Does anyone want to review my draft blog post on decentralized social networks? I'm comparing federated and p2p models, focusing on Mastodon and ssb.
Any feedback is welcome, I tried to write up an objective take on the pros and cons of each.
Good stuff, nice read! I'm going to go learn more about Scuttlebutt. :)
A couple of points that struck me as I was reading. Feel free to disregard anything you don't feel fits your writing style (or anything you just don't like, I'm not the boss of you, haha)
1. Do you want a bit of an introduction, something framing it to explain the reasons we should be interested in the discussion?
2. Is there a way to have a more clear difference between the Monetization and the Matrix sections? The header sizes are pretty similar, and at first I thought Matrix was another part of the ActivityPub section.
3. Do you want to talk about privacy and the tendency of centralized network to become centralized and focused more on monetization than user experience?
@HiRezCanDo 1. I'm targeting this at people already interested in decentralization, since all the posts that spend paragraphs reiterating why centralization is bad tend to bore me - I want to get straight into the pros and cons of specific implementations.
2. That's a good point, thanks!
@HiRezCanDo 3. Hm I'll think a bit about how to work this in, decentralized networks aren't immune to surveillance or data collection, they just haven't built business models around it.
@arcalinea If you have a preferred contact for detailed comments, please state.
Interesting, some good bits, and props on using Paul Baran's original packet network diagrams :)
The organisation overall is a bit confused, as you're switching between discussion of general principles, specific network/protocol implementations, and a few focused problem areas (moderation, identity/directory, credentials/recovery, monetisation/funding), through the course of the explainer.
@arcalinea There are some topics I'd have expected to be at least mentioned, which aren't, including:
- Network effects / graph size.
- Zipf/Power functions and affects on both registered accounts and overall activity.
- Protocols, development, and evolution or statis.
- *Hyuuuuge* rabbit hole: legal considerations / risks.
I'd approach the essay from PoV of:
- What questions am I addressing?
- Am I proposing solutions, or simply outlining the landscape?
- Is there an analytic structure?
@arcalinea You clearly can't fully address all of these, but identifying what you _do_ want to tackle, and at least _identifying_ further considerations, might be useful.
- I tried to give estimates of network sizes - a few million for activitypub, ~20k for ssb.
- As for power laws, do you mean how existing follower counts affect visibility and compounding?
- As far as I know, all the protocols I mentioned have an active community that develops and maintains them. I didn't address projects that look abandoned.
- legal considerations... I'd have to ask a friend with more law knowledge. @ross ?
@ross @dredmorbius What came to mind for me is how legal responsibility for content shifts from a company to other places in federated and p2p networks. Federated - the server admin? P2p - user's computers? If a p2p network user is distributing illegal content and that gets stored on your computer unknowingly, are you liable?
@ross Off the top of my head:
- GDPR , Calif. privacy regs.
- Child pr0n.
- Criminal activity, generally
- General liability for operations (outages, denial of service, data loss, etc.)
- Regional / local obligations & strictures, ranging from cultural to political to religious.
- Differing obligations in different jurisdictions.
- Commerce and tax / accounting obligations
- Know your Customer obligations (finance)
@ross @dredmorbius I've been following moderation discussions on Aether, a p2p Reddit, and people are like "get this nazi content off my computer! !! asdfghjkl;" Idk if it counts as being a service provider, but if you're running a p2p network sharing and serving data you don't control, yes that "race realism" pamphlet might end up being stored and shared from your computer
@arcalinea Power laws: a small number of participants account for the overwhelming majority of all activity. The 90-9-1 rule is remarkably widely held: 90% are lurkers, 9% participate occasionally (~50% of content), 1% frequently (the other 50%).
In general, *activity* counts far more than _registered users_.
I've measured this directly across several networks, most notably Google+, where the "1%" were actually the 0.2% or so:
(Expansion of my work.)
@arcalinea I've addressed legal separately. You probably don't want to do more than list issues, possibly resources (e.g., EFF, EPIC, ACLU, others -- I've a longer list).
Protocols: treating _general characteristics_, _protocols_, and _implementations_ ... at least somewhat distinctly, might be better.
Much more in my email to happening.net.
@arcalinea For discussion of platforms, protocols, characteristics, etc:
@arcalinea "Unsustainable" is maybe a harsh word for something that has been chugging along just fine for 3 years (both development on Mastodon, and multiple examples of Mastodon nodes that have been around for that long)
@Gargron Donations are funding, but not a business model, which is what I meant by "sustainable revenue". I'll change it to "sustainable business model." This space gets funded various ways, from donations to VC money, but I don't see any working business models. Being profitable businesses is what makes centralized social sites capable of pouring resources into development and growth. Not having that means alternatives may survive, but will not keep up at the same scale.
@arcalinea Reads pretty well. Might be worth at least mentioning "historic" examples (e.g. IRC, email .. maybe napster et al?)
@meejah Haha yeah there's a lot to get into, federated and p2p protocols have a long history, I could rabbithole for days... which is why I tried to stay focused on a few currently popular implementations. A historical evolution would be a good companion post, I'd be interested in tracing how centralization was a winning strategy for all the big companies we're familiar with today. IRC -> Slack, email -> Gmail, napster -> Spotify
@arcalinea Yeah .. I mean, I don't think it's that complicated (nor, unfortunately, surprising) .. but would probably still be an interesting study. "Make some popular federated thing centralized" is still probably a great VC pitch :/
(Another one for your list: Web -> Facebook .. luckily not completely successful yet. That in particular might make an interesting case, from BBSes, AOL et al to open, federated Web and back)
@arcalinea The descriptions are ok. In the blockchain post you could mention that blockchains are physically distributed but logically centralized.
@laravista I don't use Mac, but you can probably override that message in your Settings. Just double check that you downloaded a trusted version first :)
@laravista Manyverse is a mobile app for ssb which is easier to download. It's a bit slow to load anything for me at the moment though, and not great for battery, I don't think my mobile phone likes running a p2p network.
@arcalinea For the moment I do not remember the website or project name... but using the desktop client can I follow you? (I do not see your picture)
@laravista My picture and nickname might not have loaded yet, I've encountered this before. It takes awhile after you boot up the network for the backlog of posts to sync
@laravista That's probably safest, otherwise you won't be able to recover your account! I think I did that once, should do it again 🤔
How do you think, is it possible to bridge activitypub instance with ssb node, and how tricky could be that. I've made bridge to steamit blockchain, but it was pretty easy because content could be pulled from webserver API
@arcalinea nice, I thought it was a good high level overview, and appreciated the comparison between ActivityPub & SSB - I added it to my RFC where I'm collecting bluesky related stuff: https://docs.google.com/document/d/1wKQBVYfkWNTRvyBZGsNeK_Jcnrr_q2vDbxKYcmm302E/edit#
@arcalinea I am late. 😟
But I believe that's a pretty complex topic nowadays. I come from Diaspora and Matrix, and hardly ever use twitterlikes (like this).
Server run by the main developers of the project It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!