C'est vrai--Let's keep the knowledge flowing . . . and keep crediting those who contribute real benefits to gnutella users.

Dave: You harm yourself and your users by taking more than your share of the credit. Roger, Sam, Gregorio, Phillipe, Jens-Uwe and many others I am only dimly aware of deserve credit. You know best who to credit, so let your users know too. You corrected those who bashed gnutella, so do the same for those who bash LW.

(Arne, I know you try: do you get much credit for your contributions?)

If ACQX is really using the 4.0.2 core, we'll see how many people complain about too many connections as an UP.

Politics--bleagh! (RAZA is taking another hit in the GDF this morning--hope Phiippe's java optimization gets more attention).

Back to the topic: so--multinetworks (bad), which leads to clustering/ "Good Leaf" definition.

Should we ask for "Block Vendor" with the ability to identify false Vendor messages?

Currently LW clusters naturally, but bans itself from taking all the slots, right? LW will not allow *itself* to use all leaf slots--two will be reserved for non-LW leafs, but as many as 15 *could* be connected if I understood Sam correctly here http://groups.yahoo.com/group/the_gdf/message/19704
I agree--this is a very generous policy.

The definition of a *good* leaf is less clear to me
trap_jaw wrote a "greedy client" patch, which I gather has been adapted and works quite well. Is this enough?

I'm inclined to think the slow process of the GDF and the proven cooperation of LW to encourage other developers is sufficient.

Ask for "Block Vendor", or just let it be?