Timestamps are in GMT/BST.
[1:17] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) Quit ("\(^_^)/' No Soliciting!!! Unless you have legs way, way up and really, really big tits....")
[1:32] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) has joined #freenet
[1:39] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) Quit ("\(^_^)/' No Soliciting!!! Unless you have legs way, way up and really, really big tits....")
[1:39] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) has joined #freenet
[2:10] <KenMan> crap, i just realized i have a problem with feedback in my sim:
[2:10] <KenMan> as the set of globally stored (retained) keys begins to "specialize" , so does my selection of search keys !!
[2:12] <KenMan> somehow, i may need to ensure that the selection of search keys is 'uniformly distributed'
[2:13] <KenMan> Select a random value (uniformly), and use the closest existing key for the next search? hmm, that will get expensive :(
[2:15] <KenMan> But it should suffice to minimize the feedback effect, below detection levels... making it insignificant.
[2:16] <KenMan> "the act of measuring something changes the observed level" - why does that situation come to mind ?!
[2:16] <KenMan> I mean, I am only changing something in *response* to an observation ;)
[2:18] <KenMan> oh well, at least i'm trying to be objective/scientific about this...
[2:19] <KenMan> this is one of the reasons why I felt toad should stick with his chosen method of selecting keys to search for.
[2:20] <KenMan> it is very difficult to call his method unfair or biased
[2:23] <KenMan> just trying to get this one aspect of the model correct is difficult ! A gaussian distribution that favors the newest keys may be reasonable, but I am unsure.
[2:24] <KenMan> In real life, there will be popular keys which are searched for (and thus retained) over a long time period. Other keys will be highly unpopular, and they will drop quickly.
[2:29] <KenMan> perhaps each unique key should have a 'popularity coefficient' assigned ? and on that thought, zzzz
[2:59] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[3:13] * Ash-Fox (HAL-9000@acy206.neoplus.adsl.tpnet.pl) has joined #FreeNET
[3:50] * nextgens (~nextgens@d80-170-117-50.cust.tele2.fr) has joined #freenet
[3:52] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[4:14] * TLF (~francisco@90.Red-81-40-113.pooles.rima-tde.net) has joined #freenet
[5:14] * TLF (~francisco@90.Red-81-40-113.pooles.rima-tde.net) Quit ("Adi?s a todos y a todas, y que les vaya bien. || http://www.vitaliano.esp.cc")
[5:24] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[5:55] * Ash-Fox (HAL-9000@acy206.neoplus.adsl.tpnet.pl) Quit (Read error: 104 (Connection reset by peer))
[6:03] * moskau23 (~Miranda@dsl-213-023-248-247.arcor-ip.net) has joined #freenet
[6:06] * sdogi_ (~java@84-50-18-9-dsl.kjj.estpak.ee) Quit (Read error: 104 (Connection reset by peer))
[6:07] * sdogi (~java@84-50-17-231-dsl.kjj.estpak.ee) has joined #freenet
[6:08] * linagee_ is now known as linagee
[7:30] * Elly|afk (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Read error: 54 (Connection reset by peer))
[7:30] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[7:49] * TLF (~francisco@132.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[7:52] * kevloral (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) has joined #freenet
[7:53] <kevloral> g'afternoon all
[7:55] * moskau23 (~Miranda@dsl-213-023-248-247.arcor-ip.net) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[8:55] * Xenu (anonymous@h6210017003.dsl.speedlinq.nl) has joined #freenet
[8:55] <Xenu> hi all
[8:55] <Xenu> where do i go if i think I have this great new p2p idea, and wanna find out if it hasn't been tried yet?
[9:07] * Ash-Fox (HAL-9000@acy206.neoplus.adsl.tpnet.pl) has joined #FreeNET
[9:12] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[9:17] <TheSeeker> http://www.zeropaid.com/ ?
[9:31] <hobx_> just tell us about it
[9:31] <hobx_> we are honest people who never steal ideas.
[9:31] <Xenu> ;-)
[9:32] <Xenu> I'm working on a white paper, describing a freenet-like anonymous p2p network
[9:33] <Xenu> I'd like it to have redundant AND anonymous file storage
[9:33] <Xenu> what it comes down to is this:
[9:34] <Xenu> you could split a file over several servers
[9:34] <Xenu> suppose you have 8 servers on the network
[9:34] <Xenu> server one stores the file, excluding the first bit of every byte
[9:35] <Xenu> server two stores the same file, excluding the second bit of every byte
[9:35] <Xenu> etc
[9:35] <Xenu> that way, no server will contain a full copy of the data
[9:35] <Xenu> (which might be illegal in some countries)
[9:36] <Xenu> while any 2 servers together would be able to reconstruct the file
[9:36] <Xenu> it's a bit along the lines of RAID5 redundancy, only, well, much more redundant
[9:37] <Xenu> let's call it RAID11
[9:37] <Xenu> or sod
[9:37] <Xenu> so
[9:37] <Xenu> what do you say
[9:39] <Sugadude> A bit *too* redundant, isn't it? I also can't see how this is anonymous.
[9:40] * sdogi (~java@84-50-17-231-dsl.kjj.estpak.ee) Quit (Read error: 104 (Connection reset by peer))
[9:40] <Xenu> indeed, anonymity doesn't have much to do with it
[9:41] <Sugadude> <Xenu> I'd like it to have redundant AND anonymous file storage
[9:41] * sdogi (~java@84-50-21-210-dsl.kjj.estpak.ee) has joined #freenet
[9:41] <Xenu> but it has the advantage, like freenet, that a node owner cannot be legally held responsible for having certain files on disk
[9:42] <Xenu> anonymity means that you can insert files anonymously.
[9:42] <Xenu> storage itself cannot be anonymous
[9:42] <Xenu> anyway, that's not really the point
[9:43] <Xenu> I wonder if this kind of redundancy has been tried out yet.
[9:43] <Sugadude> How do you insert them anonymously?
