4 ms·
I think that what they're getting at is that in the Usenet days half of what you've mentioned would be local data. There is no central concept of "content that
by throwaway22032 2y ago
I think that what they're getting at is that in the Usenet days half of what you've mentioned would be local data.
There is no central concept of "content that have favourited" or "users they are following", that's all handled locally in that model.
- evanelias 2y ago> half of what you've mentioned would be local data Not by volume. Posts and comments make up the vast majority of the storage requirements, and none of that can be purely client-side. > There is no central concept of "content that have favourited" or "users they are following" I'm aware, I used Usenet quite a bit in the 90s, as well as dial-up BBSs. Usenet is a distributed forum / discussion board, which is related but not equivalent to the core functionality of social media applications being discussed here. With Usenet's model, there's no concept of a profile aggregating content from a single user. This means you simply cannot replicate the primary experience of Facebook, Twitter, Instagram, Tumblr, Pinterest, DeviantArt, MySpace, Friendster, or any other social media site/app with Usenet's approach. Nor can it reproduce the experience of even modern forums like HN or Reddit. Usenet also didn't actually scale massively. Every estimate I've seen of the peak Usenet userbase puts it at a tiny fraction of modern social media. In any case, Usenet essentially failed. We already have empirical evidence about how these ideas play out! Why are we even seriously discussing this?