Oh boy, these two are starting to look really similar!
I'm curious to know what the difference between the systems is in terms of procedures, governance etc. are? What lessons have been drawn from two decades of XEPs? (Or three-and-a-half decades of IETF)
How does one make sure the system still works after, let's say a decade, when it is no longer exciting and new and everyone is full of energy? When it has become yet another grueling technical bureaucracy like any other real world (open?) standard which needs to deal with legacies, weird real-world requirements and parties which have really different opinions? Not even touching upon corporate capture here.
This is not meant to disparage these processes at all, by the way. Rather, these are genuine questions on how to make this work because we need this kind of stuff to work well for us. We need protocols not platforms.
"I'm curious to know what the difference between the systems is in terms of procedures, governance etc. are? What lessons have been drawn from two decad..."
Matrix is a monolithic specification where the standardized bits are made in harmony with the whole and there is one way of doing one thing. You only need to look at one coherent specification instead of various extensions.
Only things proven in practice land in the matrix specification.
The governance is entirely focused on Matrix.
"Only things proven in practice land in the matrix specification."
If things are proven in practice before they land in the matrix specification, they still need to be specified before, so developers can pick that specification and implement it to prove it in practice.
This pre-standard specification is what XMPP people call an experimental XEP.
Of course, if you only have a single client+server combination widely used in practice (element+synapse), implementing just in those makes it already proven in practice and if both server and client are developed by the same party, they may not publish the specification early on.
In XMPP, the first iteration of an upcoming specification is typically written early, so that it can be implemented by multiple independent parties at the same time, so they can test against each other.
Further, just take a deeper look at the spec, e.g. TOC
a question: isn't the protocol building a plattform?
I would like to see what happens when the capitalistic corporations and the vc's move on.
Will there be any vollunteers at that point?
Plus the matrix protocol needs a serious rework to adopt anything.
From p2p to low bandwith requirments
Yes, this is a very good point. I think community collaboration to evolve the #OpenStandards is essential. Healthy community, active participants and good governance. It is the Achilles Heel of the #ActivityPub #Fediverse.
See the announcement here:
@rra >How does one make sure the system still works after, let’s say a decade, when it is no longer exciting and new and everyone is full of energy? When it has become yet another grueling technical bureaucracy like any other real world (open?) standard which needs to deal with legacies, weird real-world requirements and parties which have really different opinions?
We’re already at this point. The Matrix spec is large enough that implementing a client with complete support requires a huge time investment. Implementation of an alternative server backend is essentially infeasible; even official efforts at doing this are moving incredibly slowly. A spec of this size and complexity is antithetical to a diverse federated ecosystem, which means 99% of people get stuck with the one-size-fits all Element client. That said, the reference implementations have steadily improved, and Matrix remains the only logical successor to IRC accessible to normal non-hacker people, so I continue to use it in my circle.
I think what has made Matrix succeed over XMPP is the availability of decent reference clients and official servers, which allows a much lower barrier to entry at the cost of a huge amount of centralization on matrix.org and element.io. This is an acceptable compromise to me because there is absolutely zero chance of me successfully convincing anybody I know to use XMPP.
Welcome to post.lurk.org, an instance for discussions around cultural freedom, experimental, new media art, net and computational culture, and things like that.