Live data from Hacker News

Show HN: An anonymous P2P social network for Android, written in Clojure

nightweb.net

61–70 of 141 posts

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#61
post #56

Great stuff but will it really scale? See secushare.org/2011-FSW-Scalability-Paranoia Did anyone try to save Nightweb's data directory encrypted in Dropbox or ownCloud? Would it be possible to add something like a reply option to a post? This would give us twitteresque chat. Would love to be able to share posts with certain persons only. PS: Can someone provide a download link for those avoiding Google Play? TIA

It would be way better if the author just provided a download link on it's page, but in the meanwhile: https://www.dropbox.com/s/g98p4nn2j33tr8t/net.nightweb.apk (version 0.05)

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#62
post #16
post #11

Wow, this is just awesome! The app is really polished for an early beta release, especially that hidden markdown feature. Just three things that come to my mind: 1. how long does it take to propagate updates? 2. what is the traffic consumption on mobile? 3. what about statusbar notifications?

Once you've bootstrapped/integrated with the network, it takes a minute or two for a new update to reach you in my testing. The initial time to bootstrap onto the network is around five minutes. The traffic consumption is pretty high, because it becomes a full peer on the I2P network. Over a month I would expect it to consume around 1GB or more. If this becomes an issue for people, I may add the ability to be a non-c…

It might make more sense to switch between the type of peer based on the type of network connection. If the phone is currently connected to wifi, then run as a contributing peer. If the phone is only on 3G/4G/LTE/etc then switch to being a non-contributing peer.

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#63
post #59
post #57

Android 4? That's a pity. But why a mobil app? Why not a .war app for the i2p router? IMO Social networks are something that are mostly used on desktop computers. Also it seems to me like you want to fork i2p. I think that's a very bad idea. First since you should be very confident in your skills if you roll your own crypto. Second because a split seems not necessary and harmful in the long run for the network. The i…

I didn't actually modify the core or router code at all, I only needed to modify the I2PSnark code (i.e. the BitTorrent component). I did not roll my own crypto at all, other than the digital signing of meta torrents which is an application-level concern. It is technically a fork because I have a separate copy of the I2P codebase in my repo, but I will always update it to the latest versions as they come out.

>but it will pay big dividends later if I need to rewrite parts of the I2P backend.

I'm all for clojure, open source, innovation and forking. But i urge you to stay part of i2p and not create incompatible releases. Maintaining your private branch will result in a lot of extra work, and I'm 100% sure that this will result unfixed security holes. The i2p project on the other hand would probably gladly accept your java patches.

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#64
post #60
post #53

Earlier quoted context omitted.

It requires version 4.0.3 (Ice Cream Sandwich) and above, so that may be the issue.

Guess it will not turn out as killer app then. This suggests other fixes, such as apk sizes http://stackoverflow.com/questions/10475954/why-does-the-goo...

It's not like it will suddenly see mainstream adoption in any case. The project is young and the early adopters will be the nerds and the geeks who don't mind the bumps. It's always like this. Organically it can spread and by then it could both be a robust app and have prerequisites pretty much anybody fulfills.

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#65
post #63
post #59

Earlier quoted context omitted.

I didn't actually modify the core or router code at all, I only needed to modify the I2PSnark code (i.e. the BitTorrent component). I did not roll my own crypto at all, other than the digital signing of meta torrents which is an application-level concern. It is technically a fork because I have a separate copy of the I2P codebase in my repo, but I will always update it to the latest versions as they come out.

>but it will pay big dividends later if I need to rewrite parts of the I2P backend. I'm all for clojure, open source, innovation and forking. But i urge you to stay part of i2p and not create incompatible releases. Maintaining your private branch will result in a lot of extra work, and I'm 100% sure that this will result unfixed security holes. The i2p project on the other hand would probably gladly accept your java…

Understood; I actually was moreso referring to rewriting I2PSnark, not I2P itself. The BitTorrent codebase is pretty old and I know I can improve it if I had time. I agree that touching the core/router code itself is dicey, but keep in mind the I2P developers themselves have encouraged people to make alternative implementations. I don't plan on that right now, though.

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#66
> the goal is to have uncensorable, untraceable communication and file-sharing on mobile devices."

This is really interesting. How is this uncensorable? From what I've read, i2p requires defined servers to resolve addresses which themselves can be taken down.

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#67

> the goal is to have uncensorable, untraceable communication and file-sharing on mobile devices." This is really interesting. How is this uncensorable? From what I've read, i2p requires defined servers to resolve addresses which themselves can be taken down.

I do not use I2P's domain name system at all. Everything is just fully-qualified base32 hostnames underneath.

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#68
post #36

Earlier quoted context omitted.

Exactly, I've been both wanting to learn Clojure and develop for Android (but no Java thanks) for a while. So here's mine http://nightweb.net/#type=user&userhash=dtvjjjuk7eai42xx...

Here's mine, for science: http://nightweb.net/#type=user&userhash=tnaohxmxj2eg3fsf...

And mine!http://nightweb.net/#type=user&userhash=trz3t72fcwwqzdi4...

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#69
post #67

> the goal is to have uncensorable, untraceable communication and file-sharing on mobile devices." This is really interesting. How is this uncensorable? From what I've read, i2p requires defined servers to resolve addresses which themselves can be taken down.

I do not use I2P's domain name system at all. Everything is just fully-qualified base32 hostnames underneath.

What do you mean by base32 hostnames? ips? If you mean ips, then, especially on a phone, wont the ip change often. If you mean some actual hostname (presumably not the reverse lookup) then who translates that hostname? If it is provided by nightlyweb servers, then if that server gets taken down, you have no more network, right?

Re: Show HN: An anonymous P2P social network for Android, written in Clojure

#70
post #45

had the same idea a while back and was pondering how I would get past NAT if you are on a mobile network. does it work there as well? I'm assuming it will continue to work as long as your carrier doesnt block the respective ports. also, don't take this personally - but if you're looking at targeting a broader crowd, you might have to update the UI to make it look more like traditional social apps.

There are IETF standardized NAT traversal mechanisms, see http://en.wikipedia.org/wiki/STUN#NAT_traversal_solutions

I don't know how widespread NAT is on mobile networks, my operator just started using it and I've been meaning to switch.. not a real internet connection anymore IMO. And lots of people are doing self-inflicted NAT on their home wifi. So it might actually be that mobile connections are less likely to be NATted.

Post reply on HN