4 ms·
> the problem with decentralization is that, for an average Joe, it's just a bunch of loonies wearing tin foil hats who have something to hide, and besides, wha
by asdkhadsj 7y ago
> the problem with decentralization is that, for an average Joe, it's just a bunch of loonies wearing tin foil hats who have something to hide, and besides, what has Facebook done to me (average Joe) personally?
I agree, but I think there's a middle ground that fits everyone. Eg, I really like federation because I feel it's a quality middle ground _(depending on the implementation)_. It is what I am targeting my projects distributed implementation on.
Roughly, I am designing an application similar in concept to that of the ActivityPub federated protocol (and may use ActivityPub in the future). I want to support first-class self executed tooling, like git locally, while also supporting a federated multi-tenant solution to support the people who don't care.
Currently my thought is that people should be able to pick up their data and walk somewhere else with it if they like. Self hosting with federated multi-tenant, seems like a natural solution to the problem. My implementation _(goal, at least)_ doesn't "invent" anything, uses standard web tech, and supports both federated and solo usage. I'm being purposefully vague for anonymity sake on this account, fwiw - apologies for that heh, but hopefully it doesn't matter to this discussion.
> In my opinion, the worst one is the ubiquitous use of smartphones over PCs, as any decentralized stuff on smartphones is a nightmare, mostly due to systems relying on servers for stuff like push and background app refresh.
Due to the nature of what I am creating mobile matters as well. You're definitely right in that mobile is a pain here. My goal is to "first class" support mobile too, and my "protocol" would hypothetically support first class mobile, but given that storage and cpu are quite limited a client->server architecture will definitely be useful. I plan/hope to make mobile work in a hybrid manner, supporting both on-device and a federated server, as on-device will be important for offline functionality.
> Decentralization must offer some clear advantages over fb, that an average person might appreciate. Those advantages must offset the UX problems, of which there are plenty, most of them pretty hard to fix.
I agree, which is why I don't think decentralization itself will be a winning feature. Personally I don't think there's anything wrong with what Facebook is doing - what's wrong is that we can't take our data off FB and move to a platform we're more aligned with. One maybe that we pay for, or one I self host, etc.
- miki123211 7y agoWhat guarantee do you have that your thing won't turn into the "Gmail and the rest of them" problem? Basically you need money and tech skills to run a server. Small, amateur servers might have problems like much more frequent downtime, being paid or small and hard to find etc. In that situation, some big corp, perhaps Google or Facebook, could create their own server, free for everyone, but with no privacy and lots of tracking. People will migrate, and then, slowly but surely, the trap will start springing. The service will be locked down, federation will be disabled, most data export functions too, same for the API etc. The quality will slowly degrade, but this willbe the point of no return. I see that happening with Gmail already.Mail was great, but we need more. Decentrealized protocols should be based on constitutions of democratic countries, as they've learned their lesson already. It's not enough to make a country democratic now, we need to secure it against future hostile attemts at takeover. Same needs to be done for decentralized/federated protocols.
- asdkhadsj 7y ago> What guarantee do you have that your thing won't turn into the "Gmail and the rest of them" problem? Basically you need money and tech skills to run a server. Small, amateur servers might have problems like much more frequent downtime, being paid or small and hard to find etc. In that situation, some big corp, perhaps Google or Facebook, could create their own server, free for everyone, but with no privacy and lots of tracking. This is not a problem I'm trying to solve, as I have not heard/thought-of a solution I like. I also feel it is a minimal problem. Facebook is a problem in my view. Gmail, not so much. Why? Because I don't use Gmail, and I have minimal to zero issues caused by the fact that I don't use gmail. Federation seems to work well here. > In that situation, some big corp, perhaps Google or Facebook, could create their own server, free for everyone, but with no privacy and lots of tracking. People will migrate, and then, slowly but surely, the trap will start springing. I'm not sure I've seen that with something like email. Though, to be honest I'm not sure I've seen anything like email, either. It sort of happens with email in the sense that without a large email provider spam is difficult to beat. Furthermore large providers are distrusting of non-providers (ie, running your own email server), but this to me seems less of a problem with federation, and instead a problem with email itself - one I'm not sure how to solve. Spam is a difficult beast. > I see that happening with Gmail already.Mail was great, but we need more. Decentrealized protocols should be based on constitutions of democratic countries, as they've learned their lesson already. It's not enough to make a country democratic now, we need to secure it against future hostile attemts at takeover. Same needs to be done for decentralized/federated protocols. I don't understand what you mean tbh - but if your democratic countries create standards like email, federated and distributed, then I don't have any complaints. Though, to me (as a hypothetical dev) it makes little difference that you added politics on top of it - I just care that it's federated and I can own my data/etc.