Timestamps are in GMT/BST.
[0:00] * hirvox (~hirvox@cs181027153.pp.htv.fi) Quit (Read error: 104 (Connection reset by peer))
[0:18] <KenMan> yeah, ever since this show called 24 came out, I assumed it was that "7 days" show. But that one hasn't been on TV for at least a year or two...
[0:21] <KenMan> it was pretty funny, the way they ascribed time-travel and basic science to NSA. My favorite one was how they portrayed NSA as responsible for alien diplomacy and Area 51 and ...
[0:22] <KenMan> everyone has known for many years now that cows were being abducted by NSA, not by aliens. And NSA cornered the market in crop rings way back in the 60's.
[0:24] <KenMan> 7 Days was funny for how badly they ignored the actual duties of NSA. And for how easily the average american was educated on what NSA supposedly does.
[0:25] <KenMan> Now most people that viewed that show believe it is the NSA who stirs their toilets in a counter-clockwise direction when flushed. It is so obvious, when you consider that their range only covers the northern hemisphere.
[0:26] <KenMan> of course toilets in the southern hemisphere are allowed to flush in the universally Normal clockwise direction !!
[0:29] <Seed212> hehe
[0:32] <KenMan> ooh, I think my stable 5086 node is coming to life, after 5-6 hours of struggling. Man, fred had to fight tooth & nail for these 52 peers... it should be smooth sailing from here on out.
[0:34] <KenMan> fred is also beginning to have a smidgen of successes now, yet is reaching levels like 24K qph... that is about 23Kqph higher than normal for me :(
[0:35] * Anvil_Vapre (~Cichli@63-227-158-112.omah.qwest.net) Quit (Connection timed out)
[0:37] <KenMan> also , the number of backed off nodes has averaged 99% throughout. That can't be helping things very much :p
[0:41] <Seed212> I wonder if a qued system would work... where routs are always found but you have to wait for your turn to use it..
[0:44] <Seed212> I gotta get some sleep ttyl
[0:51] * Anvil_Vapre (~Cichli@63-227-158-112.omah.qwest.net) has joined #freenet
[3:55] * Fisu_ (~fisu@hoas-50dd08f7.hoasnet.inet.fi) has joined #freenet
[3:56] <Fisu_> What do the "Estimate graph"s of other nodes tell at the RTS page?
[3:59] <Iakin3> It tells what time your node thinks that it will take to retrieve a key, located in the given area, from that node
[3:59] <Iakin3> Did that make sense?
[3:59] <Iakin3> X-axis is keyspace, Y axis is time
[4:01] <Fisu_> So it (when it works) reflects directly the datastores of other nodes? Well anyway the reason I was asking is that for the first time the estimate graphs actually look like they might countain some useful information against before showing sin curves and other strange curves like that. Now they are more like bar codes
[4:05] <Fisu_> Only problem now is that my node has no specialization and according to those estimators almost all other nodes are very specialized. Which mighti indicate that the graphs aren
[4:05] <Fisu_> very accurate
[4:10] <Iakin3> Fisu: They reflect the combination of your DS and the DS's of the nodes that your node can route to
[4:11] <Fisu_> Why does the content of my DS matter when requesting data from other nodes?
[4:11] <Iakin3> For instance.. if your node dont have a specific key in its store but it has a good place to get that key from it will be 'fast for that key too' in the eyes of other nodes
[4:12] <Fisu_> ah, clear enough
[4:37] * Pascal is now known as Pascal666
[5:05] <Iakin3> thelema: You here?
[5:30] * BlackyVIP (~VIP@rsteam.org) has joined #freenet
[5:31] * BlackyVIP (~VIP@rsteam.org) has left #freenet
[5:51] * BlackyVIP (~VIP@rsteam.org) has joined #freenet
[5:51] <BlackyVIP> is this support chanel of www.freenet6.com ?
[5:59] <iip_i2p> <oOo> Not at all
[6:05] * BlackyVIP (~VIP@rsteam.org) has left #freenet
[6:31] * piranha (~piranha@192.220.223.28) has joined #freenet
[6:46] * robilad[afk] is now known as robilad
[7:01] <sanity> KenMan: here?
[7:01] * sanity changes topic to 'Upgrade to 5086 | Unstable: Upgrade to 60170 | Freenet: Taming the World's Largest Tamagotchi | Logs: http://newton.matcmp.ncc.edu/~lockej/freenet/chanlog/'
[7:02] * sanity changes topic to 'Freenet: Taming the World's Largest Tamagotchi | Upgrade to 5086 | Unstable: Upgrade to 60170 | Logs: http://newton.matcmp.ncc.edu/~lockej/freenet/chanlog/'
[7:02] <sanity> (title of my DEFCON talk)
[7:14] <mazzanet> heh tamagotchi
[7:26] <sanity> its official, my new kitten is cute ;-) :- http://www.ratemykitten.com/ratemy/kitten?image=151083
[7:29] * |UK-Monster| (~bbtt@cpc2-warr1-5-0-cust27.bagu.cable.ntl.com) has joined #freenet
[7:29] <mazzanet> uh
[7:36] * guido^pe (~unknown@dsl-213-023-243-038.arcor-ip.net) has joined #freenet
[7:36] * Hory (~Miranda@81.196.25.110) has joined #FreeNet
[7:38] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[7:38] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[7:44] * leexgx (~bbtt@cpc2-warr1-5-0-cust27.bagu.cable.ntl.com) Quit (Read error: 110 (Connection timed out))
[7:46] * guido^pe (~unknown@dsl-213-023-243-038.arcor-ip.net) Quit (Read error: 60 (Operation timed out))
[8:06] * piranha (~piranha@192.220.223.28) Quit (Read error: 60 (Operation timed out))
[8:12] * Pascal666 (Pascal@c-24-13-17-191.client.comcast.net) Quit ()
[8:40] * greycat (~wooledg@192.35.79.70) has joined #freenet
[8:54] * Hory (~Miranda@81.196.25.110) Quit (Connection timed out)
[9:06] * toad_ (toad@82-32-16-91.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[9:11] <Iakin3> Morning
[9:26] * Sugadude (~Sugadude@dsl027-178-156.sfo1.dsl.speakeasy.net) has joined #freenet
[9:28] * thelema|aikido is now known as thelema
[9:29] <Iakin3> Hello thelema
[9:29] <Iakin3> Care to check something for me?
[9:32] <KenMan> oh man, sanity has turned me on to this new kitten site, and let me tell you, some of these kitties are HOT !! :o
[9:33] <Iakin3> I think there is something broken with those strange if-statements in SBKE.report()
[9:33] <thelema> ian: what's up?
[9:33] <thelema> ok, where?
[9:34] <KenMan> I used to waste time on those "rate this girl' sites, but not for a few years now. I'll never get anything done with FreeNet, as I'm hooked on this kitten site.
[9:34] <Iakin3> ok.. there is a statement checking if(idx==0)
[9:34] <Iakin3> (I have messed up my line#'s
[9:35] <Iakin3> )
[9:35] <thelema> gimme a sec...
[9:36] <Iakin3> The comment below there states 'We are behind center #0'
[9:36] <Iakin3> Is that really right?
[9:37] <thelema> idx == 0 means the searched element is < index 0
[9:37] * Iakin3 would say that we are _before_ center #0
[9:37] <thelema> so behind means to the left of, in this instance.
[9:37] <Iakin3> Right.. so the comment is wrong at least
[9:37] <thelema> I agree that's a bad word to use.
[9:37] <Iakin3> ok, I'll use 'to the left' instead
[9:38] <Iakin3> But.. then .. the comparision below..
[9:38] <Iakin3> If we are to the left of the center.. why compare the leftEdge with the center?
[9:38] <Iakin3> Hmmm...
[9:39] <Iakin3> Ahh.. because the leftEdge might actually be somewhere to the _right_?
[9:40] <thelema> yes; leftEdge[0] can be way off the right edge of the keyspace
[9:40] <KenMan> i think 5086 success rates are good, but it's really too soon to tell. Saying it may take days or longer to join the network is appropriate !!
[9:41] <Iakin3> ok, then I assume that that clause is right..
[9:41] <Iakin3> Now on to the previous compareTo() then..
[9:42] <Iakin3> We are compairing the left edge of the first sector with the left edge of the second sector..
[9:42] <KenMan> Iakin3 I know you like to be productive (don't we all?) but your jbigi contribution is _significant_ !! thanks !
[9:43] <Iakin3> Shouldn't some test wrt the last sector be in there somewhere?
[9:43] <Iakin3> Kenman: Thanks
[9:43] <thelema> Iakin: one can compare the first to the second or to the last.
[9:43] <thelema> if the first is to the right of the second, we're wrapped.
[9:44] <thelema> if it's to the right of any sector we're wrapped
[9:44] <Iakin3> if(leftEdge[0] < leftEdge[1]) then....
[9:44] <thelema> because normally the first (well, the 0th, but whatever) is to the left of them all.
[9:46] <Iakin3> But the find says that the 'key should be inserted somewhere after the last center'
[9:46] <Iakin3> Shouldn't we then compare the key with leftEdge[0] then.. to figure out if it fell into the last or the first bucket
[9:48] <thelema> isn't the comparison compound?
[9:48] <Iakin3> Not as far as I can see..
[9:49] <thelema> hmm, that might be an error, then.
[9:49] <thelema> in guess(), right?
[9:49] <thelema> or in report?
[9:50] <KenMan> pSuccess / 16 buckets is showing highest success in f, but lowest in 0 (it doesn't seem to wrap) ...
[9:50] <Iakin3> in report
[9:51] * hirvox (~hirvox@cs181027153.pp.htv.fi) has joined #freenet
[9:51] <thelema> okay, the case where idx = 0 means that K is to the left of C0
[9:51] <thelema> Cx = center x, Ex = edge x
[9:51] <KenMan> i take that back... the numbers are not significant at 256 buckets
[9:52] <thelema> if we're in the wrapped case (E0 > C0), then there's no edge to compare with, and we know we're in bucket 0
[9:53] <thelema> if we're not in the wrapped case, we know to compare with center 0.
[9:53] <thelema> which is what we do below.
[9:54] <Iakin3> right
[9:54] <thelema> Iakin: want more explanation?
[9:55] <thelema> similarly for idx=15
[9:55] <Iakin3> No, that one is ok.. but the one above
[9:55] <thelema> not 15
[9:55] <thelema> 16
[9:55] <Iakin3> 16, right
[9:55] <thelema> if we're past C15, we need to know if we're wrapped.
[9:55] <Iakin3> Yes
[9:55] <Iakin3> i.e. E0
[9:56] <thelema> if we aren't, we don't need to worry about anything; we know what bucket we're in
[9:56] <thelema> and if we are, then we need to do the normal compare.
[9:56] <thelema> if you want, an isWarpped() function or maybe isSplit() that returns a boolean would be good.
[9:56] <Iakin3> so.. .shouldn't we compare k to E0 somewhere there?
[9:57] <Iakin3> Instead of E0 to E1?
[9:57] <thelema> yes, idx = 16 should be % accuracy to get which edge to compare with.
[9:57] <thelema> we compare E0 to E1 to determine if we're wrapped.
[9:57] <thelema> note that this case is the opposite of the first case.
[9:58] <thelema> the first one is ==1, the second is ==-1
[9:58] <thelema> the E0 <=> E1 is just for determining wrapped status.
[9:59] <Iakin3> Shouldn't we try to compare E0 and C0 instead?
[9:59] <KenMan> toad_- 5086 query rate limiting looks badly busted... the queries per hour bounces all over the place, and QPM is not really any better. It certainly doesn't appear to be "under control" :p
[9:59] <Iakin3> Just for consistency
[10:00] <thelema> as I said, feel free to encapsulate the logic of testing for a wrapped E0 into a function so you can work with a nice boolean instead of having magic -1's and 1's
[10:00] <Iakin3> private boolean isWrapped(BigInteger low, BigInteger high){
[10:00] <Iakin3> return low.compareTo(high) ==1;
[10:00] <Iakin3> }
[10:01] <Iakin3> That would work, wouldn't it?
[10:01] <thelema> Iakin: no, private boolean isWrapped() { return E0.compareTo(E1) == 1; }
[10:02] <thelema> isWrapped would be a function of the whole bucket array, it'd only need to be done with E0 and E1
[10:03] <thelema> it'd be useless to do isWrapped(E3,E8) because E3 < E8
[10:03] <Iakin3> ok.. So.. hmmm... C0 can be wrapped too then?
[10:03] <thelema> but E0 might be on either end.
[10:03] <thelema> no, the *centers* are always kept in order. That's why E0 could be wrapped.
[10:04] <thelema> because the ffff.0000 interface might end up being between an edge and a center (since the keyspace is modular)
[10:04] <thelema> and because of the way we sort centers, it's got to be between E0 and C0
[10:05] <Iakin3> Right.. so.. return E0.compareTo(C0) would work as well as compareTo(E0) then..
[10:05] <thelema> you mean E1, right?
[10:05] <Iakin3> Yea
[10:06] <thelema> then yes.
[10:06] <sanity> is freemarker in freenet-ext.jar now?
[10:06] <thelema> you can compare E0 to anything else and it should be less then them all or more than them all.
[10:06] <Iakin3> Right
[10:08] <hirvox> sanity:no
[10:09] <hirvox> BookmarkServlet, NodeStatusServlet and the splitfile request servlet need some work, so I'm not merging until I'm done
[10:09] <sanity> hirvox: fair enough
[10:09] <sanity> hirvox: what does your work so-far look like?
