~dr|z3d
@RN
@T3s|4
@T3s|4_
@eyedeekay
@not_bob_afk
@orignal
@postman
@zzz
%Liorar
%acetone
%ardu
%cumlord
%snex
+FreefallHeavens
+HowardPlayzOfAdmin
+Xeha
+f00b4r
+hk
+mareki2p
+poriori
+qend-irc2p
+r00tobo_BNC
+uop23ip
Arch
BubbRubb
Danny
DeltaOreo
Irc2PGuest44125
Irc2PGuest86233
Meow
Onn4l7h
SigSegv
Stormycloud
anontor
duck
halloy13412
maylay
nnm--
parnikkapore_
pisslord
r00tobo[2]
shaye
shiver_
simprelay
solidx66
thetia
u5657
woodwose
zer0bitz
dr|z3d
zzz: I think the estimated total floodfills on the LS debug page might be slightly off? not entirely sure about the estimayed total leasesets, either.
dr|z3d
Estimated total floodfills: 1140939 (?!)
f00b4r
Rationale for bep_0051 is that DHT indexing is possible, but requires either a lot of unique IP addresses, or polluting routing tables of other nodes. Polluting tables is impossible, because I2P DHT has "Secure Node ID Requirement", but since I2P addresses are basically just cryptographic keys, anyone can create a lot of unique addresses. Moreover, to be capable to create more
f00b4r
identities on a singe machine, one can (and is incentivized to) disable transit traffic and reduce number of hops, which is bad for the network. Two possible solutions I see are either to make DHT indexing extremely difficult or remarkably simple.
f00b4r
Reasons I prefer the last approach are:
f00b4r
1. DHT searches would reduce reliance on trackers for locating content
f00b4r
2. They would also lead to a build-up of larger swarms, increasing availability and download speed
f00b4r
3. Private filesharing in I2P is already possible and simple enough
f00b4r
cc: zzz, dr|z3d
dr|z3d
I think we had a discussion about DHT searches a while ago, don't remember the exact details, but it's not a bad idea in principle, if you exclude all the potentially nasty content out there.
f00b4r
I think, it's not possible to filter content in decentralized manner on KRPC level. IMO, that should be solved by a search engine moderator
dr|z3d
I was thinking more inline search in I2PSnark.
f00b4r
In theory we can include hashes of nasty infohashes and/or regexps for filtering in snark
dr|z3d
yeah, sure, and just as easily someone can compile snark without them, which is fundamentally the issue. discovery. some stuff doesn't want to be discovered.
f00b4r
I'm not sure if I understand the issue correctly. Isn't the whole point of i2p censorship resistance?
f00b4r
I mean, even now one can create a tracker with nasty content and other one can load it using snark
dr|z3d
it's a bit more nuanced than censorship resistance.
dr|z3d
but let's be clear. Tor has a reputation for hosting the kind of content we'd rather not have here. sure, it's about, but we'd rather not make it easier to find than it already is.
dr|z3d
it's a similar discussion to the one we're having about destination blacklists. maybe the two have some crossover.