How do you expect mobile users to utilize PoW? If the answer is Delegated PoW, even that involves payment.
Discussion
That involves payment. But the PoW can still be verified.
You can’t verify payment between third parties.
Paid relays are the mobile solution, certainly, with or without additional work. However, allowing work as the alternative to paying prevents the walled garden result. Non-paying users can still be in the community, and if the required work is correctly tuned, without the spam of an open public relay.
Precisely. Anyone can still get a laptop or other thing better than a phone and do the PoW.
And they could pay someone else to the work for them when in mobile.
Of course paid relay operators aren’t going to like it as it would make their margins even thinner. But users should have the last word
Is there a real world application example of just a bit of PoW? I could spare 500ms of Mobile GPU, but maybe a Botnet could not?
I‘m spending a lot of time typing (I‘m a very slow typer). Could we utilize that with some sort of a multiPoW?
Theoretically we could hash each substring of the content with a little bit of Mobile GPU PoW while it‘s in the writing. The slower you write, the more PoW behind your message.
I imagine the nonce being a tree structure, so that you also can edit your text.
I'm not sure how to do it without permitting the work to be re-used, but I'd love to see something like this. I can imagine a client showing you a progress bar as your work builds up, letting you know you can do another like or post.
Perhaps the relays would be queried for a string to hash. I think a big part of what is missing is how do clients figure out how much work they need to do? There's no part of the NIPS that defines this. I suppose the idea is that users would filter for work, but I don't think they will. The NIPS should be updated to presume the relays are filtering for work and so they need to provide a target and they could provide the message to hash.
The relays would each require unique work from you, to post, but you could build up that work while you're just browsing, and that would be pretty cool.
Perhaps the relays could be, each one, a line in the hashed message and so long as the hashed message included a line of text provided by the relay, it can confirm the user did unique work, without requiring the user to hash each message independently. A single message could be hashed and used as verification but still only once for each relay that provided, each, a line of text to hash.
Doesn‘t this go a bit against thin relays?
There's far more messages than users and each user would only be working on one hash for each relay at a time, so I think it's within the same order of magnitude of storage space required as running a relay and storing a constant 1 additional bit of text for each user.