[10:11] <hirvox> hirvox:I've mainly worked on infrastructure, so visually it looks more or less the same
[10:11] <hirvox> I've done some little things, like proper formatting of time intervals, bytes etc
[10:12] <sanity> cool
[10:13] <Iakin3> Thelema: Does this look ok you you:
[10:13] <Iakin3> if(idx == buckets.length) {
[10:13] <Iakin3> // We are after center for the last sector, at the right hand end
[10:13] <Iakin3> if(!isWrapped()) {
[10:13] <Iakin3> // firstSector.leftEdge < firstSector.center < lastSector.center
[10:13] <Iakin3> // We are at the high end, so we can now be sure that we are in the last bucket
[10:13] <Iakin3> return new FindResult(lastSector,FindResult.RIGHT_OF_CENTER);
[10:13] <Iakin3> }
[10:13] <Iakin3> }
[10:13] <Iakin3> wrt. the comments mostly
[10:13] <thelema> yes.
[10:14] <Iakin3> Good :)
[10:14] <hirvox> the time interval and bytes formatting is centralized in misc/functions.ftl, so instead of modifying n+1 format() functions, it's sufficient to make the changes in one place
[10:15] <hirvox> likewise, all UI widgets (like pages, titleboxes, progress bars etc) are all in misc/widgets.ftl
[10:16] <hirvox> so, to turn my simple theme into aqua, it's enough to modify widgets.ftl (and the css)
[10:17] <hirvox> s/modify/override/
[10:21] <iip_i2p> <Newsbyte> anyone knows how the paddingthingy is going toad talked about yesterday?
[10:21] <thelema> news: it's done.
[10:22] <thelema> news: it's a bit crude, but it's implemented.
[10:23] <iip_i2p> <Newsbyte> is it det at 128 or 256?
[10:23] <iip_i2p> <Newsbyte> set
[10:24] <thelema> news: 256
[10:24] <iip_i2p> <Newsbyte> pitty, the smaller the chunks the better (in regard to wasted b/w)
[10:25] <Seed212> is setting up an unstable node from scratch as simple as the instructions on the main freenet webpage?
[10:26] <iip_i2p> <Newsbyte> let's say: no
[10:26] <Seed212> cause my node cannot connect ot any of the 72 references
[10:27] <iip_i2p> <Newsbyte> welcome to the club (though I have stable, curently)
[10:27] <Seed212> Connection Attempts 1095
[10:27] <Seed212> Successful Connections 141
[10:27] <Seed212> Contacted node references 0
[10:28] <iip_i2p> <Newsbyte> going from stable to unstable is well-explained, though. It's also on the www.freenethelp.org wiki
[10:29] <thelema> Seed212: ask someone running instable for some references.
[10:30] <Seed212> anybody on unstable got a reffernce I could use?
[10:31] * Ash-Fox (Hal-9000@aad121.neoplus.adsl.tpnet.pl) Quit (Remote closed the connection)
[10:35] <Iakin3> Seed: Same thing here
[10:35] <Iakin3> thelema: It is not his fault.. something is causing it
[10:35] <thelema> really? There's connection problems on unstable?
[10:36] <Iakin3> One of my two nodes cannot connect to anything
[10:36] <thelema> did toad add one too many NBIs?
[10:36] <thelema> or is it just nodes having too many newbie connections and not allowing any more
[10:37] <Iakin3> thelema: timeouts timeouts and timeouts is what it causing the troubles
[10:37] <thelema> really? the timeouts got broken?
[10:37] <thelema> I remember seeing toad doing something to some timeouts.
[10:37] <Iakin3> I think more like some kind of link corruption or so
[10:38] <thelema> increasing them to make inserts work better.
[10:38] <thelema> the crypto issues can't be killing us this completely now; we've had them all along.
[10:38] <Iakin3> toad mentioned seeing 'the other node doesn't posess the private key for its public key'-messages lately
[10:38] <thelema> really?
[10:38] <Iakin3> And thought that that was what I am seeing
[10:39] <Iakin3> Maybe one of those new BigInteger(byte[]) -> new BigInteger(1,byte[]) changes or something
[10:40] <thelema> are that many unstable nodes running native crypto?
[10:40] <Iakin3> Toad tried using non-native BI on his testnet.. didn't help
[10:40] <Iakin3> is probably isn't the native code
[10:41] <thelema> why would a bigint be negative?
[10:42] <Iakin3> No idea.. it was just a thought
[10:42] <Iakin3> they _can_ be...
[10:43] * Iakin3 committed
[10:43] <Iakin3> Build 60171:
[10:43] <Iakin3> Refactoring; Move large parts of the SBKE code into a separate class 'BucketStore'
[10:43] <Iakin3> I am done with the refactoring now
[10:43] <Iakin3> (I hope)
[10:44] <Iakin3> authentity.g=930168de21e7fb66c0375e08e964255a0f7f0ad54507a51864afdc686f36be8bb8b7865408116060c5f34f94b5146cbef9e4adb70324fba01d34c1c60817cbadf6854d654176cb391de0d41e0f0fbbc8ceea5546c09a676b0d9a9988c7a1ce36ce31596037a18b4d540374bdf2ad071a3f8dd1015a9d8ba0f0d51cde212db6daauthentity.q=ef1f7a7a73362e526515f348075aee265e9eff45
[10:44] <Iakin3> Aren't those negative?
[10:44] <thelema> hmm...
[10:44] <Iakin3> Starting with '9' and 'e' respectively
[10:44] <thelema> they shouldn't be. do they need to be padded with a 0?
[10:45] <Iakin3> Thats the content of my 'node' file..
[10:45] <thelema> all crypto constants are positive.
[10:45] <iip_i2p> <Newsbyte> true
[10:45] <thelema> I don't know how the bigint parser works
[10:46] <thelema> whether it takes the leading bit to be sign or what...
[10:46] <Iakin3> The byte[] parser for sure does...
[10:46] <Iakin3> possibly not the String-parser though
[10:46] <Iakin3> probably not even
[10:46] <Iakin3> ok, I am off bbl
[10:46] <Seed212> cya
[10:47] <Iakin3> let me know if my routing refactoring broke something
[10:47] <Iakin3> (haven't really tested it since my node wont work)
[10:48] <Iakin3> Heh.. it didn't work
[10:48] <Iakin3> damn
[10:49] <Iakin3> ok, I'll fix it a little bit later
[10:53] <Iakin3> darn.. binarySearch(X,Y) doesn't work very well
[10:53] <thelema> what's the problem?
[10:54] <Iakin3> I wanted to binarySearch for BigInteger in an Sector[]
[10:54] <Iakin3> (which implements comparable)
[10:54] <thelema> oops.
[10:55] <Iakin3> but... didn't compare against a BigInteger..
[10:55] <thelema> so you have to ... ick.
[10:55] <Iakin3> public int compareTo(Object o) { //Define the natural ordering of Items as the natural ordering of their centers
[10:55] <Iakin3> if(o instanceof BigInteger)
[10:55] <Iakin3> return center.compareTo(o);
[10:55] <Iakin3> else
[10:55] <Iakin3> return center.compareTo(((Sector)o).center);
[10:55] <Iakin3> }
[10:55] <Iakin3> Hmmm..
[10:55] <thelema> does that work?
[10:55] <Iakin3> Not sure wheter or not that is a good think
[10:55] <Iakin3> _seems_ to
[10:55] <thelema> I think it's a very nice solution.
[10:55] <thelema> much better than what I was thinking of.
[10:56] <Iakin3> heh.. ok, then I'll commit it :)
[10:56] <thelema> make sure to bug a big warning comment on what dyou did.
[10:58] <Iakin3> bbl
[11:01] * thelema is now known as thelema|away
[11:01] <thelema|away> back in a bit
[11:24] * nextime (~nextime@danex.i-m-c.it) Quit (Connection reset by peer)
[11:42] <iip_i2p> <Newsbyte> there, back. Missed me? ;-)
[11:43] * TLF (francisco@248.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[11:43] <iip_i2p> <Newsbyte> wb, tlf
[11:44] * thelema|away is now known as thelema
[11:46] <TLF> hi all
[11:46] <thelema> hi tlf
[11:46] <iip_i2p> <Newsbyte> IIP seems rather unstable today
[11:50] <KenMan> okay, we need our best brains to focus on routedToChoiceRank and the whole backoff scheme someday soon...
[11:51] <KenMan> I believe that the mRI sent from node A to node B needs to be much smoother than it is allowed to fluctuate currently...
[11:51] <thelema> my best solution to load is to decrease mRI.
[11:52] <KenMan> you mean increase mRI, thus decreasing rate of query ??
[11:52] <thelema> does mRI fluctuate that badly?
[11:52] <KenMan> Actually, I haven't looked for a week or two. But queries/minute and queries/hour are jumping around just a bit too much.
[11:52] <thelema> no, I mean decrease mRI by affecting the factors that go into its calculation. bandwidth, memory, CPU, work in general.
[11:53] <KenMan> :)
[11:53] <thelema> getting rid of a 37K overhead on all requests is *very* important.
[11:53] <KenMan> heh, i can't argue with that!
[11:53] <thelema> teh overhead on requests should be as close to the theoretical minimum as we can make it.
[11:54] * thelema tries to figure out how anyone thought freenet could work with 37K StoreData messages and (back when) 4-5K request messages... Who knows.
[11:55] <KenMan> i think a lot of problems can arise from a situation where the individual mRIs from each route are allowed to be radically different from one another.
[11:56] <thelema> from the perspective of a single route, mRIs are irrelevant.
[11:56] <KenMan> In a way, it allows a node to 'eat' too many queries, if that node's mRI is <<< other routes' mRIs
[11:56] <thelema> oh yeah, nodes having wildly different mRIs does imbalance routing choices, but I don't think that's a problem at all.
[11:57] <thelema> I don't have any problem with some pipes in freenet being fatter than others.
[11:57] <KenMan> My node was sucking up 21Kqph, but having very few successes (percentage wise)... it seemed inefficient.
[11:57] <KenMan> no, of course some have more capacity than others. But stability in MRIs seems a worthy goal.
[11:58] <thelema> stability in mRI = between nodes, or for a single node over time?
[11:58] <KenMan> both - i am more concerned with the 'over time' part.
[11:59] <KenMan> If I can say 'mri = 10 minutes' at one time, and then 5 seconds later say 'now it is 4 seconds' and then flip-flop way back again, it doesn't really help my peers.
[11:59] <KenMan> other than to communicate to them that I cannot manage to keep my own load under control...
[11:59] <thelema> I just want a system without QRej
[12:00] <thelema> i.e. where nodes know that the peer they send a request to will be able to serve it.
[12:00] <thelema> if it takes communicating a 'available'/'not-available' flag every 5 seconds, I'm for it.
[12:00] <KenMan> yes. the two are not completely unrelated...
[12:00] <thelema> I'm just talking in general.
[12:01] <thelema> mRI does a decent job of that. But it does have the downside that ...
[12:01] <thelema> I'm not sure how to describe the downside...
[12:01] <KenMan> every 5 seconds with 100 peers seems very inefficient, but if we can demonstrate that it has positive effect, well ... :)
[12:01] <thelema> KenMan: I'd put it on every message as one bit in a binary protocol if I could.
[12:01] <thelema> with a ping every 5 seconds, if no message had been sent.
[12:01] <KenMan> we should be able to say "MRI is this" and not need to change it by more than some small percentage over 10 minutes+ time.
[12:02] <thelema> KenMan: What happens if all of a sudden we hit a pocket where a lot of requests succeed?
[12:02] <thelema> without data chording, we have to push a lot of data through our node.
[12:02] <KenMan> if we are updating the value by a significant amount very often, then we are reacting to load, rather than controlling it...
[12:03] <thelema> and if we can't say 'hold on, I'm busy', then we get even more overloaded by incoming requests.
[12:03] <KenMan> a change in load is roughly caused by a change in query rates...
[12:03] -lilo- [Global Notice] Hi all. As you may be aware, O'Reilly's OSCON 2004 conference is in progress this week. If you're attending, or if you're just interested in knowing what's going on, please stop by freenode's unofficial #oscon channel. Thanks!
[12:03] <thelema> are current requests too chunky to be controlled by mRI? would mRI work better with more smaller requests?
[12:04] <KenMan> that's entirely possible !!
[12:04] <thelema> there's evidence that as a node stays up longer, it picks up more concurrent transfers, which bog down the node, but it doesn't slow incoming load enough to let/make them finish, so it just gets slower and slower.
[12:05] <thelema> which is why people restart their node often, and the newly restarted nodes work much better for them.
[12:05] <KenMan> i don't have that evidence. I see a non-decreasing number of incoming xfers recently, but expect that is a reporting bug. Over time, number of xfers is handled very well, IMHO.
[12:06] <thelema> it levels off at what?
[12:06] <KenMan> People complain because they can't fetch 200 URIs in a matter of seconds...
[12:06] <thelema> heh.
[12:06] <KenMan> well, I usually set 75 peers, and xfers float around 10-20 active at any time...
[12:07] <KenMan> the graphs i make show the 1 minute sampling of in's and out's ...
[12:07] <thelema> how much bandwidth?
[12:07] <KenMan> 12KB/s , sometimes 18KB/s ...
[12:08] <thelema> so between 1.8K/s and 0.6K/s per xfer? ick.
[12:08] <KenMan> lately 18... seems to give me a slight advantage over the average node...
[12:08] <thelema> we need to get bandwidth limiting working for values over 20K/s
[12:08] <KenMan> yeah, but that is 10x better than it used to be. Come on, be positive :)
[12:09] <KenMan> can't you recall the average transfer rate computing to about 12 bytes/second not soo long ago ??
[12:10] * thelema still wants modem nodes to be useful.
[12:10] <thelema> not many people in china have DSL.
[12:10] <KenMan> first we need to reproduce the bw over 20KB/s problem... it should be possible to do it on a local net. I wonder if those people having troubles aren't somehow hitting the limits of their provider
[12:11] <thelema> no, it's usually on university networks
[12:11] <thelema> or business connections
[12:11] <KenMan> well, it should be simple enough to permit bw limiting on local interfaces... it is currently avoided however.
[12:12] <thelema> if you can find the code in the tangle that is the networking stack.
[12:12] <KenMan> if I can see it work reliably on a LAN between two hosts, I want to blame something other than fred code for the troubles...