[9:44] <Xenu> this should be the same as with freenet: there's no TRUE anonymity, but when a node recieves a file from some other node, it should be impossible to see if the originating node is the original inserter of the file, or that the originating node is merely relaying the file in response to some request
[9:45] <Xenu> so, file transmission shouldn't be p2p, as with gnutella/kazaa/etcetera, but p2p2p2p2p2p2p
[9:45] <Xenu> as with freenet
[9:45] <Xenu> anyway, back to redundancy
[9:46] <Sugadude> So your proposal is just a redundancy scheme, this can be implemented into Freenet and used instead of FEC.
[9:46] <Xenu> possibly
[9:46] <Xenu> would be great
[9:46] <Sugadude> (And the reason doesn't work isn't because it's not redundant, but because it's routing is nearly random)
[9:46] <Xenu> but the proposal will concern more than this
[9:49] <Xenu> sorry, i don't understand "(And the reason doesn't work isn't because it's not redundant, but because it's routing is nearly random"
[9:50] <Sugadude> Freenet sucks. The reason it sucks is because it's routing sucks. The reason isn't because Freenet isn't redundant enough. So if you implement a Freenet with your redundancy it'll be as sucky as the current Freenet.
[9:51] <Xenu> I wouldn't say freenet sucks. But I like Mute's routing scheme better than freenet's. Well, I'm a biologist, so i like dynamic systems based on natural phenomena
[9:52] <Xenu> freenet's goal was never performance
[9:52] <Xenu> and admittedly, this non existing goal has very much NOT been achieverd
[9:52] <Xenu> ;-)
[9:54] <Sugadude> MUTE's routing scheme, although efficient, has never proven it's ability to cope with large networks. And the prediction is that it will only allow a node to see the nodes close to it, not the whole network.
[9:54] <Xenu> I'm working on a new routing scheme
[9:55] <Xenu> but, unlike the redundancy scheme i just mentioned, it's something I'll keep to myself for now
[9:55] <Xenu> which is hard, because i find it quite ingenious.
[9:55] <Xenu> ;-(
[9:56] * TLF (~francisco@132.Red-81-40-116.pooles.rima-tde.net) Quit (Read error: 104 (Connection reset by peer))
[9:57] <Sugadude> OK, maybe you've found the holy grail of routing algorithms. Or maybe it's something that's been proven to be horse-shite. You never know. :)
[9:58] <Xenu> no. when I've finished writing the paper, it'll have to be scrutinized by some clever brains
[10:03] <Xenu> anyway, i'm gonna do the groceries and employ other profane activities
[10:03] <Xenu> later
[10:07] * kevloral is now known as kevlAway
[10:31] * guido^pe (~unknown@dsl-213-023-233-226.arcor-ip.net) has joined #freenet
[10:49] * Xenu (anonymous@h6210017003.dsl.speedlinq.nl) Quit (Read error: 110 (Connection timed out))
[11:04] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[11:14] * Elly is now known as Elly|afk
[11:20] * kevlAway is now known as kevloral
[11:50] * orange_ (~orange@pm2-24.bahnhof.se) has joined #freenet
[11:57] * oierw (~mathew@220.255.10.127) has joined #freenet
[12:04] * orange_ (~orange@pm2-24.bahnhof.se) Quit ()
[12:16] * moskau23 (~Miranda@dsl-213-023-252-172.arcor-ip.net) has joined #freenet
[12:22] * moskau23 (~Miranda@dsl-213-023-252-172.arcor-ip.net) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[12:31] * oierw (~mathew@220.255.10.127) Quit (Read error: 110 (Connection timed out))
[12:32] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) has joined #freenet
[12:38] * TLF (~francisco@132.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[12:41] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) Quit (Read error: 60 (Operation timed out))
[12:54] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[13:20] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[13:21] * hugh (goldfinger@12-221-93-37.client.insightBB.com) has joined #freenet
[13:23] <hugh> Its impossible to install freenet.jar considering its downloading as .2kB/s on my otherwise normal cable connection.
[13:23] <hugh> actually to install freenet because freenet.jar is only going .2kB/s
[13:23] <KenMan> yes, sourceforge (who hosts the files) has become quite slow of late ...
[13:28] <hugh> hmm, well I just skipped the freenet.jar and used the one from the linux release and it seems like it'll work
[13:28] * hugh (goldfinger@12-221-93-37.client.insightBB.com) Quit ()
[13:39] * Ash-Fox (HAL-9000@acy206.neoplus.adsl.tpnet.pl) Quit (Nick collision from services.)
[13:39] * Ash-Fox (HAL-9000@ada87.neoplus.adsl.tpnet.pl) has joined #FreeNET
[14:57] <toad_> <hugh> hmm, well I just skipped the freenet.jar and used the one from the linux release and it seems like it'll work - we should get our own hosting! he's probably using completely the wrong jar...
[15:01] * sdogi_ (~java@84-50-20-241-dsl.kjj.estpak.ee) has joined #freenet
[15:11] * moskau23 (~Miranda@dsl-213-023-252-172.arcor-ip.net) has joined #freenet
[15:11] <hobx> the wrong jar!
[15:11] <hobx> but there could be anything in there!
[15:13] <kevloral> yes. Publishing sha1s of the jars would be nice.
[15:14] * sdogi (~java@84-50-21-210-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[15:37] * TLF (~francisco@132.Red-81-40-116.pooles.rima-tde.net) Quit (Read error: 104 (Connection reset by peer))
[15:45] <kevloral> g'night all
[15:45] * kevloral is now known as kevlAway
[15:52] <TheSeeker> subbit new jars to bitzi using bit collider?
[15:53] <TheSeeker> *submit
[15:53] <TheSeeker> http://bitzi.com/bitcollider/
[15:58] * TLF (~francisco@132.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[16:00] * kevlAway (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) Quit (Read error: 54 (Connection reset by peer))
[16:01] <TheSeeker> hmm, interesting. I've been running Frost for a while doing an insert of a few dozen small files, and while I have more conenctions now than when I started, I no longer have any incoming connections.
[16:02] * TLF (~francisco@132.Red-81-40-116.pooles.rima-tde.net) Quit (Client Quit)
[16:02] <TheSeeker> will it hurt much if I simply restart my node? (will others see it as a disconnect if I'm available again within a minute?)
[16:27] * nextgens (~nextgens@d80-170-117-50.cust.tele2.fr) Quit ("bbl")
[16:34] <KenMan> TheSeeker: I think some of them will attempt to reconnect, but you will lose some of them. And gain some new ones.
