Jordan Ritter (napster co-founder) wrote this, which RAH forwarded to the DBS list. http://www.monkey.org/~dugsong/mirror/gnutella.html So I'm presuming why people compare gnutella favorably to napster is because napster is flawed at a different level: it is not robust against legal attack. The very recent developments highlight the fragility of the napster model. This is not a criticism of napster or gnutella, but comparing them directly is difficult, gnutella is trying to do a much harder thing. It is typically much easier to design, and build hierarchical systems. Fully distributed algorithms and protocols with good characteristics are hard to design. But distributed systems _can_ be designed, and to some extent their performance and scalability predicted. So Gnutella has it's problems. However I wouldn't say that distributed sharing can't work, just that the first version of Gnutella wasn't very scalable. There are other systems: FreeNet, MojoNation, which perhaps are better. For one set of commentators the interesting factor is the ability of the distributed document space to be censor resistant; it's one major part of the application which generates the hype. So Napster doomed itself by being scalable enough to get enough users that it generated enough press (mostly centered around distribution of arguably copyrighted materials) to make the media industry notice it and the predictable result follows. It may be that Napster will continue in the form of the open napster clones, even if Napster the company is prevented. Anyone can start napster servers at this point. So this is the short term easier thing to do: create lots of reasonably large individually hierarchical napster clones. We know the web is reasonably censor resistant, web pages get shut down, but the content resurfaces at other sites, in different jurisdictions, etc. As long as there is an interest (and there clearly is), new servers will pop up faster than the censors can shut them down. Adam Personal opinions only, of course