[12:12] <KenMan> I think I found it
[12:12] <Iakin3> Whats the problem with that bw?
[12:12] <Iakin3> (>20k)
[12:13] <KenMan> suposedly it becomes unstable for people, not correctly limiting the flow
[12:13] <Iakin3> Aha..
[12:13] <Iakin3> Well.. that might be true
[12:14] <KenMan> i'm just saying, if I set a value of 57KB/s for two LAN attached nodes, and they smoothly hit the 57KB/s ceiling, it would seem to me that the code is accurate.
[12:14] <thelema> Iakin: I thought I had fixed getDataLength
[12:14] <KenMan> but i haven't done it yet...
[12:14] <thelema> Iakin: but it's broken now.
[12:14] <Iakin> iakin.poweruser.org/freenetstuff/bw.jpg
[12:15] <Iakin> Thats with 50kbyte/s limiting or so
[12:15] <Iakin> thelema: Huh.. In what way?
[12:15] <Iakin> bbl
[12:16] <thelema> Iakin: it's not counting the header...
[12:16] <thelema> public int getDataLength() {
[12:16] <thelema> int l = 4 + 4;
[12:16] <thelema> that should be 4+4+4
[12:38] * lolo-laptop (~lostlogic@68.251.84.226) has joined #freenet
[12:50] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) has joined #freenet
[12:50] <Pascal> howdy
[12:51] <thelema> hi
[12:52] <Pascal> <toad_> <Pascal> toad: I wonder if L66-L72 are also in the wrong order - I don't get it... - in the same way L60 & L61 were backwards. Shouldn't message be assigned before going into logging?
[12:53] * TLF (francisco@248.Red-81-40-116.pooles.rima-tde.net) Quit (Remote closed the connection)
[12:56] <Iakin3> thelema: serial magic + version + accuracy, right?
[12:56] <thelema> Iakin: yup
[12:58] <Iakin3> i'll fix
[13:01] * thelema is tempted to re-implement key as a fixed-length byte array, including proper IO methods for it.
[13:02] <thelema> + int l = 4 + 4; //serial magic + version
[13:02] * Sugadude (~Sugadude@dsl027-178-156.sfo1.dsl.speakeasy.net) Quit ()
[13:02] <thelema> what about accuracy?
[13:06] * guido^pe (~unknown@dsl-213-023-243-038.arcor-ip.net) has joined #freenet
[13:08] <robilad> what did i break? :)
[13:08] <robilad> hi guido^pe
[13:14] <guido^pe> hi
[13:14] <Iakin3> thelema: It is written, and accounted for, by BucketStore
[13:15] <Iakin3> bbl
[13:16] <thelema> ah. that's the other change. good.
[13:26] <Seed212> is mass overloading of nodes the current bottleeck in freenet?
[13:28] <thelema> bandwidth is the bottleneck that's causing mass overloading of nodes.
[13:28] <KenMan> sure looks that way on stable network (5086)
[13:29] <lolo-laptop> on stable was _killing_ my CPU... :-(
[13:30] <KenMan> CPU not bad here, just "everything else"...
[13:30] * wind789 (wind789@6532175hfc51.tampabay.rr.com) has joined #freenet
[13:30] <wind789> sorry my connection was lost
[13:30] <thelema> lolo-laptop: try NBI
[13:30] <lolo-laptop> I've taken my node down while I serve a few hundred images for my car club any way...
[13:30] * Seed212 (wind789@6532175hfc51.tampabay.rr.com) Quit (Read error: 104 (Connection reset by peer))
[13:31] <wind789> are there any plans up raise the current preset upload limit, or set it dynamically based on teh users connection?
[13:31] <lolo-laptop> thelema: I don't know how to make NBI work with stable... I have it in my freenet-ext.jar ... ...
[13:31] <greycat> What does the "--skipDS" option on "freenet.client.cli.Main put" actually *do*?
[13:31] <greycat> Does it only skip the data store for the mapfile key, or for all the chunk CHKs, or both, or neither?
[13:33] <thelema> lolo-laptop: I'm not sure it is in stable yet, but it should be soon.
[13:33] <greycat> When I insert a splitfile with "freenet.client.cli.Main put --htl 20 CHK@ filename", I get a CHK. Good, that's better than 5084 ever managed. Then when I do it again, I get the same CHK, and it's fast. Even better! Then I do it a third time with "--skipDS" and all the chunks come back quickly (as though it's not reinserting them), but the final mapfile insert takes long.
[13:33] <KenMan> NBI in stable is on toad's todo today list ...
[13:33] <lolo-laptop> thelema: that's what I thought... hence stable kicking my CPU's ass... I guess I'd just gotten used to unstable..
[13:33] <lolo-laptop> KenMan: thanks
[13:33] <thelema> greycat: in theory it should skip for all.
[13:34] <thelema> greycat: I guess it's only actually applying it to the mapfile
[13:34] <KenMan> lolo-laptop how many peers have you ? I've only got ~60 so far, but CPU use is LOW
[13:34] <KenMan> and qpm is HIGH ...
[13:35] <sanity> yay, just got continuations working in java
[13:35] <lolo-laptop> KenMan: node is down right now, later today once everyone has the pics I'm hosting, I'll fire it back up
[13:35] <thelema> sanity: continuations?
[13:35] <thelema> sanity: how?
[13:36] <sanity> thelema: http://www.velare.com/product/atct.htm
[13:36] <sanity> its currently closed source, but there is a good chance it will be open soon
[13:36] <thelema> patent pending?
[13:37] <sanity> the company is going out of business and the guy behind it wants to open source it
[13:37] <sanity> the patent thing is probably just investors fodder
[13:37] <thelema> ah.
[13:38] <sanity> i am also talking to another guy who has another continuations implementation using a different approach, it is open source but currently tied into his software, he has promised me that he will separate it out soon
[13:39] <thelema> sounds like fun
[13:42] * TLF (francisco@175.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[13:43] <guido^pe> I'm having trouble starting Freenet stable on a FreeBSD machine with Sun Java 1.4.2. It will startup and then endlessly consume CPU cycles until I stop it, but won't show any other "lifesigns"
[13:43] <guido^pe> (like a "store" directory appearing or port 8888 becoming available)
[13:44] <thelema> guido^pe: you're emulating linux binaries, right?
[13:44] <guido^pe> After I raised loglevel to "debug", these lines started accumulating at the end of freenet.log: "Jul 26, 2004 7:58:06 PM (freenet.support.BlockingQueue, PRNG/Yarrow entropy processing thread, DEBUG): Waiting... class java.util.LinkedList:[]"
[13:45] <thelema> is /dev/random a blocking source of entropy?
[13:45] <thelema> s/random/urandom/
[13:45] <guido^pe> thelema: You mean with the Java VM? Well, I installed it from java/jdk14...
[13:45] <greycat> on linux, random blocks, urandom does not
[13:45] <thelema> greycat: what about freebsd
[13:46] <greycat> I don't know fbsd. obsd has random, srandom, urandom, prandom, arandom.
[13:46] <guido^pe> hm, if I cat /dev/urandom, it produces tons of output...
[13:47] <guido^pe> fbsd only has random and urandom
[13:47] <thelema> guido^pe: okay, that;'s not it... hmmm.
[13:47] <greycat> http://www.openbsd.org/cgi-bin/man.cgi?query=random&apropos=0&sektion=4&manpath=OpenBSD+Current&arch=i386&format=html
[13:47] <thelema> j.c.u.LinkedList:[]
[13:47] <thelema> hmm
[13:47] <thelema> oh well, it's time for me to go.
[13:49] <guido^pe> still noone here who can help me?
[13:51] * thelema is now known as thelema|away
[13:52] <guido^pe> anyone here who is or has been using Freenet successfully on FreeBSD?
[13:54] <Orasis> Anyone going to defcon???
[13:54] <KenMan> just sanity
[13:55] <Orasis> sanity == ian?
[13:59] <KenMan> yes
[14:00] <Orasis> Cool, I'll have to keep an eye out for him.
[14:00] <KenMan> where a rabbit on your lapel then
[14:00] <KenMan> wear it, i mean !
[14:01] <Orasis> santiy: If you read your backlog, maybe see you there. I'm no longer bald, so I might spot you before you spot me.
[14:01] <Orasis> Hmm...maybe I should get some Swarmcast t-shirts made up....
[14:01] <greycat> wouldn't the rabbit start to smell after a few hours?
[14:02] <KenMan> not if properly embalmed !
[14:02] <KenMan> make him happy, use a t-shirt with Hops emblazoned across the chest and back...
[14:03] <Orasis> KenMan: I don't think ian would be too happy if I coopted the freenet rabbit for the Swarmcast logo :)
[14:05] <KenMan> well, just don't wear a beehive on your person. People would probably get upset.
[14:06] <greycat> insect rights activists?
[14:07] <KenMan> no - people allergic to swarm stings...
[14:08] <KenMan> the IRA (insect rights assoc) doesn't have much of a presence in LA...
[14:09] <greycat> normally one would recommend removing the honey and bees from the hive before converting it into personal ornamentation
[14:09] <KenMan> not me... I'm not that thoughtful. :p
[14:12] <sanity> Orasis: yes, i will be doing one of my classic "bore people who know freenet, and confuse those that don't" talks
[14:37] * Hory (~Miranda@81.196.25.110) has joined #FreeNet
[14:46] * Anvil_Vapre (~Cichli@63-227-158-112.omah.qwest.net) Quit (Read error: 104 (Connection reset by peer))
[14:51] <Iakin> Anyone know if dodo will resend list-mails that bounced due to 'full mailbox'?
[15:20] * salahx (~salahx@sc1-24.217.174.147.charter-stl.com) has joined #freenet
[15:46] * salahx (~salahx@sc1-24.217.174.147.charter-stl.com) Quit (Remote closed the connection)
[15:46] * greycat (~wooledg@192.35.79.70) Quit (Read error: 104 (Connection reset by peer))
[15:47] * greycat (~wooledg@192.35.79.70) has joined #freenet
[15:59] * nextime (~nextime@danex.i-m-c.it) has joined #freenet
[16:02] * iip_i2p (~changate@host.teitel.net) Quit (Remote closed the connection)
[16:02] * iip_i2p (~changate@host.teitel.net) has joined #freenet
[16:12] * robilad (~topic@mpiat2304.ag2.mpi-sb.mpg.de) Quit (Remote closed the connection)
[16:14] * greycat (~wooledg@192.35.79.70) Quit ("Lick Bush in '04")
[16:18] * robilad (~topic@mpiat2304.ag2.mpi-sb.mpg.de) has joined #freenet
[16:39] * Pascal666 (Pascal@c-24-13-17-191.client.comcast.net) has joined #freenet
[16:41] * iip_i2p (~changate@host.teitel.net) Quit (Remote closed the connection)
[16:42] * iip_i2p (~changate@host.teitel.net) has joined #freenet
[16:45] * robilad is now known as robilad[afk]
[17:02] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) Quit (Read error: 110 (Connection timed out))
[17:02] * Pascal666 is now known as Pascal
[17:02] * thelema|away is now known as thelema
[17:04] <thelema> anyone here?
[17:07] <Iakin> me :)
[17:08] <thelema> is there any way to clean up the implementation of getBucketValue?
[17:09] <thelema> It's just painful as is. I told toad that the idea should probably be integrated into the interpolate routine
[17:09] <thelema> but he convinced me that it was easier to do it this way.
[17:09] <Iakin> Hmm... let me look
[17:10] <Iakin> Yes.. actually I have done that exact thing once :)
[17:11] <thelema> you did that once?
[17:11] <Iakin> But I had to redo all of the refactoring from scratch today since I wanted to commit the stuff in multiple phases (for peer reviewing)
[17:11] <Iakin> And in that process I dropped that particular change ;(
[17:11] <thelema> :(
[17:12] <thelema> I carefully reviewed everything but the move to BucketStore.java
[17:12] <Iakin3> It is a really good idea to try working with the .prev() and next() methods
[17:12] <thelema> does it take any work to do array.length?
[17:13] * thelema has to go
[17:13] * thelema is now known as thelema|aikido
[17:13] <Iakin3> They tend to be safer to work with than random access to the buckets using indexes
[17:14] <thelema|aikido> oh yeah, what about making a rightEdge() function?
[17:14] <thelema|aikido> that way you don't have to do next.next
[17:14] <Iakin3> I had one such
[17:14] <thelema|aikido> I guess you can't return a lvalue in java...
[17:14] <thelema|aikido> anyway, gotta go.
[17:15] <thelema|aikido> ttyl
[17:15] <Iakin3> ok, cu
[17:16] <Iakin3> Hmm..what about an fan-out iterator
[17:19] * guido^pe (~unknown@dsl-213-023-243-038.arcor-ip.net) Quit (Remote closed the connection)
[18:29] <iip_i2p> <Newsbyte> padding finished, toad?
[18:31] <toad_> <KenMan> oh man, sanity has turned me on to this new kitten site, and let me tell you, some of these kitties are HOT !! :o - ugh, bloody pornographers... especially the inter-species ones...
[18:32] <toad_> <KenMan> I used to waste time on those "rate this girl' sites, but not for a few years now. I'll never get anything done with FreeNet, as I'm hooked on this kitten site. - /me moves that the project provide a free castration service for developers... :)
[18:33] <iip_i2p> <Newsbyte> I'll castrate you myself if that means further work on Freenet, toad, don't worry.
[18:34] <toad_> <KenMan> toad_- 5086 query rate limiting looks badly busted... the queries per hour bounces all over the place, and QPM is not really any better. It certainly doesn't appear to be "under control" :p
[18:35] <toad_> hmmm
[18:35] <toad_> strange
[18:35] <toad_> Newsbyte: I meant KenMan
[18:35] <toad_> :)
[18:36] <iip_i2p> <Newsbyte> You did? Darn, I was just going to say: "thus a coder improves his sex-life: by getting turned on, not by pussies, but by kittens."
