Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

If Skype hadn’t been a thing I’d agree, but it was.

The fact that we aren’t using peer-to-peer encryption in our message apps today is solely because companies want to charge you money for sending messages.

Because Skype had the capability to operate as a decentralized encrypted messaging app, and now 10-15 years later, there isn’t an app like that. Not because it’s not possible but because giant tech companies have become the equivalent of fossil fuel companies, killing innovation by buying it and absorbing it, altering it or shutting it down.



Do you remember how skype worked in 2011? You could send messages anytime, but they would only be delivered if you and your conversation partner were both online at the same time. If you used two computers, the conversation history would sync up if both computers were online simultaneously -- or you would sometimes get it from your conversation partners if they were online.

That's fine in the world of desktops with always on connectivity (often not even NATed), but it doesn't work on iPhones were when your app isn't on screen, it isn't running (they have a limited amount of background execution these days), and anyway connectivity is spotty.


so offline messaging was such a difficult thing to solve without selling? I doubt it. Skype hasn't improved much, not only because MS is not doing its best, but also ISPs typically throttle that.


Centralization was more or less required to do offline messaging. Not so much the selling bit (but note that Skype had already been sold twice before Microsoft bought it in 2011). I agree Microsoft hasn't done a good job with it, though.


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).


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)


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.


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?


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.


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 ;)


there's plenty of people willing to accept less features for more booty - check out tinder


Now, this would be a thing I'd get excited about - if a non-profit company gets behind developing it. It was a given that whoever bought Skype would milk it for the money.

Imagine an App on every platform/device that forms a mesh amongst its users and communicates short messages in a P2P way. Perhaps some simple state could be stored in some special nodes that aid in co-ordination. The whole thing may come with a Wikipedia like openness goal. I'd donate generously for such a cause.


matrix.org


Yep. Main issue with Matrix is that Riot.im, the main client, needs a lot of UI/UX work for it be a seemless as Whatsapp or Signal.


>The fact that we aren’t using peer-to-peer encryption in our message apps today is solely because companies want to charge you money for sending messages.

Encryption export laws. And most importantly, pressures by three lettered agencies to target individual bad actors.

https://arstechnica.com/information-technology/2018/09/skype...


there are strong encryption apps though, and the bad actors are already probably using those.


Before the MS acquisition Skype was not completely free. It had subscriptions and you could buy credit to reach the PSTN, however it was not profitable. Money would eventually run out, as it happens to every company that keeps an unsustainable product alive.

Of course we as consumers would love to have everything free, but someone is footing the bill.


> buying it and absorbing it, altering it or shutting it down

and making it free, bankrupting any possible competitor




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: