First and formost it is now time for Hydranode to get the existing functionality realiable from the user perspective. What is there has enough potential to make the gros of users happy. It has to be rock solid before actually doing so.

You are absolutely on the right track if you pursue the stabilisation path first and then approach the next biggest user community. But that time you will do it with a reputation already.

Better make that reputation a good one.
 

The "old" Gnutella network isn't called "G1" (only G2 users do so), it's just "Gnutella".
So, it would be cmod_gnutella instead of cmod_g1 :)

Gnutella2 is also known as "G2"
 

Instead of implementing a new protocol for more anonymity, why not try and support TOR (needs proxy support?, two birds, one stone =P) and I2P *cough* like Azureus. Besides there is no such thing as perfectly anonymous P2P.

Getting HTTP fixed, FTP working and then Gnutella implemented seems like the way to go, though =P. [However I would love support for HUP signal. =P]

As for the GUI, it would be useful (for me at least) to not have it start the core locally if it can't connect to my server =P. P2P is not looked on favorably at the work-place =P. [Even better would be decoupling the GUI from the core completely, but I don't think the security people detect an attempt to run hydranode-core.exe... Yet. If there is ever an installer that makes you install both (core+gui), that might be bad.]

As I see it, Hydranode is trying to be the best P2P program out there, but in order to draw in eMule users, you'll need to support all the features they but do it better =P. Kad and proxy support might be good additions. (Even if it seems Overnet/Kad will not add a lot of sources or greater file-type diversity, it may add a lot of users =P.)

One more point against implementing anonymous networks (for now): high anonymity networks are probably more useful for publishing files, not for general P2P file-sharing. And lets face it, Newsgroups work well for publishing =P.

That said, I can't wait to see and read about how Hydranode develops from here =P.
 

excuse me, but i dont see the point in proxysupport. it can accelerate http, but only if hispeedclients ask the same data from a lowspeedserver. in a p2pnet in most cases the clients dont ask the same data. for p2p proxies hinder the network, and tors only surplus is its anonymity. tors anonymityarchitecture isnt build for p2p, p2p needs anonymity through spreading, not through tunneling.

btw. new networks (like gnut or g2) are good as addition, but only as long as the old ones are implemented near perfection.


so long,
 

Many people don't see the point in proxy support... However some people do... Mldonkey, Shareaza, Azureus, uTorrent, Limewire, eMule, Overnet/eDonkey, even Napster, heh.

HTTP caching proxies like Squid are mostly only useful for browsing websites (or in reverse for taking load off your webservers processing), not P2P.

If you read How TOR works (http://tor.eff.org/overview.html.en) You'll see that it routes data through a bunch of servers (Outgoing-Initiated connections, requests for files, etc), however if you're using eD2K you might want to force LOWID, because the HIGHID contains your IP address =P. No doubt there are a few catches like that.

And here is a link comparing TOR, I2P, Freenet, MUTE/AntsP2P (and others):
http://www.i2p.net/how_networkcomparisons