[18:37] * The_Saint (~SimonT@66-133-132-188.nas2.roc.ny.frontiernet.net) has joined #freenet
[18:38] <toad_> <iip_i2p> <Newsbyte> pitty, the smaller the chunks the better (in regard to wasted b/w) - yes BUT the packet overhead is at least 24 bytes (TCP), and probably more like 40 - it doesn't matter that much. I will make it 128 bytes though soon.
[18:38] <iip_i2p> <Newsbyte> you know, like in pussy (ambigue) - cat / cat - kitten ? (wordplays are not easy in another language, you know :-)
[18:39] <iip_i2p> <Newsbyte> I thought you couldn't?
[18:39] <toad_> <Seed212> cause my node cannot connect ot any of the 72 references - yeah, well known bug, I was working on it on Saturday, it's not related to NativeBigInt and affects both branches
[18:39] <iip_i2p> <Newsbyte> wasn't there 4 bytes too much, somewhere?
[18:40] <toad_> <Iakin3> thelema: timeouts timeouts and timeouts is what it causing the troubles
[18:40] <toad_> <thelema> really? the timeouts got broken?
[18:40] <toad_> <thelema> I remember seeing toad doing something to some timeouts.
[18:40] <toad_> <Iakin3> I think more like some kind of link corruption or so
[18:40] <toad_> it's link corruption
[18:40] <toad_> one end times out because the other end died with a corruption error
[18:40] <Iakin3> coulnd't we close gracefully?
[18:40] <Iakin3> (and fix of course...)
[18:41] <toad_> <Iakin3> Maybe one of those new BigInteger(byte[]) -> new BigInteger(1,byte[]) changes or something - I doubt it, since the numbers should NEVER be negative, but maybe. I've tracked it down to an identity transfer getting corrupted...
[18:41] <iip_i2p> <Newsbyte> actually, there must be an equilibrum: if you have to cut a trailer into a zillion chunks, it's rather the question if it's really better then to cut it up in less but bigger chunks.
[18:41] <iip_i2p> <Newsbyte> I mean, at some point, if you make the chunks smaller and smaller, the overhead is the only thing remaining ;-)
[18:42] <toad_> thanks Iakin3, you just blew your node's private ID :)
[18:44] <toad_> <thelema> I think it's a very nice solution.
[18:44] <toad_> <thelema> much better than what I was thinking of.
[18:44] <toad_> it's not
[18:44] <toad_> use a Comparator
[18:44] <iip_i2p> <Newsbyte> he did? I missed it!! Can you post it again? ;-)
[18:44] <toad_> otherwise you'll have BigInt.compareTo(sector) not working
[18:44] <toad_> but sector.compareTo(bigint) does work
[18:45] * tessier_ (~treed@wsip-68-224-172-77.sd.sd.cox.net) Quit ("Client exiting")
[18:45] <toad_> wonderful, now I get to debug iakin's changes too :)
[18:46] <toad_> <KenMan> okay, we need our best brains to focus on routedToChoiceRank and the whole backoff scheme someday soon... - not yet
[18:46] <toad_> we have some big issues to deal with first
[18:46] <toad_> specifically 1) connection negotiation is broken!! MUST FIX NOW!!!
[18:46] <toad_> 2) more work on padding
[18:46] <toad_> probably more
[18:47] <Pascal> <toad_> <Pascal> toad: I wonder if L66-L72 are also in the wrong order - I don't get it... - in the same way L60 & L61 were backwards. Shouldn't message be assigned before going into logging?
[18:48] <iip_i2p> <Newsbyte> Did I hear best brains? did anyone call me?
[18:48] <iip_i2p> <Newsbyte> Oh, you meant *coders* brains
[18:49] <iip_i2p> <Newsbyte> using both of these words like that...isn't that a contradictio in terminis?
[18:49] <Pascal> I've been thinking about a new backoff scheme. How about when I node starts getting overloaded it determines which requests to accept based on the store size of the requester. So if the request comes from a node with a store size of only .25GB, it only has a .25% chance of being accepted. A store size of 5GB would give it a 5% chance. etc.
[18:50] <iip_i2p> <Newsbyte> that said, apart from moving to 128, padding was ok, I thought?
[18:51] <iip_i2p> <Newsbyte> pascal; so, basically, rewarding people with big HD?
[18:51] * The_Saint (~SimonT@66-133-132-188.nas2.roc.ny.frontiernet.net) has left #freenet
[18:51] <wind789> anbody can set the store size to anything they want
[18:53] <Pascal> Newsbyte: reqarding people who contribute to the net
[18:53] <Pascal> Newsbyte: rewarding people who contribute to the net
[18:53] <wind789> I'd rate the mount of data in their store if I was gonna do it that way
[18:54] <iip_i2p> <Newsbyte> besides, with this sheme, 100 GB would give 100% chance...but in the not-too-distant-future, 100GB will probably be commonplace, so everyone will have 100%, thus it will solve nothing, then.
[18:54] * jay (~jay@198.38.10.200) has joined #freenet
[18:55] <toad_> KenMan: What about the KenManHack ? I think if we widely deployed it it might be interesting. ?
[18:55] <Pascal> Newsbyte: I was talking about "when a node starts getting overloaded" not when it was overloaded
[18:55] <toad_> KenMan: Is your MRI really fluctuating THAT much? Changing from 10 minutes to 5 secs and back regularly, not during startup, indicates a BUG.
[18:55] <toad_> thelema: The MRI is included on every packet now.
[18:55] <toad_> <thelema> are current requests too chunky to be controlled by mRI? would mRI work better with more smaller requests? - good point!
[18:55] <Pascal> wind789: thats a good idea too
[18:55] <toad_> <thelema> there's evidence that as a node stays up longer, it picks up more concurrent transfers, which bog down the node, but it doesn't slow incoming load enough to let/make them finish, so it just gets slower and slower. - is there? are you sure? if they are coming in at a reasonable speed then they WILL slow down the requests coming in. is this supported by KenMan's graphs?
[18:56] <toad_> <thelema> we need to get bandwidth limiting working for values over 20K/s - are we ABSOLUTELY SURE that it doesn't work for 20kB/sec+? It seems totally incredible...
[18:56] <toad_> <thelema> not many people in china have DSL. - not many have computers. those who do are students. they have LAN access.
[18:56] <toad_> <KenMan> well, it should be simple enough to permit bw limiting on local interfaces... it is currently avoided however. - we do, there's an option... localIsOK=true.
[18:56] <toad_> <Iakin> iakin.poweruser.org/freenetstuff/bw.jpg
[18:56] <toad_> <Iakin> Thats with 50kbyte/s limiting or so
[18:56] <toad_> hmm, what's the scale?
[18:56] <Iakin3> [01:04] <toad_> wonderful, now I get to debug iakin's changes too :)
[18:56] <Iakin3> pft
[18:57] <Iakin3> I fixed it
[18:57] <iip_i2p> <Newsbyte> hmm...if I set my store size to 200 GB, but I set my firewall not to let almost anything as outgoing...have I then contributed more or less to the network, then someone who has a smaller storesize, but doesn't try to obstruct the dataflow?
[18:58] <toad_> <wind789> are there any plans up raise the current preset upload limit, or set it dynamically based on teh users connection? - if you know a way to autodetect the user's connection, please tell us
[18:58] <toad_> <greycat> What does the "--skipDS" option on "freenet.client.cli.Main put" actually *do*?
[18:58] <toad_> <greycat> Does it only skip the data store for the mapfile key, or for all the chunk CHKs, or both, or neither?
[18:58] <Pascal> Newsbyte: more, because more requests still terminate at your node
[18:58] <toad_> it SHOULD skip it for everything
[18:58] <wind789> send as much junk as you can and see fast it flows out?
[18:58] <toad_> in practice it doesn't
[18:59] <toad_> wind789: regarded as very antisocial...
[18:59] <wind789> it doesn't have to be done often
[18:59] <wind789> I've seen I program that constantly monitors some kinda ping time and constantly adjust the bandwidth to compensate
[19:00] <wind789> emule..
[19:00] <Pascal> wind789: look at messageSendTimeRequest
[19:00] <iip_i2p> <Newsbyte> speaking of b/w limiting...couldn't you make something that makes it obvious and simple for every newbie to set those limits (meaning: NOT by meddling in config and such)?
[19:01] <Iakin3> [01:15] <toad_> hmm, what's the scale?
[19:01] <Iakin3> percentages of 10 MBit/s
[19:02] <iip_i2p> <Newsbyte> I mean, if you look at a myriad of other OS projects, they have a straightforwardmanner to set the limits in the most obvious way. Like shareaza, emule, etc.
[19:02] <toad_> <thelema> is there any way to clean up the implementation of getBucketValue?
[19:02] <toad_> <thelema> It's just painful as is. I told toad that the idea should probably be integrated into the interpolate routine
[19:02] <toad_> <thelema> but he convinced me that it was easier to do it this way.
[19:02] <toad_> uhh, what's the problem with it?
[19:03] <toad_> Newsbyte: yes, the windows GUI has an option to set the bandwidth limitsa
[19:03] <toad_> if you can't use that you probably can't use the web
[19:04] <iip_i2p> <Newsbyte> spoken as a true coder ;-)
[19:07] <toad_> <Pascal> I've been thinking about a new backoff scheme. How about when I node starts getting overloaded it determines which requests to accept based on the store size of the requester. So if the request comes from a node with a store size of only .25GB, it only has a .25% chance of being accepted. A store size of 5GB would give it a 5% chance. etc. - uh, no. :)
[19:07] <toad_> firstly, backoff sucks.
[19:07] <toad_> secondly, how do we tell what their store size is?
[19:08] <toad_> we could base it on the likelihood of their requests succeeding, but I proposed that a while ago and ian didn't like it
[19:08] <toad_> or on their routing utility to us...
[19:09] * greycat (rfc1413@wooledge.org) has joined #freenet
[19:09] <toad_> <wind789> I've seen I program that constantly monitors some kinda ping time and constantly adjust the bandwidth to compensate
[19:09] <toad_> <wind789> emule..
[19:09] <toad_> hmmm
[19:09] <toad_> that MIGHT be a possibility...
[19:10] <jay> wind789: counter strike?
[19:10] <iip_i2p> <Newsbyte> you mean rightklicking on the tray, 'configure', 'advanced settings' and then trying to set those rather oscure numbers out of the blue? Because, I did say: simple and easy for a joe doe newbie, right? I think we fall in the fallacy of thinking that because we (or you - in any case, geeks, coders, or more then average IT-minded people) know what to do, the average newbie-user is going to to know.
[19:10] <toad_> <toad_> Newsbyte: yes, the windows GUI has an option to set the bandwidth limitsa
[19:10] <toad_> <toad_> if you can't use that you probably can't use the web
[19:10] <toad_> <iip_i2p> <Newsbyte> spoken as a true coder ;-)
[19:10] <wind789> jay: sorry havn't played that much
[19:10] <toad_> WELL WHAT DO YOU WANT ME TO DO, NEWSBYTE??
[19:10] <toad_> oh come on, it's not advanced settings
[19:11] <toad_> it's regular settings, right?
[19:11] <toad_> Newsbyte: the fallacy is that the user knows what a byte per second is
[19:11] <toad_> if they do they can use the config
[19:11] <toad_> if they don't, they shouldn't use it
[19:11] <iip_i2p> <Newsbyte> well, make a GUI like on kazaa lite, or shareaza, or emule or a zillion other OS-projects that have far more intuitive ways of setting b/w limits, that even a complete IT-idiot can set :-)
[19:12] <toad_> wind789: one problem would be to have a host to ping
[19:12] <jay> make a gui
[19:12] <toad_> Newsbyte: the most intuitive slider will not help at all if they have no idea what value to set it to
[19:12] <jay> Newsbyte: do you know how diffucult it is to make a real gui?
[19:12] <toad_> if they do know, then they can use the most primitive command line interfaces
[19:12] <toad_> jay: we HAVE a GUI
[19:12] <Pascal> I wonder if we can utilize the dslreports speed test to auto-detect on install the user's connection speed.
[19:12] <jay> Newsbyte: why don't you make one using FCPlib
[19:12] <wind789> sombody mentioned messageSendTimeRequest what does that do?
[19:13] <toad_> Pascal: I doubt itt
[19:13] <toad_> very much
[19:13] <jay> toad_: that's a web based gui
[19:13] <toad_> but perhaps one day with UP&P
[19:13] <toad_> jay: no, we have a proper GUI for windows too
[19:13] <Pascal> what is holding up upnp?
[19:13] <toad_> the web based GUI won't let you change settings
[19:13] <toad_> Pascal: nobody is doing it, that's all
[19:13] <jay> Pascal: and it's totally insecure so far
[19:13] <Pascal> toad: oh, I thought I heard someone was
[19:14] <toad_> also, I don't see why a web based GUI can't be easy to use ... look at google!
[19:14] <Pascal> jay: insecure?
[19:14] <iip_i2p> <Newsbyte> jay: C'mon, let's not exagerate; it's done a multitude of times before, on a dozen other opensource projects. I suspect you wouldn't even have to begin from scratch, with all these examples around.
[19:14] <toad_> google is used by a LOT of users, most of them idiots
[19:14] <toad_> Newsbyte: what does it matter if we have a nice pretty slider if nobody has any idea what to do with it?
[19:14] <greycat> newsbyte: let us know when you're done coding it.
[19:14] <wind789> I'm not sure how emule does it, it might not be a true ping because the times are often below 100ms
[19:15] <Pascal> I've been using another open source java app for some time now that support upnp. I'll look at porting over the code.
