> "1. We should use a single seednode, do an announcement through it to get > a set of initial connections, and then drop it. > > + > > Even if currently there is no script to download the seednodes > and block all the IP:port's, if we ever make any progress we have to > assume that THERE WILL BE." > > And one seednode will help how, in that case, exactly? If it's easy to block > a lot of seednodes, it's even more so to block one seednode. If you can't > anounce through it (say some dude in china) you don't get a set of > connections. > > I mean, I understand the concept, and maybe it's not a bad idea, but your > additional 'safety concern' is not really adressed by moving from dozens of > seednodes to one or two, on the contrary. If it gets blocked, it gets > blocked.
"No, there would certainly not be one central seednode. That would be much worse than the current situation. I think that quote is directed towards #2..." Maybe I didn't understand you after all, then. What did you mean with one seednode, then? Something in the lines of what the noob said? (Which, btw, I think you misunderstood). I don't think his idea was all to bad. I think he meant this: say, you have 200 seednodes, instead of putting them all up, you chose randomnly 3 seednodes (or something) out the list and put them in the download (for each download). Thus, all nodes that get installed get only 3 other nodes, from which they can get connections, but those nodes are chosen at randomn and thus are different for each node, and thus won't overload. > > > "2. We should allow the network to grow slowly by invitation, not via > slashdottings." > > That's like saying there shouldn't be hunger in the world. Well, no, there > shouldn't, but such a remark is rather wishful thinking then anything else. "Possibly." Certainly. In fact, even today, we already produce more then enough food for every person on the face of the planet, only the distribution (and the willingness to distribute) isn't there.Some will think this is a pinko remark again, but I'm not saying this to condemn it (well, maybe a bit ;-), but it just shows the nature of humans: we work best for profit, and there isn't much profit sending food to starving poor people on another continent. If the government would subsidize human aid, there would be next to nothing to distribute to those area's. As in regard to Gmail; there people perceive a profit for themselves too: an emailaccount that is fast, easy searchable, with a whopping 1 GB and for free... If you can elevate Freenet to the status that it is perceived by the public at large as being beneficial for themselves, then you might try the invite-system. Alas, IMHO, it neither works good enough yet, nor has enough mainstream recognition/hype/appeal yet, to be succesful in growing with an invite system. Our current appeal lies with people that are the-government-is-after-me-paranoids, nerds that come for the technical aspect of the project, free-speech fanatics and a handful of pedo's and chinese (etc) dissidents (no link between the two). In fact, you only have one normal, sane, intelligent and good-looking user: me. ;-) Point being: you won't get broad appeal outside the slashdot crowd untill the network is easy to use and good working, and offers content that the public wants. Seen where the succes lies in other P2P systems, we all have a good idea what content that is, exactly. Of course we wouldn't want to promote any illegal behaviour like downloading copyrighted works, but it's a simple observation that many P2P-systems are used for that, as our friend Mr.Riaa already points out. It's also an observation that our network doesn't offer that much content yet, and that it's not very good in retrieving/inserting it anyway. > Fact is, we NEED slashdot. Not only for the influx of new users, but also > for the input of new money. We all know that, so whether we like that or "So that every two years we get slashdotted, raise maybe $4K, get 5000 new users, and the network collapses, and 4,800 of those users leave, along with 200 of our existing users?" Hmm..yes, well, who's fault is it, that it's only every 2 years? I have said many times before, that we should have brought out a 0.6 version much sooner. It's beta, for christ sake, you don't have to wait untill it works perfect. Yes, yes, it would be nice if the network worked much better, as I've said myself, but it's not a matter of one every 6 months where the network hasn't improved, and one after 2 years, where it works fantastic, is it? After 2 years, the network STILL hasn't improved much. So, even if we do it every 2 years, the network still sucks, plus you have the result of being totally overwhelmed too, as you said. Every six months may cause 2000 new users, which wouldn't overload the network that much, which may cause not so many to leave, etc. > not, it's a basic fact that it has provided us with extra funds more then > once already. The moment you make a /. article, even if it's only for asking > donations, you WILL receive a sudden wave of users also. It just can't be > helped. So I think it's better to make the network cope with such an > influx/drop of nodes then to argue we shouldn't slashdot anymore. "How, exactly?" You're the one with the simulator! ;-) > Point is, > we're far from being at the point where we can ignore slashdot as a means of > getting extra recognition, users and money. > > Thinking that we can replace that with an invite-system looks > over-optimistic, certainly in the short run. I have strong doubts if it > would equal the succes of Gmail. Gmail is, ultimately, being portrayed as a > new, better webmailer: people know and use email already, and they will > always find it interesting to get something better (apart from the fact that > Google also hyped it). It's far from certain that everyone knows, uses or > needs something like Freenet (at least in their own mind), especially with > its more controversial issues. And we can't afford advertisements, exept > maybe in the form of slashdot-articles. > > Besides that, if the network worked considerably better, and maybe a bit > more content, I don't think you would have a 150% rebounce of users. "Sure but there is a simple fact of life here: Tripling the size of the network in a day is bound to cause major problems. Especially if most of the nodes then leave. Even if we can make it survive the slashdotting, it still won't meet most of the users' expectations." True. But some things may help, and others not. I have said before, that for the crowd, even the slashdotcrowd, the perception matters a lot. There are basically two things to reduce the rebounce of users: improve the perception, and lower the expectations. As for the first: well, I've already given some suggestions in the past. Create a google-like searchengine (yesyes, the js one, there I am again ;-); it doesn't have to work exactly like google, as long as it gives a familiar feeling for people and makes it easy to find something. Secondly; maybe improve the seednodes thingy before a slashdotting (such as what we've discussed above), so that they don't become overloaded as much, or maybe try to ensure of some hardcore-freenetters that they'll leave a well connected (T1-like :-) node running non-stop for 3 days during and after a slashdot. Thirdly, improve the pre-caching of links on freenet, so that users may have the impression it works faster. Forth, maybe we can send downloads that have a partially filled datastore already (just the most (legal) popular content), like the indexes and some popular flogs or pronsites - only (be)for the slashdotting, to ease the pain. Fifth: implement the..., how was it called again...TUK's? We need something better then DBR, I've said this before, and I think this one is really important too. Six : make an appeal to the regular freenetters to put some extra content in the network, prior to a slashdot-release. As for the second part; an effort should be made to diminuish the expectation of the slashdotcrowd; make it very clear, in the article itself, that it is only beta, that the network will suffer due to the slashdotting, and that they shouldn't give up to soon; that they need to let their node connected for 2 or 3 days. Maybe even put it blunt and say it will be crap at first, but that it will get better. All these suggestions could at least help ease the pain for a slashdot a bit. > > My conclusions thus, is, that in the short (and medium) run, we need and > will continue to need slashdot and the like. Speaking of which; how's the > simulation/slashdot article comming along? And, what's up with the simulator-page-/.-article? ;-) _______________________________________________ Support mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