[16:36] <TheSeeker> I have a gig of RAM, how much should I give Freenet? it's currently using 95 of the 128 am giving it, but has allocated it all...
[16:37] <TheSeeker> I have a 6 gig datastore (full) and am limiting upload speed to 20 KB/s I suppose I could go to 30 without affecting my download speed ...
[16:38] <TheSeeker> The only other modificatiosn I've made to default are to give it 100 connections (upped rtmaxnodes to 100 as well) and 200 threads...
[16:39] <TheSeeker> rtmaxrefs is 50, but I don't know what that does, and the onyl reference I saw for it said something like "if you don't know what it is you shouldn't be messing with it"
[16:43] <TheSeeker> I was thinking of upping the rtMaxNodes sicne currently I have: Connections open (Inbound/Outbound/Limit) 91 (0/91/100)
[16:44] <TheSeeker> but don't know what effect that might have on things, so I'm hoping to get some insight from people that actually understand the code somewhat...
[16:49] <TheSeeker> anyone know if restarting a node will kill a Frost message upload?
[16:52] * moskau23 (~Miranda@dsl-213-023-252-172.arcor-ip.net) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[17:00] * Adylas (~Adylas@terra.infomtl.qc.ca) has joined #freenet
[17:05] <TheSeeker> hmm... interesting or not? http://theseeker.bounceme.net:8888/servlet/nodestatus/ds_histogram_detail.txt
[17:09] <TheSeeker> hmm, ok... so with rate limiting, the node refuses incomming conenctinos, but not outgoing ones... ?
[17:10] <TheSeeker> Pooled threads running jobs: 40 (20%) [Rejecting incoming connections and requests!]
[17:10] <TheSeeker> Pooled threads which are idle: 20
[17:10] <TheSeeker> It's normal for the node to sometimes reject connections or requests for a limited period. If you're seeing rejections continuously the node is overloaded or something is wrong (i.e. a bug). Current estimated load for rate limiting: 544.2%.
[17:11] <TheSeeker> which would be ... messageSendTimeRequest
[17:15] <Adylas> Hello, There is a way to "Hot modify" the outbound speed (upload) ? Whitout restarting the node ?
[17:15] <TheSeeker> Last I heard: no
[17:15] <TheSeeker> yesterday morning...
[17:15] <TheSeeker> [08:14.34] <skimpIzu> its possible to change the max upload rate in an live freenet without restarting
[17:15] <TheSeeker> [08:15.38] <d-ArkAngel> nope
[17:18] <Adylas> Hehe
[17:18] <Adylas> Great reporting ..
[17:18] <Adylas> Thanks
[17:19] * Adylas (~Adylas@terra.infomtl.qc.ca) Quit ("Leaving")
[17:29] <TheSeeker> someone on the network detected my address as one that belongs to mail.teitel.net o.O
[17:43] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[17:50] * toad_ (toad@toad-with-underline.active.supporter.pdpc) Quit (Remote closed the connection)
[18:02] * linagee (~linagee@netblock-66-245-227-35.dslextreme.com) Quit (Read error: 110 (Connection timed out))
[18:05] <TheSeeker> Connections open (Inbound/Outbound/Limit) 100 (0/100/100) :(
[18:07] * Sugadude (~Sugadude@ERR.COS.CS.CMU.EDU) Quit (Remote closed the connection)
[18:08] <TheSeeker> ok, just upped my connections and rtmaxnodes to 200, and restarted...
[18:08] <TheSeeker> Connections open (Inbound/Outbound/Limit) 69 (51/18/200)
[18:10] * Sugadude (~Sugadude@pakastelohi.cypherpunks.to) has joined #freenet
[18:13] <TheSeeker> just restarted again to give freenet 256M instead of 'default' ... back to 86 nodes connected... restarts dont' seem to have much of an effect on integration with the network (seen as a hickup by other nodes more than anything?)
[18:13] <TheSeeker> 86 (58/28/200) vs 100 (0/100/100) before the restarts...
[18:33] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) Quit ("\(^_^)/' No Soliciting!!! Unless you have legs way, way up and really, really big tits....")
[18:36] * linagee (~linagee@netblock-66-245-229-151.dslextreme.com) has joined #freenet
[18:51] * MrNaughty (MrNaughty@d199-126-25-30.abhsia.telus.net) has joined #freenet
[19:03] * skimpIzu_ (~albo@pcp0010469677pcs.sntafe01.nm.comcast.net) has joined #freenet
[19:09] * skimpIzu (~albo@pcp05055555pcs.sntafe01.nm.comcast.net) Quit (Connection timed out)
[19:18] <KenMan> TheSeeker: i think your datastore is uninteresting, but I'm not an expert. I looked at the 256 bucket version and the 16 bucket version.
[19:19] <KenMan> The highest peak for 16 buckets is only 1.12 times average. I saw a 1.6 for 256 buckets, which is sort of interesting.
[19:19] <TheSeeker> I seem to be 'specialized' around 4 areas, but have no other nodes to compare to...
[19:20] <TheSeeker> hmm, wait... the graph now is nothing like it was before... hmmm
[19:20] <KenMan> we don't know what peak levels indicate a 'strong' specialization, but usually, my rule of thumb was anything over 1.4 in the 16 bucket version.
[19:21] <KenMan> yeah, that is another problem (or, according to some, a good thing)
[19:21] * skimpIzu_ (~albo@pcp0010469677pcs.sntafe01.nm.comcast.net) Quit (Read error: 110 (Connection timed out))
[19:21] <KenMan> how many keys are in your datastore ? around 12000 ?
[19:21] <TheSeeker> in the 256 bucket version before, it was mostly flat with 4 of them being 5-7 over median
[19:22] <TheSeeker> 15685
[19:22] <TheSeeker> http://theseeker.bounceme.net:8888/servlet/nodestatus/ds_size_histogram.txt
[19:22] <KenMan> oh, people have observed strong specialization during the initial hours/days after startup, but usually it quickly fades... :(
[19:22] <TheSeeker> ah, hmmm
[19:22] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit (Client Quit)
[19:23] <TheSeeker> well, my datastore has been full for months now, so I thought that specialization of the keys in it took a while...