[19:15] <toad_> Newsbyte: we need to be able to autodetect the connection speed... the pinging idea is reasonable
[19:15] <jay> Newsbyte: you said Graphical User Interface.. i imagine any program that uses a Graphical environment.. which could be a full-blown insertion and retrieval client or just a config utility
[19:15] <iip_i2p> <Newsbyte> lol...coders code, and I'm no coder
[19:15] <toad_> Pascal: yeah, I know, there's a nice java up&p impl on some app or other
[19:15] <jay> Pascal: there are implentation issues with it (for example) with XP's firewall
[19:15] <toad_> in order to make the current system significantly more user friendly w.r.t. bandwidth we need to either
[19:16] <toad_> a) Auto-detect the bandwidth of the connection, or
[19:16] <toad_> b) Use something like the pinging option
[19:16] <jay> Pascal: i can't paste a link to the article, but it's out there
[19:16] <iip_i2p> <Newsbyte> toad, while I agree auto-detecting is preferable, I recall seen that discussion before, and IIRC, it gave a lot of troubles to implement it
[19:16] <toad_> I don't think a) is easily feasible... we could do a naive implementation but there'd be all sorts of problems with it
[19:17] <toad_> Newsbyte: exactly
[19:17] <Iakin3> [01:21] <toad_> uhh, what's the problem with it?
[19:17] <toad_> if a luser sees a box to configure bandwidth usage, what is he supposed to set it to? he probably has little clue how fast his connection is, and will be happy with the default as long as his conn is 256kbps or more uplink
[19:18] <iip_i2p> <Newsbyte> well...wouldn't it just be easier to make a GUi-ish thingy where you can set the limits?
[19:18] <toad_> Iakin3: huh?
[19:18] <Iakin3> It is more than the recommended 15 lines of code?!
[19:18] <iip_i2p> <Newsbyte> toad, really, that's not true
[19:18] <toad_> Newsbyte: WE HAVE A GUI
[19:18] <toad_> WE HAVE A FUCKING GUI
[19:19] <toad_> the bandwidth limit is not "advanced". if it is that's a bug in the gui and can be rectified - but if it is, then all the options are "advanced"
[19:19] <iip_i2p> <Newsbyte> most internet users, even when no-IT mi?nded at all, will have an idea what to set it to, thanks to popular P2P such as kazaa and the likes
[19:19] <toad_> Newsbyte: so why can't they use the existing GUI?
[19:19] <iip_i2p> <Newsbyte> but not a simple, intuitive one
[19:19] <toad_> the only way a slider would be meaningful is if we can calibrate it by the actual connection speed
[19:19] <toad_> if we don't know that there is zero difference to a box in bytes per second
[19:20] <jay> Newsbyte: do you know what he's talking about? because i don't run windows and haven't used any GUI
[19:20] <toad_> if they're smart, they know what to put there
[19:20] <toad_> if they're not, they're screwed anyway
[19:20] <greycat> you could use a slider with little markings for "modem" "cable (128k)" "DSL (256k)" "DSL (384k)" etc.
[19:20] <toad_> please show me a screenshot of it
[19:20] <iip_i2p> <Newsbyte> for instance, I think shareaza or something similar gives a little list of settings by default, I believe
[19:20] <toad_> greycat: umm, how many Consumers know their UPLINK ?
[19:20] <iip_i2p> <Newsbyte> a drop-down menu, or how's it called in english
[19:20] <toad_> greycat: in any case, the above vary WIDELY
[19:21] <greycat> toad_: if they don't know their uplink, then they can't set bandwidth limits intelligently.
[19:21] <toad_> e.g. you can get cable at 3Mbps here, the default is 512kbps
[19:21] <toad_> greycat: well duh. and most of them don't. and those that do... oh I give up
[19:21] <toad_> if Newsbyte can find somebody to "fix" the GUI, he's welcome to
[19:22] <toad_> the basic problem: there is NO WAY for us to get the bandwidth - not even with a slider
[19:22] <toad_> because broadband is so variable
[19:22] <toad_> we could have a database of all broadband connections in the world
[19:22] <toad_> and we'd need to update it constantly
[19:22] * lolo-laptop (~lostlogic@68.251.84.226) Quit ("Client exiting")
[19:22] * toad_ thinks something based on ping times is much more plausible
[19:23] <greycat> Here's a thought: run the node for a while without any bandwidth limits. Over that period of time, keep track of the actual number of bytes per second that get sent out. On some part of the web interface page, if there is no configured bwlimit yet, then offer a suggestion to the user based on the measured output.
[19:23] <jay> ICMP pings or something application level?
[19:23] <toad_> but we still have to figure out who to ping and what to set the default limits to
[19:23] <toad_> greycat: won't work
[19:23] <iip_i2p> <Newsbyte> actually, I'm not sure if toad and I are speaking about the same ;-) I do see something that indicates it is for limiting b/w (and then you always have the possibility of going to config/ini etc ofcourse), but neither seem remotely comparable with what you find on most current P2P systems today.
[19:23] <toad_> it'd have to be run for aaaaaages
[19:23] <toad_> Newsbyte: SHOW ME!
[19:23] <toad_> GET ME A SCREENSHOT
[19:23] <greycat> how long does it take for a new node to get up to the point where it can saturate a typical 256k uplink?
[19:24] <toad_> I KNOW WHERE YOU LIVE! YOU CAN'T HIDE BEHIND ANONYMITY! GET ME A SCREENSHOT!
[19:24] <toad_> greycat: the typical cable here is 128kbps uplink
[19:24] <jay> Newsbyte: dude if i knew what you were talking about i could help you.. but im not installing any of those programs to find out what you're referring to..
[19:24] <iip_i2p> <Newsbyte> of what, shareaza?
[19:24] <toad_> typical DSL is 256
[19:24] <jay> toad_: haha
[19:24] <toad_> however in other parts of europe it's a lot different
[19:24] <toad_> and in the states it is too
[19:24] <wind789> I'm attemting to find the emule source code used in their dynamic limiter...
[19:24] <toad_> THEY HAVE TO KNOW THEIR UPLINK!
[19:24] <toad_> giving it a fancy GUI won't help if they don't know their uplink
[19:25] <toad_> wind789's comments were probably the most useful in this entire conversation
[19:25] <iip_i2p> <Newsbyte> tsskk...toad!
[19:25] <toad_> use the ping
[19:25] <toad_> not the bandwidth
[19:25] <toad_> don't use explicit bandwidth limits AT ALL unless the user configures them manually and is an expert
[19:25] <toad_> Newsbyte: what part of "they have to know their uplink" doesn't make sense to you?
[19:26] <iip_i2p> <Newsbyte> that's silly...all the other P2P use it, and they are not complaining about it
[19:26] <toad_> if they know that they need to set it to 20kB/sec so it doesn't interfere with all their other stuff they're capable of using a dialog box asking for a limit
[19:26] <toad_> Newsbyte: so show me a comparison of screenshots
[19:27] <toad_> and tell me what it is that makes it so much easier, an order of magnitude easier, to use one than the other
[19:27] <toad_> <wind789> I've seen I program that constantly monitors some kinda ping time and constantly adjust the bandwidth to compensate
[19:27] <toad_> <wind789> emule..
[19:27] <toad_> see, emule doesn't use explicit limits
[19:27] <toad_> explicit limits are for experts
[19:27] <iip_i2p> <Newsbyte> how many times do I have to tell you? They are used to it by now, thanks to all the other P2P progs that DID and DO use a simple intuitive GUI
[19:27] <toad_> who know what their uplink is and what else is running
[19:27] <toad_> Newsbyte: we shouldn't need to make the user configure it
[19:28] <jay> newsbyte you are helpless sometimes
[19:28] <toad_> s/helpless/hopeless
[19:28] <iip_i2p> <Newsbyte> ok, true, I agree with that; auto would be easier still
[19:28] <iip_i2p> <Newsbyte> but, remaining realistic, we all know that that's not going to be for tommorow
[19:29] <jay> auto configuration is easier?
[19:29] <toad_> if they are smart enough to decide "well, I'll need 10kB/sec for emule, and 5kB/sec for kazaa, and I can go up to 30 without big problems... so I can give Freenet 15kB/sec", then they are smart enough to TYPE IN A FUCKING NUMBER
[19:29] <jay> didn't you say that it wouldn't be easy to *code* ?
[19:29] <iip_i2p> <Newsbyte> nono, the word you are looking for is 'hopeless', jay ;-)
[19:29] <toad_> now, if they are not, then giving them a slider won't help
[19:29] <toad_> autoconf will help
[19:29] <jay> no i mean helpless.. you cannot be helped because you do not listen
[19:29] <toad_> no, autoconf can be done easily enough
[19:29] <jay> there's hope still
[19:29] <toad_> based on pings
[19:30] <toad_> as has been repeatedly suggested and ignored
[19:30] <jay> Newsbyte: i myself would like to see screenshots of what gui's you're referring to
[19:30] * toad_ repeats jay's request
[19:30] <iip_i2p> <Newsbyte> well, aparently toad does too
[19:30] <jay> because im a "linux geek/coder/hacker" i don't run windows
[19:30] <jay> heh
[19:30] * TLF (francisco@175.Red-81-40-116.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[19:30] * toad_ resists the urge to paste newsbyte's IRL address to force him to stop hiding behind I2P's "Oh I can't post screenshots without losing my anonymity" :)
[19:30] <greycat> I have a windows box, but I sure as hell don't run windows P2P apps on it.
[19:30] <iip_i2p> <Newsbyte> hasn't anyone here used kazaa light, or shareaza, or emule or sort? geez.
[19:30] <jay> toad_: lol
[19:31] <greycat> Newsbyte: no. Never.
[19:31] <jay> greycat: that's more accurate for what i do as well
[19:31] <iip_i2p> <Newsbyte> well, that explains everything, then
[19:31] <jay> Newsbyte: never
[19:31] <greycat> I've never even run the original Napster. I've used TekNap on Linux, though.
[19:31] <jay> if you listened to us we were saying this before
[19:32] <toad_> Newsbyte: I don't break the law without a hell of a good reason, therefore I have only very rarely run any of them
[19:32] <iip_i2p> <Newsbyte> you know, that would be a good idea to have an idea how other progs/networks do it, you know
[19:32] <toad_> once I ran a gnutella client for a while, I ran edonkey for a longer while
[19:32] <toad_> but that was years ago
[19:32] <toad_> Newsbyte: absolutely. So get us some screenshots. You have the power to do something useful, for once.
[19:32] <jay> he wants us to do all the work
[19:33] <iip_i2p> <Newsbyte> breaking the law? You can use it fully legally you know...certainly if it's only for viewing the settings/GUI :-)
[19:33] <jay> ok ill install all those stupid apps, and deal with the spyware and popups and assorted nonsense for penis pills and the new billy huang cd
[19:33] <jay> just so i can "learn" about kazaa's gui?
[19:33] <iip_i2p> <Newsbyte> jay; that is true. But, it's to the benefit of Freenet, so that shouldn't matter :-)
[19:33] <toad_> Newsbyte: GET THE FUCKING SCREENSHOTS
[19:34] <jay> Newsbyte: while my windows machine comes to a grinding halt, what should i used for Cakewalk Sonar ?
[19:34] <toad_> you have repeatedly implied that you run such things
[19:34] <toad_> so SHOW US
[19:34] <Pascal> http://pascal.rockford.com/azureus.bmp
[19:34] <Pascal> http://pascal.rockford.com/azureus2.bmp
[19:34] <jay> Newsbyte: we're programmers.. we dont have to *use* the program to understand how it works, esp from a configuration standpoint
[19:34] <jay> but all this you wish to ignore
[19:34] <Pascal> no idea if thats what Newsbyte is talking about, but is what I've seen
[19:35] <iip_i2p> <Newsbyte> jay; indeed...that's why you have difficulties imagining how non-coding average joe doe users react to progs and GUIs
[19:35] <jay> when i get Wine running on linux maybe ill install kazaa
[19:35] <jay> Newsbyte: i've written GUI's professionally.. based on specs from the client
[19:35] <toad_> oh nice, a drop down
[19:35] <toad_> still assumes people know their uplink
[19:36] <jay> Newsbyte: i have plenty of experience making GUI's.. in case you haven't noticed im busy getting fcptools/FCPLIb working
[19:36] <Pascal> oops, 2nd is same as first, 1 sec
[19:36] <toad_> LOL, upload speed only counts per torrent!
[19:36] * toad_ can't take it seriously if the limit doesn't even WORK
[19:36] <greycat> Pascal: it's absolutely *trickling* the first .bmp here...
[19:36] <jay> Newsbyte: which specifically doesn't have a GUI.. *and* im willing to entertain your ideas if you help me out a bit and get me screenshots
[19:37] <jay> wow lag on that BMP
[19:37] <greycat> same here. I have approximately the bottommost 1.5 inches of it.
[19:37] <Pascal> there. ignore 2 and get 1 and http://pascal.rockford.com/azureus3.bmp
[19:37] <jay> Newsbyte: since you are the relative noob, you are in the best position to spell out exactly what you mean by an intuitive GUI
[19:37] * greycat aborts it. :)
[19:37] <jay> Newbyte: and since you keep referring to other P2P apps, it seems reasonable to ask you for a screenshot
[19:38] <jay> is that bitmap in freenet or something ?
[19:38] <Pascal> sorry, my router needs a reload
[19:38] <iip_i2p> <Newsbyte> jay, it isn't a contest to prove you can or can not make GUIs. The point I'm making is, that, compared to a lot of other P2P progs, we don't have an easy, intuitive manner to b/w limiting. Even if you were the supreme creator of the best GUI ever made for fcptools, that wouldn't change that fact.
[19:38] <jay> Pascal: ah
[19:38] <Pascal> I was about to do so when I saw this convo
[19:38] <Pascal> should be better now, I killed fred
[19:38] <greycat> Newsbyte: you're going round in circles.
[19:38] <toad_> if you want us to implement a pulldown with "128kbps, 256kbps, 384kbps, 512kbps, 768kbps, 1024kbps..."... sure, that's possible
[19:38] <jay> Newsbyte: are you going to provide me with a screenshot or are you going to keep trolling ?
