4 ms·

You could cache encrypted messages on a P2P network until the offline party comes online. There is no money in that, but from a technical PoV it is very doable.

You could cache encrypted messages on a P2P network until the offline party comes online. There is no money in that, but from a technical PoV it is very doable. Systems like the lightning network on Bitcoin or Raiden on Ethereum are working on something similar for outsourced channel monitoring (aka watchtowers).

toast0 8y ago
You could, but in order to make that reliable, you have to use a lot of other people's space storing other people's messages... It's a lot simpler to keep it reliable if the service owner owns this (this is also part of the reason the Skype supernodes moved to be company hosted)
- EthanHeilman 8y ago
  How much replication do you need? 10x? 20x? If these are text messages that's probably less than 10KB per person per week of being offline. Delete the messages if the recipient hasn't come online in a week. I believe you could build a highly reliable P2P offline texting app without any centralized services. I'm not sure anyone would want it and it seems difficult to monetize, but it is doable.
  - uxcolumbo 8y ago
    Messaging apps nowadays also allow you to send images and even videos. How would you handle those? Or only allow sending text when the other person is offline?
    - EthanHeilman 8y ago
      I agree that once you start adding features like images and videos that it tips the balance more toward centralization. > Or only allow sending text when the other person is offline? I think that is an excellent idea. Also I'd propose a vector-based image format with a maximum file size and fixed resolution (for online and offline messages). That way you could send drawings and diagrams without increasing the data requirements. It would also give the app a unique look and feel and would make it more difficult but not impossible for people to send hideous photographic content. The real question to me is how many people are willing to have less features for more decentralization. I suspect that is a small number of users.
      - t_akosuke 8y ago
        there's plenty of people willing to accept less features for more booty - check out tinder
      - uxcolumbo 8y ago
        Yeah I like your vector-based image idea. It would make it look unique and still be usable. I've already got some ideas about potential UX flows for this... And I think it there is a big enough community that would use this kind of messaging system. Build things that don't scale - so they say ;)

toast0 8y ago

You could, but in order to make that reliable, you have to use a lot of other people's space storing other people's messages... It's a lot simpler to keep it reliable if the service owner owns this (this is also part of the reason the Skype supernodes moved to be company hosted)

EthanHeilman 8y ago
How much replication do you need? 10x? 20x? If these are text messages that's probably less than 10KB per person per week of being offline. Delete the messages if the recipient hasn't come online in a week. I believe you could build a highly reliable P2P offline texting app without any centralized services. I'm not sure anyone would want it and it seems difficult to monetize, but it is doable.
- uxcolumbo 8y ago
  Messaging apps nowadays also allow you to send images and even videos. How would you handle those? Or only allow sending text when the other person is offline?
  - EthanHeilman 8y ago
    I agree that once you start adding features like images and videos that it tips the balance more toward centralization. > Or only allow sending text when the other person is offline? I think that is an excellent idea. Also I'd propose a vector-based image format with a maximum file size and fixed resolution (for online and offline messages). That way you could send drawings and diagrams without increasing the data requirements. It would also give the app a unique look and feel and would make it more difficult but not impossible for people to send hideous photographic content. The real question to me is how many people are willing to have less features for more decentralization. I suspect that is a small number of users.
    - t_akosuke 8y ago
      there's plenty of people willing to accept less features for more booty - check out tinder
    - uxcolumbo 8y ago
      Yeah I like your vector-based image idea. It would make it look unique and still be usable. I've already got some ideas about potential UX flows for this... And I think it there is a big enough community that would use this kind of messaging system. Build things that don't scale - so they say ;)