[19:23] <KenMan> yeah, the breakdown by keysize is normal. Should be very similar to that for all nodes.
[19:23] * skimpIzu (~albo@pcp0010469677pcs.sntafe01.nm.comcast.net) has joined #freenet
[19:23] <KenMan> that is a logical thought. You need to know how many new keys are being injected per hour / day.
[19:23] <TheSeeker> I think I remember Toad saying he would like a single small chunk size ... not sure how much I like that idea for large files though...
[19:24] <TheSeeker> is there some diagnostic thing that would tell me?
[19:24] <KenMan> well, small constant sized keys is predicted to offer some good benefits
[19:25] <KenMan> you would need to count successful fetches (from further routing) and successful inserts to your node.
[19:25] <TheSeeker> http://theseeker.bounceme.net:8888/servlet/nodestatus/diagnostics/graphs?image-type=image/bmp&period=hour&type=1&var=pcacheAccepted <- like that?
[19:27] <TheSeeker> I wonder if it's because I restarted my node? though I would think that those histograms are purely based off of actual ds contents, and so should be the same after a restart, yes?
[19:28] <KenMan> i think you found an appropriate diag there. So count how many keys were done over a 24hr period and you can begin to measure the turnover rate of your datastore...
[19:28] <KenMan> yeah, the diags data should be retained after a restart (it is stored in the stats subdirectory)
[19:29] <TheSeeker> I had onyl been running for 5 or 6 hours when I observed that... but it may have something to do with Frost... (also running the whole time)
[19:29] <KenMan> i changed 'hour' to 'day' in that last URL, and got a weird shape (not sure I trust it)
[19:29] <TheSeeker> I just don't see how that many keys could have appeared or dissapeared so quickly...
[19:30] <KenMan> oh, wait. You had the right graph to begin with.
[19:30] <TheSeeker> I haven't been running Freenet much lately
[19:30] <KenMan> per day, you had a maximum of 1420 new keys in a 24 hour period.
[19:30] <KenMan> probably about 11/24/04
[19:31] <KenMan> okay, in the hourly graph, you inhaled 694 new keys. That astounds me (it sounds a little high)
[19:31] <KenMan> 694 new keys over a 1 hour period !
[19:32] <TheSeeker> probably due to Frost.
[19:32] <KenMan> quite likely. I can believe that.
[19:32] <TheSeeker> I haven't run that program in over a year, and re-installed today ...
[19:33] <TheSeeker> there's a LOT of keys on Frost I wouldn't be very likely to have since I wasn't requesting them... and have been running on Unstable up until 5100 came out, so wouldn't be likely to have routed those keys either...
[19:34] <KenMan> if you are curious to watch DS specialization, shut down your node, rename 'store' to 'fullstore' , set DS size to 256MB and restart.
[19:34] <KenMan> A small store can shift quicker, assuming a fairly constant rate of new key injection
[19:34] <KenMan> ie. the histograms will dance more.
[19:34] <KenMan> also, it should be easier to achive clear peaks.
[19:35] <KenMan> But this is just for curiousity. No one denies that providing more store is a good thing.
[19:35] <TheSeeker> I'm still confused about what I saw earlier though...
[19:35] <KenMan> you had bigger peaks in the histogram, right ? and now they are smaller ?
[19:36] <TheSeeker> in the 256 bucket view there were two peaks over 5.0 and 2 peaks over 7.0 while the rest of it was pretty flat (no other peaks above 1.5)
[19:37] <TheSeeker> I wish there was a historical histogram view :/
[19:37] <KenMan> wow. Yeah, it seems announcements (which mainly go out in the early mins/hours) can create a strong spec. But then it fades :(
[19:38] <KenMan> announcements cause multiple nodes to pull on a shared peer with the same estimated spec
[19:39] <TheSeeker> how could that affect the ds the way I had described?
[19:39] <KenMan> i forgot your description :( you mean, how it had many new keys at one period and then not so many during a different period?
[19:40] <KenMan> or, the stuff related to frost ?
[19:41] <KenMan> i have nothing on which to base this, but i suspect frost generates a very large percentage of all requests in the network.
[19:41] <TheSeeker> well, the amount of data in my store hasn't changed since then (6 gigs constant use) so I'm finding it hard to figure out where all those extra keys at those 4 specific specializations came from... and then went to...
[19:41] <KenMan> so whether you use frost or not, the majority of the keys swimming around are due to frost. At least, that's my theory of the year.
[19:42] <KenMan> you have a limit of 6 gigs, right ?
[19:42] <TheSeeker> looking at the complimentray "keys rempved from the ds
[19:42] <KenMan> there is a category of key that 'expires' and thus will go away on its own. I think.
[19:42] <TheSeeker> " chart, I see that only a few hundred keys have left
[19:42] <KenMan> well, many small keys can replace one big one, and vice-versa.
[19:43] <TheSeeker> http://127.0.0.1:8888/servlet/nodestatus/diagnostics/graphs?image-type=image/bmp&period=hour&type=1&var=pcacheRejected
[19:43] <TheSeeker> yes, but the datastore histogram is the number of keys, not the amount of space taken up by them
[19:45] <TheSeeker> I knew I should have taken a screenshot :/
[19:46] <KenMan> if your node specializes, that means it will traffic in keys that occur in a more concentrated area(s) of keyspace.
[19:46] <KenMan> so, more of the keys coming in to be stored will cluster around tight spec-points. And keys outside of those areas will drop out over time.
[19:46] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[19:46] <KenMan> the second one takes longer than the first
[19:47] <KenMan> so someone decided it would be nice to insert 'b' keys to you, and use you to search for 'b' keys.
[19:47] <KenMan> Which drove up your 'b' count for a while. Then that peer went away or stopped trying to focus you on 'b'
[19:48] <KenMan> in fact, it is likely that several nodes (due to announcements) obtained a common spec point for you.
[19:49] <TheSeeker> I still thinkk it was a bug.
[19:49] <KenMan> But then their use of you dropped over time, so the effect faded.