[19:38] <Pascal> normally QOS in my router limits fred, but it has a memory leak and just ran out of heap
[19:38] <toad_> it's a matter of getting somebody to do it
[19:39] <iip_i2p> <Newsbyte> yes, it seems reasonable. But at the end, toad will still say: "but they need to know their b/w limiting" ;-)
[19:39] <toad_> I fail to see how it is intuitive though
[19:39] <toad_> if Freenet is the only app installed, then it might work, but in that case there's a good chance the user doesn't know their uplink bw
[19:39] <toad_> because they don't know
[19:39] <toad_> and it's not generally advertised
[19:39] <toad_> and they don't NEED to know
[19:40] <toad_> also I've never hacked on the WinInstaller
[19:40] <toad_> I hope eventually we'll be able to eliminate it
[19:40] <toad_> and put everything through the browser
[19:40] <jay> i suspect other apps don't have the same bandwidth issues since they are not trying to provide anonymity and censorship resistance
[19:40] <toad_> WI is a hack, and it's barely maintained
[19:41] <Pascal> why don't we have a java GUI?
[19:41] <Pascal> get rid of WI
[19:41] <jay> Pascal: good question
[19:41] <iip_i2p> <Newsbyte> jay; making a suggestion is not trolling. And if it is, I would be glad to troll. In any case, it's rather late, and I don't feel like starting the progs up and taking screenshots at 2.00 am or later
[19:41] <Pascal> (yes the state, I hate those cheeseheads)
[19:41] <Pascal> :)
[19:41] <jay> Newsbyte: you know how to contact me if you have screenshots to show me
[19:42] <iip_i2p> <Newsbyte> and no doubt other progs are not anonimous and all that, but it still has no bearings on the GUI for setting limits
[19:42] <toad_> LOL jay
[19:42] <iip_i2p> <Newsbyte> I do?
[19:42] <toad_> Newsbyte: the screenshots from Pascal are what you wanted, right?
[19:42] <iip_i2p> <Newsbyte> Well, I know how to contact toad, in any case
[19:42] <jay> Newsbyte: ilnero@gmx.net
[19:43] <iip_i2p> <Newsbyte> I'll try to make some screenshots of some different P2P in tommorow
[19:43] <jay> heh
[19:43] <jay> serious lag on pascal's box
[19:43] <jay> or router
[19:43] <iip_i2p> <Newsbyte> Screenshots from Pascal?
[19:44] <Pascal> jay: should be no lag now
[19:44] <jay> Pascal: i can't get em
[19:44] <toad_> http://pascal.rockford.com/azureus.bmp and http://pascal.rockford.com/azureus3.bmp
[19:45] <Pascal> ok, if your still having problems I'll go ahead and bounce it. brb
[19:45] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) Quit ()
[19:46] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) has joined #freenet
[19:46] <toad_> Newsbyte: http://pascal.rockford.com/azureus.bmp and http://pascal.rockford.com/azureus3.bmp
[19:47] <Pascal> jay: work now?
[19:50] <Pascal> but its still irrelevant if the user doesn't know their speed
[19:50] <Pascal> toad: is there a reason we don't have a java GUI? andthing WI does we can't do in java?
[19:50] <jay> Pascal: nope
[19:51] <jay> Pascal: no; i can't access the screenshots
[19:51] <Pascal> geez, 2.5MB per screen shot.
[19:51] <jay> can't convert to png or jpg?
[19:51] <Pascal> oh well. it doesn't matter. Just shows a simple drop down box in azureus (a java torrent client)
[19:52] <toad_> Pascal: why not just do it on the web interface?
[19:52] <iip_i2p> <Newsbyte> well, yes, it's more or less like that that I meant. things may differ here and there a bit according to the prog, but something like that, yes.
[19:52] <jay> a seperate java-based gui for configuration makes sense
[19:52] <jay> i don't know about a java installer
[19:52] <jay> java-based installer
[19:53] <jay> web-based config makes sense too
[19:54] <Pascal> toad: users are stupid?
[19:54] <toad_> Pascal: we could do a web based config
[19:55] <toad_> it would be much more portable and less nightmarish
[19:55] <toad_> and easier
[19:55] <toad_> and would have some major advantages
[19:55] <toad_> e.g. online updates
[19:55] <toad_> i suppose that's elegance rather than functionalit
[19:55] <toad_> y
[19:56] <iip_i2p> <Newsbyte> maybe we could just take over the defaults setting of (one of) those other P2P? while they won't cover everything, no doubt, they should cover a lot of 'standard' upload/download limits. Heck, even to view a movie one has to indicate the approximate speed, these days. And I'm quite confident a dial-up modem user won't set it on 300 kb/s (certainly not after one time trying ;-)
[19:56] <jay> but a seperate java-based config app would allow you to do things like update the node
[19:57] <iip_i2p> <Newsbyte> It's true, in a sense, that users are stupid (in the sense that they are to some extent ignorant to even basic functiuons). But not THAT stupid as toad seems to think.
[19:57] <toad_> null sunt multipliciam entiam...
[19:57] <toad_> Newsbyte: wherever you have an interface, Murphy's Law applies
[19:57] <toad_> much better to avoid making the user make decisions which are usually based on worse than no data
[19:57] <jay> i'd say that ignorant users are still users, and their ability to use freenet shouldn't depend on their level of understanding
[19:57] <jay> no more than it does any of the other P2P apps
[19:58] <toad_> right
[19:58] <toad_> we need it to JUST WORK
[19:58] <jay> toad_: well yeah
[19:58] <iip_i2p> <Newsbyte> mostly, it has nothing to do with intelligence or anything, just with having to use it (often). B/w limits settings is, for most, not that magical-ignorant anymore since P2P became popular.
[19:58] <toad_> we don't need the user to have to mess around with bandwidth limit settings unless they really want to
[19:58] <jay> before any of this wild gui config app ideas will even work the node and routing have to work
[19:59] <toad_> jay: no, routing is irrelevant, except for the need for the test network. also, bugs are irrelevant. what matters is what newsbyte wants.
[19:59] <jay> and frankly it's not toad's top priority to make it user friendly at this point in time
[19:59] <jay> eheh
[19:59] * toad_ thinks it's a good idea to make freenet user friendly
[20:00] * toad_ doesn't think that implementing a pulldown is the best way to do this
[20:00] <jay> yeah i just mean it's not top on the list
[20:00] <toad_> however if you think differently, MAIL THE LIST
[20:00] <toad_> somebody may be out there willing to hack on the configurator
[20:00] <jay> true
[20:00] <toad_> YOU WILL NEED TO INCLUDE A POINTER TO A SCREENSHOT to demonstrate what you want
[20:00] <iip_i2p> <Newsbyte> well, it's the best way in the shortrun, surely.
[20:00] <toad_> (you could use pascal's ones though)
[20:00] <jay> Newsbyte: do you use Outlook Express to post messages to the list through gmane?
[20:00] * Tafs (~tafs@s01i30-0211.no.powertech.net) has joined #freenet
[20:00] <Pascal> don't post pascal's poor slow server to a mailing list
[20:01] <iip_i2p> <Newsbyte> because I don't see that auto-detect feature working good any time soon
[20:01] <toad_> short run? it'll take as long to implement one of those shitty GUIs (well it'd take ME as long...) as to implement a proper ping-dependant-load thing
[20:01] <iip_i2p> <Newsbyte> call me a pessimist ;-)
[20:01] <jay> Pascal: i got the bitmap finally
[20:01] <Pascal> jay: was it worth it?
[20:01] <toad_> we could in fact simply reduce the limits on messageSendTimeRequest and have almost the same effect...
[20:01] <jay> Pascal: every megabyte of incredibly redundant bitmap data :)
[20:02] <jay> i think i counted 368259 consecutive 'zeroes'
[20:02] <Pascal> toad: I think that should be the default. no bandwith limit, and 500ms messageSendTimeRequest
[20:04] <Pascal> jay: what is the "right" way to take a windows screen shot? I just did a ctrl-printscreen and pasted into paint. It would only let me save as a bmp though.
[20:04] <iip_i2p> <Newsbyte> come, come, toad. Ofcourse I would like to see what I proposed, but the point is, I always give arguments for why I think Freenet is better of with it then without it.
[20:05] <iip_i2p> <Newsbyte> As I recall, it wouldn't be the first of my suggestions that are disregarded at first, while being implemented later on anywayz. (Some didn't, ofcourse, like the searchengine)
[20:05] <oierw`> Pascal: you then feed it to another program other than paint that can save as a jpg
[20:07] <oierw`> or rather, you just paste into that program...
[20:08] <jay> Pascal: yeah paint sucks.. lview will save it as a jpeg
[20:09] <iip_i2p> <Newsbyte> My premisses are rather straightforward: I see it as a benefit to make Freenet more accessible to the average newbie user. In that option, it's better to have simple, easy to understand/implement b/w limiting GUis, then what we currently have. It's that simple, basically. (Not in terms of coding, perhaps, I'll grant you that). It's not that I want to have it my way at any costs, purely because it's comming from me
[20:09] <iip_i2p> <Newsbyte> if it were, I wouldn't agree with you that auto-detecting is the way to go, ultimately
[20:10] <jay> Newsbyte: this really isn't the place to argue development efforts
[20:10] <jay> Newsbyte: it's too real time
[20:10] <toad_> it's easier to auto-detect based on ping
[20:10] <jay> Newsbyte: i would like to test out ClientPut messages with RemoveLocalKey and have it fixed (if it's still broken)
[20:10] <toad_> of course there are technical issues to be debated with that
[20:10] <jay> Newsbyte: which is what this channel is for
[20:10] <toad_> but nobody wants to debate them
[20:11] <iip_i2p> <Newsbyte> I just fear, knowing the time-scales involved before anything gets created/implemnted, that the latter is not going to happen any time soon (to use an euphemism)
[20:11] <jay> the issues should be hammered out on the Dev list and this channel used for realtime communication
[20:11] <jay> debates are pointless here.. most people on the list who would give advice won't see the conversation
[20:11] <toad_> (Some didn't, ofcourse, like the searchengine) - I'd like to have a search engine. i just don't want to have to mess with executing foreign code in fred, for pretty obvious reasons!
[20:11] <jay> even if they're pointed to the irclogs
[20:12] <jay> Newbyte: well we're very far from 1.0
[20:12] <toad_> somebody should implement a nice dropdown selector for bandwidth usage
[20:12] <toad_> sure
[20:12] <toad_> not me
[20:12] <jay> i guarantee that as wel approach 1.0 it will become a top priority
[20:12] <toad_> I don't touch the wininstaller
[20:12] <jay> what's the installer written in?
[20:13] <jay> ive done some installshield in the past
[20:13] <Pascal> http://cvs.sourceforge.net/viewcvs.py/freenet/Contrib/wininstall/
[20:14] <iip_i2p> <Newsbyte> but, again, in the short run, it would be a valuable adition
[20:14] <jay> Pascal: ah nullsoft is ?
[20:14] <iip_i2p> <Newsbyte> infact, in the long run, a client-based js-searchengine would become unwieldy, if you ask me
[20:15] <jay> Newsbyte: there will be no client based js search engine ever
[20:15] <jay> Newsbyte: i am developing FCPLib so that a client-based *anything* can be easily written
[20:15] <toad_> Newsbyte: client based local code execution will be a really bad idea
[20:15] <iip_i2p> <Newsbyte> No news there, jay :-)
[20:15] <Pascal> jay: huh?
[20:15] <jay> Pascal: NullSoft Installer
[20:15] <Pascal> jay: what about it?
[20:15] <toad_> doing it well will either take a monstrous level of maintenance and centralization, or a monstrous amount of effort trying to sandbox it
[20:16] <jay> Pascal: nothing really.. i just commented
[20:16] <toad_> and probably failing (look at all the java applet exploits over the years)
[20:16] <toad_> making a file format and including a client in fproxy for it, I'm all for
[20:16] <jay> any search and/or spider app can be coded with FCPLib
[20:16] <toad_> but nobody wants to work with me... there are very few spiders running... the Freegle guy may get somewhere.
[20:16] <jay> if it can be done in javascript it certainly can be done with a clever FCP app
[20:17] <toad_> jay: which would run cross platform?
[20:17] <jay> toad_: i don't want to touch fproxy
[20:17] * toad_ would prefer that a search client is integrated into fproxy
[20:17] <jay> toad_: fcptools run on most anything afaik
[20:17] <iip_i2p> <Newsbyte> toad, it's like the limitingthingy: sure, it's better if it auto-detects...and sure, it's better if a 'real' anonimous searchengine would work on freenet. But...those are rather unrealistic to expect to work any time soon (probably meaning years)
[20:17] <toad_> but I can't write one unless we have somebody with a spider interested in doing the server side
[20:17] <jay> toad_: BSD, Linux, Win, OSX at least
[20:18] <toad_> Newsbyte: well, running foreign code in Fred is TOTALLY unrealistic
[20:18] <toad_> as I point out above
[20:18] <toad_> and lists of verified recipes would be a maintenance nightmare and a major source of centralization
[20:18] <iip_i2p> <Newsbyte> when it runs on the clientside, it's not that impossible
[20:18] <jay> toad_: i went through great pains to ensure fcptools/fcplib ran on any standard system with a C compiler
[20:18] <toad_> Newsbyte: no, it's impossible
[20:19] <toad_> see java applets
[20:19] <toad_> it may be possible but it will take aaaaages
[20:19] <toad_> jay: it'd be command line then
[20:19] <iip_i2p> <Newsbyte> a list of verified 'code' could easily be limited to a handful
[20:19] <toad_> Newsbyte: fuck off, I'm not doing it
[20:19] <jay> toad_: it's a library.. *any* program that can link to a C library can use it
[20:19] <toad_> jay: you need GUIs for each platform
[20:19] <toad_> anyway if you are searching freesites, doing it in fproxy is logical
[20:19] <iip_i2p> <Newsbyte> I mean, apart from a searchenginecode, what else is there to implemnt in that manner?
