Continuing the discussion from Pushing RSS to pump.io (how do I do this?):
I had an aggregated account up at https://microca.st/maiki, but that instance hasn’t had it security cert renewed since
01/06/2017 07:59 PM. So there is that.
I am probably going to write about the Micro.blog project soon (spoiler: it doesn’t get me excited), and in thinking about it I was reminded of the effort it took to get the pump.io account sharing my content, and the underwhelming results from it.
I am obviously trying out different ways to share info and have discussions, but I just don’t feel energy around federation.
I want to remain open to being wowed, but it also takes a lot of effort to maintain a server that holds communication/knowledge indefinitely.
So we’ll see…
The pump network has had terrible uptime. identi.ca is down right now. I think that has to do with it not being a priority for its maintainer rather than federation. Hopefully it’ll get handed off soon.
Indeed. But this seems like a problem for self-hosting, not specifically federation.
I tend to think people who think self-hosting and federation are cool (I’ve been and many times still am one of those people) have over-valued those things and under-valued commons-oriented institutions that can maintain things long-term in the interest of people whose stuff is hosted and working with the rest of the world (latter sometimes entails federation).
I know I conflate those two often, but it isn’t without reason. In order for federation to work, self-hosting is required, but so far federation in web-based social networks have failed to consider the traits that make other self-hosted services successful.
I’ve always wanted to write this in more of an essay form, but here are the high-level issues with how federated network services work and what people apparently expect from social networks (meaning the free systems developed, based on non-free systems):
- Federation is stateless (yeah, I know how that sounds). Social networks are designed to give a near complete index and history of interaction. We don’t expect this in email or other federated systems. And this immediately becomes an issue with self-hosting, because one can store years of email in the space of one day of social graph from a well-connected person.
- Privacy is fruit on a different tree. If we just remove the requirement of private messages, all our self-hosted systems get so much easier (and hence better). But the worse part of it is that we don’t even have privacy on the non-free systems. It isn’t feature parity, and it hurts adoption.
- We’ve learned nothing from email. We didn’t get it right at first, it took a while. But it only started to work when we had standards. Not an API. Standards. This is why a partial project based on XMPP is more interesting to me than a fully developed app that can’t talk to anything else.
I need to write out some thoughts on this in longer form, but I want to put this thought out there, in light of DreamHost’s intent to build a WordPress-based micro.blog thing: we’ve had all the ingredients to make federated systems for years. For months I had people using OStatus to follow and comment on my WordPress blog. It was amazing. And then it just kinda stopped working.
Was it complicated? Sure! It used like five plugins and a little tweaking to work. But none of those core concepts have changed, and here we are now with a new “micro.blog” thing claiming to fix the problems of Twitter and Medium by allowing people to run something not as flexible as WordPress.
To clarify my original statement, I was referring to the “federated social networks” that have come and stagnated over the last several years. I’ve been running WordPress for over a decade and it works fine. I know more than a dozen people that pay around $15 a year to run their infrequently update site, and we have a thriving ecosystem to assist with it.
And in a way these are all federated, via embeds and RSS and various other little features that make it easier to share and discover, beyond ad-driven 140-character bursts. If someone wanted to wow me, they’d take what obviously works and improve on it.
Oh yeah, I’ve been meaning to say this somewhere for a while: in my opinion the single greatest impact in recent history on web communication in general, and I think we will see in federation specifically, is the creation of Let’s Encrypt.
Once browsers start warning about insecure connections we are going to have a jabber manifesto moment. Except it will be smoother because LE makes it easy for people of all skill levels to utilize it (those just learning can rely on their hosting companies).
In the spirit of the season, I am raising this thread from the dead!
@mlinksva, do you know about https://social.coop? There was a write up on Medium. It reminded me of what you had written, so FYI.
I’ve noticed social.coop and it seems they’re doing a lot of smart things.