[19:49] <KenMan> sure, a bug, or code doing what it was meant to. Either way, it didn't last :(
[19:49] <TheSeeker> no, I meant a display bug
[19:50] <KenMan> that's possible, but very unlikely for the histograms
[19:50] <TheSeeker> I no longer believe my datastore ever had such data in it.
[19:50] <KenMan> it is hard to know what to trust, because we know a few of the stats demonstrate acct'ing bugs.
[19:50] <TheSeeker> it would have required many thousands of keys to have created spikes like that...
[19:53] <TheSeeker> if the count/mean is 6, that means that area has 6 times the average number of keys in it, correct?
[19:53] <KenMan> divide 16K keys over 256 buckets... 64 keys. It would only take about 64*7=448 keys to reach a peak of 8 (7 ABOVE average)
[19:54] <KenMan> that is still approaching your 'thousands of keys' level though.
[19:54] <KenMan> but if 4 nodes did it, that would only require 448/4 keys per node.
[19:55] <TheSeeker> hmm
[19:56] <TheSeeker> currently my scaling factor is .59xxx ... I believe my scaling factor then was .13xxx that could be used to determine the largest peak...
[19:56] <TheSeeker> 492 ...
[19:57] <TheSeeker> pretty close to your 488
[19:57] <KenMan> well, depending on how narrow a specialization is, it could spill over into neighboring buckets.
[19:58] <TheSeeker> it didn't (in the 256 bucket view)
[19:58] <KenMan> We've never talked about specialization with this level of specificity, i don't believe.
[19:58] <KenMan> people just went looking for 'mountains' and 'peaks' and 'tits' and stuff...
[19:59] <KenMan> and if they saw some, they got excited.
[20:00] <TheSeeker> that's why I was excited about this... it was just flat, with four huge single bucket spikes.... (mountainous in the 16 bucket view though, with peaks at the four points)
[20:01] <TheSeeker> usually when I see a mountainous range on the 16 bucket view, inspection in the 256 shows what it does now, just a bunch of random goop with rarely any mean/count over 2
[20:01] <KenMan> oh, there is another explanation for that, I can't do it justice. But basically, your node 'invites' other nodes to search for some invented candidate keys on which to search.
[20:02] <KenMan> So your node cooked up a key starting with A7191B... and told lots of other nodes about it.
[20:02] <KenMan> No, that wouldn't do this to the datastore, ... or could it ?
[20:02] <KenMan> A single key value wouldn't do what you describe. If you are looking at DS histogram, and not request histogram.
[20:03] <TheSeeker> that's just one key isn't it? this would require on the order of 350 keys
[20:03] <KenMan> yeah. There is something else going on. Fishy smelling.
[20:03] <TheSeeker> they were more or less equally spread out...
[20:03] <KenMan> we may have problems with the estimation algorithm, narrowing in too quickly or something.
[20:04] <TheSeeker> 06 4something 9c bsomething
[20:05] <TheSeeker> or was it d? well, not much use now :P
[20:05] <KenMan> I think you should expect something a little wider than a single column shoot. The neighbor buckets should form a 'mountain' , rather than the male phallic condition you describe.
[20:05] <TheSeeker> if it happens again I'll take screenshots...
[20:06] <TheSeeker> yeah, I know, nice bell curve, not what I was seeing...
[20:07] <KenMan> well, there are bell curves and there are bell curves. Some are skinny and attractive, others are wide, short, and repulsive.
[20:07] <TheSeeker> hmm, or should it be an exponential rising spike?
[20:07] <KenMan> that HAS been observed in the past.
[20:07] <KenMan> should it be that, i don't know :(
[20:08] <TheSeeker> an exponentially rising spike centered around one point int he DS is what I would expect from perfect specialization....
[20:08] <TheSeeker> I may be thinking too DHTishly though...
[20:09] <KenMan> we've never come to an agreement that there should be one or many spikes for a single node.
[20:09] <KenMan> I modelled FreeNet as a DHT, just what you described, and it generally panned out.
[20:10] <KenMan> I mean, the one catch22 is that the nodes simply told each other what their singular spec point was.
[20:10] <KenMan> But routing was able to function, and success was comparable to non-DHT/multi-spec-point routing.
[20:10] <TheSeeker> Unless a node routed each area of specialization discreetly among nodes which thought that was the 'only' area of specialization, that would be fine... otherwise, I'd say just have a single point.
[20:11] <KenMan> well, i think we all agree that the majority of a node's peers need to converge on a common estimation of the central node's spec, in any case (single point or not).
[20:11] <TheSeeker> I think a node should start off with a spec determined by some random ly entered entropy thing (like IIP gateway startup)
[20:12] <TheSeeker> the node then knows where it's specialized, and can tell anyone else that it connects to.
[20:12] <KenMan> i think nodes should just be able to tell one another what their spec is. Censorship be damned. Well, not damned. Rather, addressed through smart redundancy.
[20:13] <KenMan> the counter-argument says "if a node is allowed to choose his spec, he can censor a specific key within his spec"
[20:13] <KenMan> My argument is that nodes can choose their spec anyway, like it or not.
[20:13] <TheSeeker> I dont' see how telling a node your specialization is any sort of privacy concern, considering that all illegal content has an equal chance of being made up in part with a key stat starts anywhere in the keyspace.
[20:14] <KenMan> statistically speaking, I will just favor storing and routing those keys near my chosen censor-attack point.
[20:14] <KenMan> so that argument doesn't hold much weight with me. But...
[20:14] <KenMan> I don't think we will see nodes outright telling each other their specs anytime soon.
[20:15] <TheSeeker> you mean, like if you wanted to get someone arrested for child porn, pad a bunch of content until they all have keys that would be sotred in their cache?
[20:16] <KenMan> i wasn't thinking along those lines at all...
[20:16] <KenMan> sure, if you know a node's spec, you can place keys there. But so can anyone else.
[20:17] <KenMan> I mean, if you have a reasonable estimate of a node's spec, you WILL place keys there if you route through them.
[20:17] <TheSeeker> court argument: "Freenet only stored information that you request in your own cache if it is near your specialization, but when we examined the contents of the cache, a disproportionaly large amount of data (as examined in other nodes) was found to be of [specific illegal] nature"
[20:17] <KenMan> no, that doesn't make any sense. If you want to set someone up, I guess there are lots of ways to do it.