[20:19] <Pascal> toad: not if you do it in java
[20:20] <jay> toad_: it's totally possible
[20:20] <toad_> I will write the client if somebody else writes the server/spider end
[20:20] <toad_> Newsbyte: everything anyone wants that would be Really Cool for his site
[20:20] <toad_> the verification load would be such that we would not be able to guarantee safety
[20:21] <toad_> and the search engine code can be incorporated into fproxy
[20:21] <toad_> for the 27th time
[20:21] <toad_> if the freegle guy gets somewhere, maybe we can do something with that
[20:21] <toad_> Pascal: if you do it in java you don't use fcplib
[20:21] <Pascal> so be it
[20:22] <toad_> jay was talking about fcplib
[20:22] <jay> toad_: if you link against fcplib do it for C purposes
[20:22] <Pascal> ah
[20:22] <toad_> Newsbyte: I recommend you post to devl begging somebody to update the wininstaller with a better configuration system for bandwidth
[20:23] <iip_i2p> <Newsbyte> hehe...I am sympathetic to your cause of saying fuck-off to every whining wishlist :-)
[20:23] <toad_> maybe somebody will oblige
[20:24] <toad_> okay, what is to prevent us from using messageSendTimeRequest as the main bandwidth limiting feature?
[20:24] <toad_> if there's a local bottleneck, it'll cause message sends to be a lot slower, right?
[20:24] <KenMan> aw, damn. My testnet node is taking insertions slower and slower. Seems like a bug. The first 74,000 went in with 18 hours, but with another 8 hours it is only up to 74,088 :(
[20:25] <KenMan> CPU use is near 100% ...
[20:25] <toad_> so we just set requestSendTimeCutoff=300ms or something, right?
[20:25] <Pascal> toad: I think that should be the default. no bandwith limit, and 500ms messageSendTimeRequest
[20:25] <KenMan> make it adjustable , maybe ?
[20:25] <iip_i2p> <Newsbyte> hmm...well...I proclaim to be a profesional begger for freenet, so I could do that
[20:26] <toad_> KenMan: when I ran a testnet, it wouldn't connect the nodes up
[20:26] <toad_> that's what I was working on
[20:26] <KenMan> yeah, that's another bitch I have...
[20:26] <iip_i2p> <Newsbyte> good!
[20:26] * KenMan keeps it to himself...
[20:26] <iip_i2p> <Newsbyte> I love to see work on testnetworks :-)
[20:26] <toad_> now, will messageSendTimeRequest approximate ping times given plenty of CPU?
[20:27] <KenMan> Newsbyte - I love to see testnets that work :p
[20:27] <Pascal> but maybe negotiations should be fixed first :)
[20:27] <toad_> well, somebody needs to try it
[20:27] <Pascal> toad: yes. fred never gets near taxing my cpu and I get high MSTRs
[20:27] <toad_> on a node with a normal link
[20:27] <iip_i2p> <Newsbyte> hmm, c'mon, even toads' little testnetwork has already shown it's worth
[20:27] <toad_> I have the KenManHack enabled, and besides, my node has heavy logging
[20:28] <toad_> Pascal: and your link is saturated?
[20:28] <Pascal> yes
[20:28] <toad_> when you get high MSTRs?
[20:28] <toad_> cool
[20:28] <toad_> please try it out
[20:28] <Pascal> then it drops back down
[20:28] <Pascal> try what out
[20:28] <toad_> set in the config file requestSendTimeCutoff=500
[20:28] <Pascal> thats how I run
[20:28] <KenMan> just FYI, 5086 only manages to have 61 peers, yet qph has been 6K-10K /hr today :(
[20:29] <Pascal> been working fine
[20:29] <Pascal> till me router runs out of ram :)
[20:29] * KenMan hasn't visited stable for a few months... maybe 5K/hr is normal ??
[20:30] <toad_> okay, it's rigged for overloadLow = requestSendTimeCutoff, 1.0 = successTimeCutoff
[20:30] <toad_> errr successfulSendTimeCutoff
[20:30] <toad_> so you'd need to set that too
[20:30] <iip_i2p> <Newsbyte> if I may give some further criticism (Sanity isn't there, I hope? ;-), my stable node, after improving a bit when I reseeded it, has again fallen down considerably
[20:30] <Pascal> hmmm...try that again?
[20:31] * toad_ recommends requestSendTimeCutoff=200, successfulSendTimeCutoff=500ms
[20:31] * KenMan has been suspecting that toad enabled KenManHack for a few weeks now, without telling anyone...
[20:31] <toad_> KenMan: no, if I did that, then routing would work better
[20:31] <toad_> :)
[20:31] <jay> i heard that KenMan Linux is actually a r00tk1t
[20:31] <KenMan> well, remove my name from it before you make that claim !!
[20:32] <KenMan> jay - you listen to too many people then
[20:32] <jay> heh
[20:32] <iip_i2p> <Newsbyte> infact, it's like as if it's back to it's old (non-working) ways
[20:32] <toad_> Pascal: please try out the above settings for a while then report back to devl
[20:32] <toad_> if it works we can set it
[20:32] <jay> KenMan: im running it now.. works great ;)
[20:32] <iip_i2p> <Newsbyte> statusbar at 4%, no activelinks visible
[20:32] <Pascal> toad: will do. setting now
[20:33] <toad_> also it'd be interesting to set requestSendTimeCutoff=100, successfulSendTimeCutoff=200, and log ping
[20:33] <toad_> pings for a while
[20:33] <toad_> for people who want really low pings e.g. gamers that might be an option
[20:33] <KenMan> jay - so you are the mysterious operator who has been screwing my node for the past month. Now things are starting to make sense :)
[20:33] <Pascal> just noticed I have %doRequestTriageBySendTime=true in my config file twice for some reason
[20:33] <jay> KenMan: ive been ejecting your cdrom for weeks now
[20:34] <jay> toad_: when u say ping, you mean icmp ping or a freenet 'ping'
[20:34] <KenMan> damnit all, can't you use the videocam to sense when I need to rest my beer ? you've knocked it over twice now :(
[20:34] <toad_> or even lower limits, but we need some empirical research on that...
[20:34] <jay> eheh
[20:34] <toad_> jay: well, we can't do an ICMP ping as part of loads
[20:34] <toad_> load
[20:34] <toad_> I think messageSendTimeRequest approximates it though
[20:34] <jay> so it's app level then
[20:34] <toad_> I meant comparing messageSendTimeRequest to actual ICMP pings empirically
[20:35] <iip_i2p> <Newsbyte> I do shut the node of for the night...but still...it shouldn't become completely numb after some hours of non-activity? Or am I wrong?
[20:35] <jay> i wonder if COD and those online games do actual pings
[20:35] <toad_> to determine whether we can reliably get freenet's usage low enough to make gaming feasible
[20:35] <toad_> I doubt it, spikes will be too high...
[20:35] <jay> ill sniff the connection next time my roomate is playing call of duty online
[20:35] <toad_> I don't know how it would be possible to fix THAT
[20:36] <Pascal> whats is we set TOS on freenet packets to 0
[20:36] <Pascal> what if we set TOS on freenet packets to 0
[20:36] <toad_> uhhh
[20:36] <KenMan> localIsOk=true does not enable bw limiting on local interfaces, it just permits their use as an interface
[20:36] <toad_> isn't it largely ignored?
[20:36] <toad_> KenMan: yes, there's another setting that limits everything
[20:36] <Pascal> then when the users complains they can't game with fred we tell them to buy a decent router
[20:36] <toad_> surprisingly enough it's called limitAll
[20:36] <KenMan> ahhh
[20:37] <Pascal> toad: can't hurt right?
[20:37] <toad_> Pascal: you have to worry about incoming too but generally I'm not convinced that OS level bandwidth limiting will work well for us
[20:37] <toad_> also how do you set TOS on a TCP connection?!
[20:37] <toad_> anyway
[20:37] <toad_> it's been ummm interesting talking to you all
[20:37] <jay> in C i think you call set_socket_option or something
[20:37] <KenMan> just flip those flags that all current hardware currently ignores, yeah.
[20:38] <toad_> I still haven't got any REAL work done i.e. debugging, but maybe I wouldn't have got far as I am quite tired
[20:38] <jay> the problem is that you're in irc
[20:38] <toad_> tomorrow maybe I fix connections
[20:38] <jay> no real work gets done here in multi-user notepad
[20:38] <KenMan> yup, just stay up until 2:00 and the talk gets really lively then !!
[20:38] <toad_> no, there are other reasons
[20:38] <jay> yeah it's only 9PM here
[20:38] <toad_> local ones
[20:39] <toad_> bbl
[20:39] * KenMan senses siblings nearby
[20:39] <Pascal> Exception in thread "main" java.lang.ClassCastException
[20:39] <Pascal> at freenet.config.Params.getInt(Params.java:350)
[20:39] <Pascal> at freenet.node.Node.init(Node.java:3556)
[20:39] <Pascal> at freenet.node.Main.main(Main.java:471)
[20:39] <toad_> Pascal: take out the "ms"
[20:39] <toad_> :)
[20:39] <KenMan> UUH - maybe toad has a chick in his bedroom ?? 8)
[20:39] <Pascal> d'oh
[20:40] <jay> it's some sort of "local" reason.. perhaps there was a tornado eariler
[20:40] <Pascal> toad: last time I cut n paste what you say :)
[20:40] <jay> that would tire me
[20:40] <toad_> no, I went to aikido, then came back, brat was being bratty, and he needed to talk to the police some hours later so I didn't want to go to bed until afterwards as he wanted it to be very very very private
[20:40] <KenMan> naw, he's got a woman in his bed, and he doesn't know what to do...
[20:40] <toad_> LOL KenMan
[20:41] <KenMan> just kick her out, tell her to come back tomorrow after you are fully rested. Don't sweat it.
[20:41] <jay> yeah chicks like being rejected for computer reasons
[20:41] * toad_ doesn't think that raising kids on his current financial situation would be a good idea
[20:41] <jay> it's sexy
[20:41] <iip_i2p> <Newsbyte> well...it's getting late, I'm off to sleep. I'll let the node run, this time. And Ian can say what he wants, there IS something wrong with my build. I've never had such unresponsiviness before, not in to that extent. It seems it begins rather promising, but then dwindles down to a grind...idem after reseeding. It just seems to get less good in routing or finding data, instead of better. Very strange. Maybe I should run it more dor longer per
[20:41] <jay> toad_: who's Brat?
[20:42] <toad_> Newsbyte: I'm sure there are many problems...
[20:42] <toad_> not least are the connection issues
[20:42] <KenMan> toad's lawless little bro'
[20:42] <jay> ah
[20:42] <jay> a young rebellious tadpole eh?
[20:42] <toad_> Newsbyte: per day?
[20:42] <Pascal> http://www.javafaq.nu/article460.html
[20:43] <iip_i2p> <Newsbyte> what per day?
[20:43] * toad_ barfs at that url
[20:43] * toad_ will have absolutely nothing to do with charging-for-qos
[20:43] <iip_i2p> <Newsbyte> I'm talking about the new build (5085)
[20:43] <toad_> Newsbyte: the above statement from you ended "Maybe I should run it more dor longer per"
[20:43] * guido^pe (~unknown@dsl-213-023-243-038.arcor-ip.net) has joined #freenet
[20:43] <toad_> you probably meant per day, right?
[20:44] <toad_> Pascal: isn't the default 0?
[20:44] <Pascal> 4
[20:44] <Pascal> ICMP default is 0
[20:44] <toad_> hmmm whatever
[20:45] <toad_> implement it (as a config option) if you want
[20:45] <iip_i2p> <Newsbyte> "Maybe I should run it more for longer periods, I dunno"
[20:45] <toad_> Newsbyte: what periods do you run it for?
[20:46] * guido^pe (~unknown@dsl-213-023-243-038.arcor-ip.net) Quit (Remote closed the connection)
[20:46] <iip_i2p> <Newsbyte> Well, the 5085 I ran for say, 5-6 hours a day
[20:46] <toad_> okay well
[20:46] <toad_> there's lots to fix
[20:46] <toad_> lots of more or less immediate stuff to fix
[20:47] <toad_> after that, perhaps a network reset from scratch will help
[20:47] <jay> toad_: indeed.. first you need to fix the gui configurator
[20:47] <toad_> but lets take it one step at a time
[20:47] <toad_> LOL jay
[20:47] <jay> eheh
[20:47] * toad_ goes to bed
[20:48] <iip_i2p> <Newsbyte> but actually...it's all crap...this has nothing to do with not being connected well enough. I know how that goes, on a node. It's not like this, where I get absolutely nothing. I think this is another problem.
[20:48] <jay> Mewsbyte: leave it on for at *least* a day straight and then judge it
[20:48] <jay> if it still sucks then say so afterwards
[20:48] <iip_i2p> <Newsbyte> at least untill 8084. When I restarted my node, it could take 10 minutes of waiting and refreshing, but slowly more and more activelinks became visible
[20:49] <jay> Newsbyte: but your post to support recently after using the stable build for a few hours doesn't help
[20:50] <jay> Newsbyte: i recall the older stable build working (around 5028 even) for reasons that wouldn't scale
[20:50] <iip_i2p> <Newsbyte> point being, toad, then when I first started it, it worked better then after some hours. At the end it was total crap (remember the post on devl ;-). After I reseeded it, it was better again...but now it's again totally broke, it seems.
[20:50] <jay> it's like 2am where he is.. let em sleep
[20:50] <iip_i2p> <Newsbyte> that is not normal behaviour, even not for a not well-connected node
[20:50] <iip_i2p> <Newsbyte> I think
[20:51] <jay> Newsbyte: it's not normal relative to the classical routing algorithm on the older stable builds
[20:52] <jay> so now the newer code attempts to make decisions that will scale, but it isn't working well
[20:52] <jay> the estimators perhaps.. i dunno..
