I am new to the forum, but following the threads for a while (didn’t read the full documentation though, because I am not a computer scientist). I could find answers for most of my questions this way, but there is a complex of a few questions that still remain:
-as far as I understood, if I want to share data with another user I will perform some kind of (not safecoin) transaction, and a shared folder (or stream in case of real time communication) will pop up in this users file system. Is this the only way deduplication works? Or does deduplication also work when two people happen to upload the same file, with the same hash(es)?
-if someone wants to make data entirely public (like for instance a whistelblower) do they need to create an extra account and make the login information public, or can you share data with “all accounts in existence”?
-if people make some illegal files entirely public, could this have legal consequences for farmers in certain juristictions, since law enforcement could argue that they “could have known what they are hosting”?
You will be able to create a share (not in beta, that is basically a new ID and share only the public key (read only) or private key (full access) with friends in private.
No, they just upload to their public drive, the only difference is we normally encrypt directory entries. Public data does not, so anyone can traverse the dir tree but not alter it (as you have signed, possibly with an anonymous id).
“Could have” can be beyond any reasonable effort. If the effort is extremely significant then it’s OK, there are a ton of threads on this. Hard to conceptualise, but we are good there for sure, like saying you could have encrypted shards anywhere and create rainbow lists of all known data and if it caches in your Netscape cache then you could find it an decrypt (with effort and a lousy encrypt algo, which a lot are).
I have a notion though and want to test vaults taking a completely new id every reboot. So increasing security and also making this sort of thing even further beyond the realms of realistic (I believe it already is way past that though).
My intent on new ID’s but leaving old ones on line is a further scurity issue. If we have over 50% attack rate (we do) then all we need is more than one new user per attacker to join, if all users are really new then the attack effectively goes very close to impossible (never use that work in crypto or security, so forgive me, I mean from a specific attack I am investigating).
Means if you have a valid ID (got some rank) we keep the old id and you get a new one and a rank transfer. For ID’s with no rank we do not refresh them (so they die). This would appear to give inordinate security. I am still investigating though.
Indeed, I should have written: “could have known what they are POSSIBLY serving” instead of “could have known what they are hosting”, since a madesave node is more like a regular TOR node.
As far as I know, only TOR exit nodes can have legal problems at the moment depending on juristiction (e.g. Netherlands, Sweden <=> Germany), and not regular TOR nodes.
But just to clarify what I had in mind: let’s say the police creates a maidsafe account, monitors the peers that are directly connected (I assume it is possible to see their IP’s somehow) and then downloads some publicly available, illegal file. Even though they do not know where it was hosted (and probably it is not even mathematically possible to tell that), they could still say to the owner of this IP: you served us (part of) this file and even though you could not know at the moment, you could have known there was this chance, because it is publicly available.
I hope that not many juristiction will take this approach (as it would be close to a complete censorship of the internet, see recent quotes of your beloved PM), but I am just wondering if it is something farmers would have to worry about (maybe in even more repressive parts of the world).
Correct me if I’m wrong but I don’t believe SAFE uses IP addresses. The closest equivilant would be the PMID am I correct? And that’s private and not publically viewable. @dirvine do you mind fielding this?
Before there is something like CJDNS or BitPay FoxTrot available, I do not understand how two machines on the internet can communicate without IP addresses. I am only talking about the first hop from you to your peers. After that of course, communication becomes obscred… I believe TOR and Bittorrent and UDP and everything on the internet uses IP addresses at the moment, or am I wrong?
I think the attack is valid as described, you do know the IP address of several close nodes.
if this were illegal, an ISP would also be breaking this law, as would anyone unknowingly transporting contraband or illegal digital content (couriers, airlines, taxis, as well as any part of a computer or telephone or postal network). So legally unworkable I think. They might as well pass a law saying we can lock you up for any reason, in which case we can’t expect technology to protect us!
Also, the nodes you are connected to are randomly located, so rarely going to be in your jurusdiction. This makes this a very poor approach, not targeting a criminal, essential just trying to mass censor or mass criminalise in a rather roundabout manner.
I just finished a whiteboard session explaining clients will not connect to data retrieval groups (routing_v2), but randomly connect all over the place. The consensus attacks, group monitoring attacks and others now, would appear to be insignificant.
Even now targeting a group will prove fruitless, I promised @happybeing we would resolve the Snowden type whistle-blowers safety concerns. Even though there were extremely limited this now makes that attack even more difficult.
No I am heading now for a few days off, you really don’t want a video of me in this state I am sure the guys will update the forum whilst I am off line (heading into the hills with my tent and dog). I will still be connected fro a few days I think.