[20:18] <KenMan> Hopefully the FreeNet fund will be big when the first court case is brought. It would help to establish the scientific principles correctly the first time around.
[20:18] <TheSeeker> it's a lot easier to ruin someone's day by killing them than trying to bend Freenet AND the legal system to your whims...
[20:18] <KenMan> Or EFF , or whomever might care about this matter.
[20:19] <KenMan> Yeah, killing people is very effective.
[20:19] <TheSeeker> ACLU would defend the case, I'd bet my portfolio on it.
[20:19] <TheSeeker> (assuming american legal suit)
[20:20] <KenMan> Well, I don't have any wish to be the first 'rabbit' or 'guinea pig' or whatever...
[20:21] <TheSeeker> the first Freenet suit will be against someone with political ties. Not me.
[20:22] <KenMan> i don't know. it could be chosen for various political reasons. Let's shoot down something that works. Let's (legally) shoot down this person who we suspect of being involved in something bad.
[20:22] <KenMan> Let's pick a random user and see how far we can pervert the American Patriot Act, just for later reference.
[20:22] <TheSeeker> heh
[20:23] <KenMan> but all that assumes that the network becomes useable enough to a wide enough audience...
[20:23] <TheSeeker> those legalized illegal data taps won't be of much use against freenet ;)
[20:23] <TheSeeker> mmm, Carnivore...
[20:23] <KenMan> 'Datavhore' comes next.
[20:23] <TheSeeker> http://computer.howstuffworks.com/carnivore.htm :D
[20:24] <KenMan> oops... shh!
[20:24] * guido^pe (~unknown@dsl-213-023-233-226.arcor-ip.net) Quit (Remote closed the connection)
[20:25] * KenMan must go stare at a TV for a while.
[20:25] <TheSeeker> have fun rotting your brain and subjecting yourself to consumerism.
[20:26] <TheSeeker> "I'm a consumer whore!"
[20:26] <KenMan> wheee! Americana of the 21st century (advertising)
[20:28] <TheSeeker> yay http://www.bitterfilms.com/
[20:29] * Elly|afk is now known as Elly
[20:29] <Elly> Wow, why does Freenet use so much memory?
[20:30] <TheSeeker> because it can *nod*
[20:30] <Elly> 40MB used =\
[20:30] <TheSeeker> only?
[20:31] <Elly> Yesterday it was at 125 when I killed it
[20:31] <TheSeeker> Maximum memory the JVM will allocate 260,160 KiB
[20:31] <TheSeeker> Memory currently allocated by the JVM 149,412 KiB
[20:32] <Elly> only 11 connections, all outbound
[20:32] <Elly> rats
[20:32] <TheSeeker> Maximum space for temp files appears to be 1/3 of the datastore... is this correct?
[20:33] * Elly has no idea
[20:33] <TheSeeker> Elly: does it say your correct IP under 'Addresses Detected by the Network' ?
[20:33] <Elly> I configured the IP manually because I have a NAT
[20:34] <Elly> but where would I find that?
[20:34] <TheSeeker> there we go... http://www.bitterfilms.com/rejected-4-300.jpg
[20:34] <TheSeeker> http://127.0.0.1:8888/servlet/nodeinfo/internal/env (assuming defaults)
[20:34] <Elly> yes, I have the correct address
[20:35] <TheSeeker> I think the IP most detected on the network should be the external one
[20:35] <Elly> I tis.
[20:35] <Elly> *It is.
[20:35] <TheSeeker> hrm
[20:35] <Elly> as I said, I set it manually since it was using my internal IP before.
[20:35] <TheSeeker> how many nodes are in your routing table?
[20:36] <Elly> Hmm, it has two addresses detected by the network. How odd.
[20:36] <TheSeeker> http://127.0.0.1:8888/servlet/nodestatus/nodestatus.html
[20:36] <TheSeeker> I ahve two addresses too, one with 1 hit, and the other with 99 hits
[20:36] <Elly> 395 known routing nodes
[20:37] <KenMan> that was enough TV for me ! Memory use seems to scale with number of active connections.
[20:37] <TheSeeker> and only 11 connections? o.O
[20:37] <Elly> now I'm on 15 outbound
[20:37] <Elly> 0 inbound
[20:37] <Elly> ooh
[20:37] <Elly> 20 outbound, 1 inbound
[20:38] <TheSeeker> yey, it works!
[20:38] <Elly> now 22 out / 1 in
[20:38] <Elly> and still nothing will load.
[20:38] <TheSeeker> whats' your uptime?
[20:38] <KenMan> Elly has been up 37 minutes
[20:38] <Elly> node uptime or system uptime?
[20:38] <KenMan> node uptime, silly !
[20:39] <Elly> Not high.
[20:39] <KenMan> minutes, or hours ?
[20:39] <Elly> 9 minutes
[20:39] <TheSeeker> heh
[20:39] <KenMan> oh (giggle, giggle. GuffAW)
[20:39] <Elly> I can't leave Freenet running while I'm playing a game
[20:39] <TheSeeker> were you previously connected for any length of time?
[20:39] <Elly> yes
[20:40] <Elly> maybe two days
[20:40] <TheSeeker> with the manually configured IP? or was that just done now?
[20:40] <Elly> no, I configured it when I installed Freenet
[20:40] <TheSeeker> *nod*
[20:40] <TheSeeker> ok, have you set an outbound rate limit?
[20:41] <Elly> not that I know of, let me check
[20:41] <Elly> aha
[20:41] <Elly> it's set to 12288
[20:42] <TheSeeker> that's fine, assuming you are capable of uploading more than 12 KB/s
[20:42] <Elly> I have 768kbit up, so it should be
[20:43] <TheSeeker> I'd go with 1/2 to 2/3 avaliable upstream for freenet unless you're also running other p2p apps or are acting as a server
[20:43] <KenMan> Elly: well, jack it up then !
[20:43] * Elly calculates
[20:44] <KenMan> 40000 would be good...
[20:44] <TheSeeker> 48000 to 64000
[20:44] <TheSeeker> is what I would use.
[20:44] <Elly> okay, I set it to 49152
[20:45] <TheSeeker> that will allow more routing to go through your node and allow you to get better estimates on other nodes.
