Abaris, you are inconsistent here:

You say "even if all bears were clustered together, that wouldn't necessarily mean that it hurts the network." So how does it hurt the network if we cluster all non-Bears in response? In fact, arguably, we are achieving nothing, since the bears are already clustered 'away' from us, there is no real need for us to cluster 'away' from them.

I think the same argument extends to limewire, but see the sixth paragraph below because I don't think I need to argue the point further.

I can certainly understand people not wanting to take part in opensourcep2p, but all this uproar and fear and worry seems unwarranted. So unregistered is nearly as rude as Vinnie - but do you guys (afisk included) go harrassing Vinnie to change his ways? I wish you would, he is a w@nker.

Another note, Xolox and Phex are not the only ones which were blasting search packets. gtk-gnutella was sending them off (optionally) pretty damn often <I>[Edit note (I think from memory - maybe I'm confused with download retries)]</I>. Is this practise really so harmful to the network? I don't think so. It has been greatly overstated and is no reason to block a whole client.

There is a simple solution to this for any client who does think it's a problem - simply do not propogate a search packet if you just saw the same search from the same client 5 seconds ago. And this should be implemented in any good client anyway, to prevent deliberate abuse of the network. These types of sanity checks are totally standard practise for any sane internet client.

No need to block the Xolox, this was just an excuse by Vinnie to block a client, if you ask me.

So we want to block him back? Is that so unusual? Look at the West Bank - go try to talk them out of their revenge killings back and forth, not us out of a few little data packets we want to drop - there are a lot more lives at stake there. I think your concern is misplaced.

Most users when given the choice to block clients of their own choosing will block nothing, or will block bearshare (at the moment). After all, the more clients a user blocks, the less clients they can search, right? So they will only block the ones they really hate.

Unless a real lot of people choose to block something, you won't even notice the effect above the normal network noise of people logging off, swamped connections, etc etc. So it is only clients who do something really extreme that will ever find themselves 'suffering' from people blocking.

As for whether splitting the network altogether is a bad thing, I have argued this point before. The number of clients on the network is around 30-50 x generally accepted horizon size of 10000 machines. If some of these machines disappear to another network, then your horizon size hasn't changed. OK, Ultrapeers are <I>meant to</I> increase the horizon by 80 - 500 times, but are the users leaving limewire users? If not, limewire won't experience a difference.

And the bigger question - would they stay - would anyone joining opensourcep2p stay - if opensourcep2p wasn't there? People who are this annoyed with BearShare and feel threatened that more BS-like behaviour is going to encompass the gnutella .. many of them are going to leave gnutella anyway, I think. We are just the movement of those people who would like to continue to work on the existing protocol and clients, instead of trying some new protocol - there are plenty out there I'm sure.

You should be happy to have our input.

Nos
<I>[Editted 12.04.02 to add note about gtk hammering possible confusion],/I.