Yes, I agree. However I am concerned that I am not even seeing signs of recovery 48 hours later. The network cannot take that long to converge if it is converging at all.
The network will only converge when this becomes the most effective way for it to deliver data - while the network remains small it is better for each node to have broader expertise.
Think of Freenet as a soap-bubble. A soap bubble will always adopt whatever shape minimizes its surface area. In most cases this is a sphere. Similarly, a large Freenet will try to adopt whatever configuration will minimize retrieval times, and we have seen in pre-NGR that this generally results in pretty obvious specialization.
Don't get the causal relationship mixed up though. Soap bubbles are round because this minimizes their surface area, and pre-NGR Freenet specializes because this minimizes search time (according to its primitive measurement), but if Freenet is not specializing, or not specializing in an obvious way, then this doesn't mean its broken, merely that it has found another optimal way to minimize search times. This is to be expected since NGR Freenet has to consider network topology (latency etc) and node performance, where pre-NGR ignored both of these factors.
If Freenet isn't specializing it is probably because the network is small enough that it doesn't need to specialize yet, or, at least that it doesn't need to specialize in such an obvious manner. As the network grows it is likely that more obvious specialization will occur.
Ian.
_______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