[20:45] <TheSeeker> I think
[20:50] <TheSeeker> hmm, is there any way to tell on average what percent of my rt nodes are transferring data?
[20:51] <KenMan> Hey Seeker, what do you think of this histogram ? http://mywebpages.comcast.net/jkcorson/FreeNet/147.png
[20:52] <KenMan> Seeker: there used to be small arrows drawn on the routing table to show who had transfers going, but it has broken for a long time.
[20:52] <KenMan> So, just assume a single transfer per link (faulty assumption) and compare # transfers to # peers.
[20:53] <TheSeeker> Is it safe to assume that all nodes will constantly be transferring data and that half of them will be recieving?
[20:53] <KenMan> not really...
[20:53] <Elly> KenMan: pretty.
[20:53] <TheSeeker> I'm trying to calculate an optimal number of nodes to connect to based on a desired minimum level of upstream performance to each node...
[20:54] <Elly> "Freenet has determined that you are using Internet Explorer or Opera."
[20:54] <TheSeeker> KenMan: asside from neing pretty, it looks pretty evenly distributed...
[20:54] <KenMan> good. That is what I'm after ;)
[20:54] <Elly> I wish the readme would load.
[20:54] <TheSeeker> looks a lot like my datastore does now :P
[20:55] <TheSeeker> readme for what?
[20:55] <Elly> Freenet.
[20:55] <Elly> I need to read the readme to figure out how to configure Opera so it doesn't 'compromise my anonymity'.
[20:56] <Elly> data not found, yay
[20:57] <Elly> looks like time to install Firefox!
[20:58] <TheSeeker> personally, I've never seen anythign on freenet designed to invade my privacy ... but then, I never went and looked for trouble either...
[20:58] <TheSeeker> KenMan quick, before it changes: http://theseeker.bounceme.net:8888/servlet/nodestatus/ds_histogram_detail.txt :P
[20:59] <TheSeeker> it's starting to do that thign again...
[20:59] <KenMan> i see. That effect has been seen before, don't remember the explanation though :(
[21:00] <Elly> okay. Firefox installed and added to my LiteStep menu.
[21:00] <TheSeeker> how do I locate that keyspace on my hdd? I'd like to find the avg file size for those keys...
[21:01] <KenMan> linux ? find store -name 5D\* -exec ls -l {} \;
[21:01] <TheSeeker> no, win
[21:01] <TheSeeker> 1-5d* ?
[21:01] <KenMan> okay, search for files that start with 1-5d . Yeah.
[21:02] <Elly> hmm. Can you configure firefox to open a new tab by default instead of a new window?
[21:02] <KenMan> yes, under configuration, tabbed browsing is an item on the left hand tree-of-choices
[21:02] * goatee (~goatee@ip216-239-84-60.vif.net) Quit ("Leaving")
[21:02] <Elly> Aha, cool.
[21:02] <TheSeeker> I only count 48 files...
[21:03] <KenMan> whhaaa ?
[21:03] <TheSeeker> average size of 460K
[21:04] * skimpIzu (~albo@pcp0010469677pcs.sntafe01.nm.comcast.net) Quit (Read error: 110 (Connection timed out))
[21:04] <KenMan> you should have around 300 ! perhaps some of them are held in temp, under an assumed name ?
[21:04] <TheSeeker> 327 to be exact
[21:05] <TheSeeker> I have 213 objects in temp
[21:05] <TheSeeker> even if all of them were 5d (doubtful) it's still not enough
[21:05] <KenMan> are any of 5D* under temp ? or do they all live under 2 character subdirs ?
[21:05] <KenMan> oh, sorry. You calculated 327, you didn't count up 327 objects :(
[21:05] <TheSeeker> 66 too few
[21:06] * goatee (~goatee@ip216-239-84-60.vif.net) has joined #freenet
[21:06] <TheSeeker> I took the scaling factor*64
[21:06] <KenMan> hi goatie !
[21:06] <KenMan> so something is fudged up then...
[21:06] <TheSeeker> since 5d is my largest peak, that should tell me how many keys the histogram thinks I should have for 5d
[21:07] <KenMan> i know this is silly, but you are looking at the DS histogram and not some other histogram, right ?
[21:07] <TheSeeker> http://theseeker.bounceme.net:8888/servlet/nodestatus/ds_histogram_detail.txt
[21:07] <KenMan> right. Silly me :)
[21:08] <Elly> wow
[21:08] <Elly> I did dir /B /S in my Freenet 'store' directory
[21:08] <Elly> 2064 lines
[21:08] <TheSeeker> heh
[21:09] <KenMan> sounds like you have a bonafide bug there, Seeker !
[21:09] <TheSeeker> keys: 16916
[21:09] <TheSeeker> 15624 files in my freenetds folder...
[21:10] <TheSeeker> which includes the temp folder...
[21:10] <TheSeeker> that leaves me with the histogram thinking I have an extra 1292 keys somehow...
[21:11] <TheSeeker> enough to make up for those impossible peaks...
[21:11] <KenMan> the stock answer is to shut down, delete the index file under store, and restart your node. Oh, and close your eyes while you do it.
[21:12] <TheSeeker> index is rebuilt on launch right?
[21:12] <KenMan> Too bad we can't get a dump of the keys Hops believe he holds...
[21:13] <KenMan> yeah. rebuilt on restart.
[21:13] <KenMan> but, you want to diagnose and understand the problem...
[21:13] <TheSeeker> well, it's not the index file that's corrupt, I dont' think... because on startup the histogram is correct.
[21:14] <KenMan> index is rebuilt if it is not present. I don't know whether it gets reconstructed otherwise.
[21:15] <KenMan> you can see the raw numbers used to generate the histogram, but not the raw key values .
[21:16] <Elly> I wonder how this was done:
[21:16] <KenMan> I can believe there is an accounting bug, but i don't know why they would cluster like that...
[21:17] <Elly> My friend handed me a CD this morning, saying he couldn't get the zip on it to open. I took it home and unrared it, and the 450MB RAR had a 200KB rar inside it.
[21:17] <Elly> and that's it.
