On the problem of data-sharing in adversarial environments
I've been thinking about the server-centric centralization issue with regard to manga & such (no reason why this would be a specific limit).
I can fairly easily imagine setting a NetNews system with low-retention alt.binaries subgroup & some announcement group for chapters/batches with inline magnet links & article alt.binary NZBs/references to mitigate storage costs without completely ruining retrievability.
This is already possible, why is it not the standard?
Granted, there are issues with this model that make it applicable but non-ideal for the kind of integrated experience users tend to expect nowadays, but it's hardly as if that cannot be addressed by a bespoke user-agent using those in the background.
That isn't quite as distributed as I would prefer, rather being decentralized, but it would seem to me a net improvement over the current state of affairs. Especially since it is 100% compatible with anonymizing overlay network use.
Fully-integrated p2p distributed programs would have to content with tradeoffs on generalized compatibility or reinvent the wheel(/reimplement) several of those things internally otherwise.
I can fairly easily imagine setting a NetNews system with low-retention alt.binaries subgroup & some announcement group for chapters/batches with inline magnet links & article alt.binary NZBs/references to mitigate storage costs without completely ruining retrievability.
This is already possible, why is it not the standard?
Granted, there are issues with this model that make it applicable but non-ideal for the kind of integrated experience users tend to expect nowadays, but it's hardly as if that cannot be addressed by a bespoke user-agent using those in the background.
That isn't quite as distributed as I would prefer, rather being decentralized, but it would seem to me a net improvement over the current state of affairs. Especially since it is 100% compatible with anonymizing overlay network use.
Fully-integrated p2p distributed programs would have to content with tradeoffs on generalized compatibility or reinvent the wheel(/reimplement) several of those things internally otherwise.
