start to caress it from a solution space standpoint.
I’d like to share it with you. I think we have the same ideal, but it is always good to put words on things we assume.
Since some year, Internet is accessed from mobile more than from desktop. So let’s discuss about a user that is using Internet from his free software android.
This user heard about matrix, and it sounds good for chatting with friends. (in the coming discussion, you can replace matrix with xmpp, smtp, gnu social, wordpress…).
The user goes to it’s favorite store (F-droid) and search for a matrix client.
In the proprietary and spying capitalism world, this is easy. The person that:
- develop the mobile app
- develop the backend
- host the backend
Is the same evil company. Hence the sign up process is easy.
In the free software world, 3 services can be offered by 3 different group of people.
For instance, for my emails, I’m using K9, with Gandi, and they probably use postfix and dovecot.
The complexity is a lot higher.
How to reduce this complexity, improve the user experience, and make it in a standard way?
I think the 2 best examples out there are mastodon and Nextcloud.
And Nextcloud did the extra step to embed this in their mobile app.
So, what do we need? We need:
- a list of hosting providers
- s standard sign up mechanism
For the list of hosting provider, we started to solve it with https://github.com/libresh/catalogue/ and this is what we are working on with https://lab.libreho.st/librehosters/librehost-api
(this is just the beginning, if you are or know a schema.org expert, it would be really helpful! )
For the standard sign up mechanism, I’m more and more considering https://www.openservicebrokerapi.org/ .
It is a standard way to:
- describe a catalogue of services
- the different associated plans (a Nextcloud account on a shared instance or a full Nextcloud instance)
- the creation of the instance of that service
- the creation of the binding (to keep Nextcloud, the creation of an app password, and the configuration of davdroid for instance)
I think we all agree on the premise of what we want in the future, a standard sign up mechanism for mobile application backed by a free software backend.
I think the solution space is mature with the proposed solution.
If you all agree on that premises, then we can get started to work on making this happen, but this is another discussion 
Thanks for reading me, and happy new year!