[21:17] <Elly> Somehow 200KB got 'compressed' to 450MB.
[21:18] <KenMan> odd
[21:18] <Elly> Very.
[21:21] <TheSeeker> ok, of those extra 1292 keys, 730 of them are in 4 of the 256 buckets, the rest are presumably evenly distributed considering that the rest of the keyspace is pretty flat...
[21:21] <TheSeeker> err 5
[21:21] <Elly> Wow, you guys are smart.
[21:21] <TheSeeker> 5 of 256 buckets
[21:21] <Elly> 'keyspace'.
[21:22] <TheSeeker> 'histogram'
[21:22] <Elly> histrograms I can do.
[21:22] * Elly shrugs
[21:22] <TheSeeker> I miss the 3d barcode chart :(
[21:25] <TheSeeker> hmm, just refreshed and watched the chart morph a little (flatter in the valleys, peaks didnt' change... 15900 files now in the freenetds dir, and 17021 reported in freenet ... still nonly 49 5d files though and 327 reported
[21:26] <KenMan> barcode was only 2.5D
[21:28] <Elly> I need another half-gig of memory for christmas
[21:28] <TheSeeker> it was still fun, and you could zoom!
[21:28] <TheSeeker> something to do while waiting for pages to load...
[21:29] <KenMan> 1X optical zoom, and 40kX digital zoom...
[21:29] <TheSeeker> hehe
[21:31] <Ash-Fox> God damn it, it just takes one large channel to leave a IRC network to set it into chaos
[21:31] <TheSeeker> o.O
[21:31] <TheSeeker> #help ?
[21:31] <TheSeeker> #lobby ?
[21:31] <Ash-Fox> One large channel decided to move, then the server admins say "Well, we're not going to continue paying for shell accounts if there is going to be so little users", then the users hear this and scream "OMG! WE MUST TEH MOVE NOWZZZ!"
[21:32] <TheSeeker> I found a frost bug, yay
[21:32] <Ash-Fox> Now 4/5 users of my network left
[21:32] <Elly> wow
[21:33] <Elly> out of how many total?
[21:34] <TheSeeker> tell Frost to hide unsigned keys, then refresh a board and unhide the unsigned keys... once you read all the new posts you'll have a negative number of posts to read, and the board won't highlight as "updated" again untill the number > 0 ...
[21:34] <TheSeeker> on that note ... zZzz
[21:34] <KenMan> night seeker
[21:34] <Elly> night
[21:35] <KenMan> Ash-Fox - aren't those 4/5 unpaying users ? Shouldn't you care more about the paying ones ?
[21:36] <Ash-Fox> KenMan, users pay to get on IRC?
[21:36] <Elly> We're talking about an IRC network? That you have to pay for?
[21:36] <KenMan> Sorry, I assumed that you were running the irc daemon, and paying shell account members were responsible for it.
[21:37] * Ash-Fox is taking over payment for all servers
[21:37] <Elly> Ash-Fox: so, how many users now?
[21:37] <Ash-Fox> and now, the users hear this, they're all comming screaming back :
[21:37] <Ash-Fox> 66 atm
[21:37] <Ash-Fox> normally it's around 100
[21:38] <Elly> how many servers?
[21:38] <Ash-Fox> 3
[21:38] <Elly> wow
[21:38] <Elly> I'm on a 'network' with 1500 connected users now and 1 server
[21:38] <Ash-Fox> I wanted to have 50 originally
[21:38] <Elly> 50!?
[21:38] <Ash-Fox> So DDOS attacks wouldn't work
[21:38] <Elly> wouldn't it be better to get good filtering far enough upstream for it not to matter?
[21:38] <Ash-Fox> Elly, if you only have one server, it makes a easy target for DDOS attacks
[21:39] <Ash-Fox> Elly, I'm not a ISP
[21:39] <Elly> oh, that explains it
[21:39] <Elly> the guys that run that server are
[21:39] <KenMan> he just plays one on IRC ;)
[21:39] <Elly> well, they aren't an ISP, but they run a datacenter =)
[21:39] <Elly> I think they have 80 gigabyte ethernet links to various backbones or so
[21:40] <Ash-Fox> Trust me, it's hard to find a datacenter that actually let's you run a IRC server
[21:40] <Ash-Fox> You'd be amazed how many say in the contracts/agreements that "you may not run a IRC server" or such
[21:40] <Elly> Ah
[21:40] <Elly> They pay for it.
[21:40] <Elly> and they sell web hosting, too
[21:41] <Ash-Fox> Err no, it's colocation
[21:41] <Elly> hmm?
[21:41] <KenMan> with bw limit requestations , huh
[21:41] <Ash-Fox> They simply take servers we built etc..., and run it in their datacenter
[21:41] <Elly> mhm
[21:42] <KenMan> and point a knife at the largest 3 bw eaters
[21:42] <Ash-Fox> KenMan, at the moment we have 60GB max on each server, more than enough :)
[21:42] <Ash-Fox> monthlu
[21:42] <Ash-Fox> Any more, and the server is immidately disconnected from the net, instead of paying extra, and it's IP becomes unroutable
[21:43] <KenMan> so why their personal hatred of IRC ??
[21:43] * Ash-Fox prefers that kind of agreement, instead of keeping it connected and then we pay large bandwith bills
[21:43] <Ash-Fox> KenMan, some datacenters don't like IRC servers, because of trojans, ddos etc.. etc...
[21:44] <KenMan> ahh, just giving in to IRC's bad-boy reputation.
[21:46] <KenMan> everyone knows only crooks and criminals use IRC !?
[21:47] <Ash-Fox> Amazing thing is that datacenters often don't care about phishing scams etc... which propogant from their networks
[22:08] * SpiKee (~SpiKee@ABordeaux-251-1-14-124.w82-125.abo.wanadoo.fr) has joined #freenet
[22:09] <SpiKee> Hi
[22:09] <Elly> Hi
[22:49] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[22:58] <Elly> anyone have any sort of gues on how many users there are for Freenet?
[23:30] * SpiKee (~SpiKee@ABordeaux-251-1-14-124.w82-125.abo.wanadoo.fr) Quit ("Leaving")
[23:43] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[23:51] * Elly is now known as Elly|afk
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.
