> There is no relative difficulty. To run the client (in Windows, for > example), you run frequest.bat. If you are going to run the node first, > you run fserve.bat. In fact, it's easier to run the node first because > fserve; frequest; frequest > (that was two requests) > is less to type than > frequest -serverAddress tcp/blah:2222; frequest -serverAddress > tcp/blah:2222
I think you're overestimating the knowledge of our users. Even typing http:// is a strech ;) All I'm expecting from users is the ability to open up a web browser and type in a URL into netscape/ie. This should be our goal. > We can't take this attitude or else encryption == illegality in the eyes > of the persecutors. By encrypting everything we avoid making the paranoid > persecuted. Sorry, I'm a bit lost in the encryption deal. Isn't it transparent between nodes? We really need a single place to keep all protocol proposals and a list of ones that are going to be implemented. > Okay, I admit that people don't usually NEED to run a node. But it's > better in every way to run one. First, there is the security issue. Then > there is the issue of giving back to the network. If everyone has to run a > node then, unless they change the code to make it less friendly, they will > be acting as a datastore for the network. If we're talking about naive > users, ones that can't take the trouble to launch the node before they > launch the client, then I doubt they'll modify the source. Hopefully by 1.0 the lines between client and full fledged node will go away. Idea: Theoretically we could use the browser's cache to serve as the datastore. That would be quite interesting. > The success of Napster/Gnutella rests on the fact that everyone connected > is a datastore, even though it's possible to configure it so that no one > finds any files on your machine. If we make it easier to run a node than > to just run a client, then more people will run nodes, which is better for > the network. I think the biggest difference between us and Napster/Gnutella is that we're targeting something bigger. (At least I am ;) ) I dont see netscape opening up gnutella://<mp3> anytime soon. > So I'm not saying that people shouldn't allowed to run just a client > (there's no way to stop them anyway), I'm just saying that the premise > that most users will just use clients shouldn't be accepted. Rather, the > implementation should be changed to make it false. The fate of this question really lies with the browser makers. I see two divergent roads: 1. Majority of major browser makers embrace freenet and allow for the web browser to be a full fledged node, serving data off of cache. Of course this feature can be turned off, but is by default on (hopefully). If this happens, freenet will be big. 2. Major browser makers shun freenet for creating an open port unnecessarily and dont want to bother with this infant protocol that doesn't help sales. Then we'll have smaller, open source web browsers that can be nodes and the major browsers will at most be just clients. We'll probably have a full lynx node (official or unofficial). Netscape is a maybe. Mozilla, probably. IE is a wildcard. If this happens, the state of freenet in 3-5 years is unpredictable. Thoughts? -Larry _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