[20:52] <jay> but it isn't totally broken
[20:52] <KenMan> 5086 is just hosed right now. But I will try to gather some details before doing my rant...
[20:52] <iip_i2p> <Newsbyte> come, come...classical routing was abandonned long before 8084, so it can't be purely that
[20:53] <KenMan> let it be Newsbyte, it will probably get better. And if it gets worse, it will get more attention !!
[20:53] <jay> kenman: the situation now is that the node needs to route and learn about good peers as opposed to bad ones
[20:53] <jay> err i meant that -> Newsbyte
[20:54] <KenMan> I'm backing up NB by saying it runs like crap, but since I haven't run stable for a while, this may be the norm ...
[20:54] <jay> Newsbyte: so leave it on for a day and then be critical
[20:54] <KenMan> yes, I'm giving mine 2 or 3 days before I can figure out what the heck is most broken... i think it gets better, but slowly !
[20:55] <jay> some of the bugs recently have been in the node for months
[20:55] * Hory (~Miranda@81.196.25.110) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[20:55] <jay> the insert bug for one, then the recent one from Dev have been in there who knows how long
[20:55] <KenMan> i am a little disappointed that I don't have more peers by now, but some things appear to be getting better.
[20:55] <jay> yeah the peer number is slow for me as well
[20:56] <KenMan> i just hope that new bugs are being exchanged for old ones ;)
[20:56] <KenMan> are NOT being exchanged
[20:56] <iip_i2p> <Newsbyte> hmm...but I did leave it on for 12 hours, and it doesn't explain how it deteriorates instead of improving. I'm willing to let it on longer, no rpob, I'm just postulating my current ptoblems seem not to be related to the normal woes one would expect, if it was purely a mater of my node needing to get bettr connected.
[20:56] <jay> Newbyte: how do you read the mail from the Dev list may i ask ?
[20:56] <KenMan> expect to have 'different
[20:56] <KenMan> behavior is all we are saying...
[20:56] <jay> KenMan: heh true.. these bugs were pretty much easily fixed.. once they were found
[20:57] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) Quit ()
[20:57] <jay> Newsbyte: do you use Outlook Express?
[20:57] <toad_> oh, old bugs ARE being fixed... and new bugs ARE being created :)
[20:58] <iip_i2p> <Newsbyte> So, even while it could be that 2x 6hours it to few, it begs the question if Freenet has to run continiously for it to work properly. If it is, that will bode badly in regard to the average internet-user.
[20:58] <jay> that's the nature of software development
[20:58] <jay> Newsbyte: not true.. especially if Freenet runs on cell phones which are left powered on 24/7 no problem
[20:58] <iip_i2p> <Newsbyte> I'm not subscribed, if that is your question. I look at the archives.
[20:58] <toad_> btw re girls... "I don't believe in cars, I live with my parents and I work from home from the generosity of various maladjusted perverts and crypto geeks on the internet"
[20:58] <toad_> maybe not the best line ever :)
[20:59] <jay> Newsbyte: if you used Outlook Express to read News from GMane, you could reply to posts and have them threaded properly
[20:59] <toad_> <jay> Newsbyte: not true.. especially if Freenet runs on cell phones which are left powered on 24/7 no problem - well, give it 30 years... :)
[20:59] <jay> toad_: heh who knows..
[20:59] <iip_i2p> <Newsbyte> and where do I come in, toad? ;-)
[20:59] <toad_> jay: i left out the best bit :)
[21:00] <toad_> bbl
[21:00] <iip_i2p> <Newsbyte> 'maladjusted perverts' - 'crypto geeks' /me feels left out ;-)
[21:00] <jay> im maladjusted but in a good way
[21:00] <toad_> :)
[21:01] <iip_i2p> <Newsbyte> cell phones? Freenet?
[21:01] <toad_> feel free to compile a longer list :)
[21:01] <iip_i2p> <Newsbyte> I ,*was* talking about the 21st century, jay ;-p
[21:01] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) has joined #FREENET
[21:01] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) Quit (Remote closed the connection)
[21:01] <jay> Newsbyte: a stretch i admit
[21:01] <jay> but certainly something worth working towards
[21:01] <jay> that's our 24/7 platform right there
[21:01] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) has joined #FREENET
[21:02] <jay> minus bandwidth and memory concerns
[21:02] <toad_> jay: what makes you think that user programmable cell phones will ever be commonplace?
[21:02] <jay> Newsbyte: as far as i know they run winny in Japan on their cellphones
[21:02] <toad_> or vaguely affordable?
[21:02] <jay> toad_: J2ME runs on them already
[21:02] <jay> toad_: i don't think it's very far off
[21:03] <toad_> sure, but they'll be super-PDAs
[21:03] <jay> most people live out of their cellphones
[21:03] <toad_> you can buy them now
[21:03] <jay> well yeah.. that's where cell phones are evolving
[21:03] <toad_> 99% of the population just wants voice and text
[21:03] <iip_i2p> <Newsbyte> yes, well *cough* being a bit more near-future-realistic; most people are not letting their puter run non-stop, not even with broadband. So Freenet better work when puters constantly fall off and gett added (as it was supposed to, I believe).
[21:03] <toad_> and keep the price down
[21:03] <jay> Newsbyte: true
[21:03] <toad_> Newsbyte: that
[21:03] <toad_> that's pretty damn hard
[21:03] <toad_> but we do try
[21:04] <toad_> I don't see why broadband connected PCs need to be turned off regularly though
[21:04] <jay> toad_: noise and power
[21:04] <iip_i2p> <Newsbyte> I might agree
[21:04] <toad_> jay: power's an issue. noise is only an issue if it lives in your bedroom.
[21:04] <iip_i2p> <Newsbyte> but, for instance, I'm pretty sure my mom would disagree ;-)
[21:04] * Tafs (~tafs@s01i30-0211.no.powertech.net) Quit (Connection timed out)
[21:04] <jay> toad_: well that's an issue for me and will be as long as i live in new york
[21:04] <toad_> which is only an issue if you live with your parents, or have an absurdly small apartment
[21:05] <iip_i2p> <Newsbyte> Luckily, *I* don't live with my parents anymore
[21:05] <iip_i2p> <Newsbyte> hehe ;-p
[21:05] <jay> ideally i would like a linux based router somewhere else so i don't have to listen to the fan
[21:05] <toad_> well, I'll talk to you all tomorrow. Newsbyte is in GMT+2 iirc at the moment, so it should be ~ 1:23
[21:05] <jay> and run all kinds of useful stuff on it including freenet
[21:06] <toad_> jay: I have my main workstation sitting in a closet under the stairs
[21:06] <iip_i2p> <Newsbyte> seriously, though, I think most people, especially the bit elderly, just look at that as a waste of money (current)
[21:06] <toad_> Newsbyte: not if they have Frost constantly downloading whatever it is
[21:07] <iip_i2p> <Newsbyte> I doubt that's that much of an incentive
[21:07] <toad_> one thing that scares me is that if freenet is ever popular, 99% of our nodes will run on Windows XP
[21:07] <jay> of course
[21:07] <toad_> whether anonymity means anything in such circumstances is debatable
[21:07] <iip_i2p> <Newsbyte> I have friends my age that DO see the benfit of some P2P progs, but even they let it seldom on for 24/24
[21:08] <iip_i2p> <Newsbyte> If I wasn't running some servershit, even I wouldn't probably
[21:08] <toad_> Newsbyte: freenet is servershit
[21:08] <toad_> but I don't know what we can do to make short lived nodes more useful
[21:09] <iip_i2p> <Newsbyte> we're just a people that don't like to let things waste energy, when we're not around to watch it, I guess ;-)
[21:09] <toad_> nah, most people are pretty lazy about that sort of thing
[21:09] <toad_> it's more the noise
[21:09] <toad_> and the fire risk
[21:09] <toad_> and the worry of one's toddlers fiddling with it
[21:09] <jay> definitely noise
[21:09] <toad_> and so on
[21:09] <iip_i2p> <Newsbyte> I differ in opinion on that, but let's say it's both
[21:09] <jay> i can't really practice music with the fan noise in the background
[21:10] * oierw` is now known as oierw
[21:10] <jay> it's like constant static
[21:10] <toad_> jay: displace the computer then
[21:10] <toad_> as I have done
[21:10] <jay> i live in a one-room apartment
[21:10] <toad_> ouch
[21:10] <jay> very small
[21:10] <iip_i2p> <Newsbyte> I rather got used to it
[21:10] <toad_> cost of living in Gotham City, I suppose
[21:10] <jay> i don't mind living in a coffin really
[21:10] <jay> toad_: heh
[21:10] <jay> there is a Gotham in manhattan actually
[21:11] <toad_> London's the same
[21:11] <toad_> you can buy a parking space for $XXX,XXX
[21:11] <KenMan> toad MUST have a buxom blonde in his bed. He simply cannot leave the keyboard !!
[21:11] <iip_i2p> <Newsbyte> though I wouldn't mind some puter with a zalman fan in it, or something
[21:11] <toad_> well, there are quiet-ish cooling systems
[21:11] <jay> i just want a closet to put a puter in
[21:12] <toad_> the big problem is that CPUs just use up sooo much power...
[21:12] <jay> when my income increases things will change in that area
[21:12] <toad_> they're really hard to cool
[21:12] <KenMan> this desktop case with a zalman 'copper flower' is just wonderful. Seagate HD, I can't hear a darned thing.
[21:12] <KenMan> oh, and zalman PS is quiet too
[21:12] <toad_> does it work? I heard the cooling performance was pretty dire...
[21:13] <KenMan> it works great. Great enough anyway.
[21:13] <toad_> even with 100% cpu usage around the clock?
[21:13] <iip_i2p> <Newsbyte> yeah, heard great things of the 'silence' products of zalman
[21:13] <KenMan> It is mounted on an athlon XP 1600+ , 68 watts of heat...
[21:13] <iip_i2p> <Newsbyte> didn't they have a passive cooling for even the powerthingy as well?
[21:13] * toad_ has tons of cooling, and it still gets worryingly hot sometimes...
[21:13] <KenMan> about the same wattage as the 2600+ making all the noise over in the corner :(
[21:14] <toad_> passive cooling isn't going to keep it cold enough unless you have some big chassis fans
[21:14] <jay> i have an athlon xp 2.5
[21:14] <iip_i2p> <Newsbyte> imagine having really ALL fans gone and replace with passice cooling....you couldn't get much more silent then that
[21:14] <toad_> otoh they can be big, slow, quiet
[21:14] <KenMan> zalman is cool , but that darned "Reserator" has been on my mine ever since Iakin put it there...
[21:14] <toad_> Newsbyte: easy, just get something like an Eden
[21:14] <toad_> or whatever they called it
[21:14] <iip_i2p> <Newsbyte> toad, I think you are mistaken
[21:14] <toad_> VIA's cheap low-power no-fan X86 system
[21:15] <toad_> fully integrated
[21:15] <KenMan> i think toad is not mistaken
[21:15] <toad_> more or less
[21:15] <toad_> and pretty slow
[21:15] <KenMan> C3 (used to be cywix) may be the same processor
[21:15] <toad_> C3 is an earlier version
[21:15] <toad_> Eden ~= C5
[21:15] <toad_> more integrated, slightly faster
[21:15] <iip_i2p> <Newsbyte> I definately recall passive cooling for the poweroutletthingy
[21:15] <toad_> less power
[21:15] <KenMan> yeah. I could deal with that. Fred probably could too.
[21:15] <toad_> crap floating point
[21:15] <iip_i2p> <Newsbyte> and for the CPU, it exists even longer
[21:15] <toad_> definitely not for gaming
[21:16] <toad_> Fred increasingly uses quite a bit...
[21:16] <toad_> still a 700MHz Cyrix probably could run a node acceptably
[21:16] <KenMan> I want quiet stable computing that let's me read my email and slashdot. Maybe some emacs for programming. And if it can run freenet too, great.
[21:16] <toad_> (of course there's no optimized GMP for it..)
[21:16] <iip_i2p> <Newsbyte> it wasn't probably meant for the latest-and-greatest game-playing PCs, I dunno
[21:16] <toad_> (so it'll be slow...)
[21:16] <toad_> Newsbyte: that's the general idea, yeah
[21:17] <iip_i2p> <Newsbyte> but then again, a simple webserver doesn't need that
[21:17] <toad_> sometimes
[21:17] <toad_> we can probably improve further on freenet's cpu usage
[21:17] <toad_> iakin's work has been quite significant
[21:17] <iip_i2p> <Newsbyte> hooray!
[21:17] <toad_> GCJ would probably help a bit too
[21:17] <iip_i2p> <Newsbyte> long live iakin!
[21:18] <toad_> 5087 will probably have the nativebigint stuff
[21:18] <toad_> which will reduce cpu usage significantly
[21:18] <iip_i2p> <Newsbyte> wait...did he die yet? 'the king is dead, etc)
[21:18] <toad_> on supported platforms
[21:18] <toad_> with an updated freenet-ext.jar
[21:18] <iip_i2p> <Newsbyte> native...cool
[21:18] <toad_> otoh maybe that'll wait for 5088, or 5089.. lots of bugs to fix first
[21:19] <iip_i2p> <Newsbyte> and that after choosing a multi-platform thingy like java ;-)
[21:19] <toad_> only for x86 at the moment (and not all of them)
[21:19] <toad_> Newsbyte: well, if Sun would only open source the JVM, then we'd just use GMP, and it'd be even faster
[21:19] <toad_> and portable too
[21:19] <jay> i haven't gotten a sig 11 for at least a few days
[21:20] <toad_> as it is we have to compile our own platform specific GMP wrappers
[21:20] <jay> toad_: i think we might see more sig 11's in the near future.. something crept in that causes them occasionally
[21:20] <toad_> if Sun open sourced java, Sun could use GMP itself
[21:20] <iip_i2p> <Newsbyte> kaffe?
[21:20] <toad_> jay: JVM bug...
[21:20] <toad_> Newsbyte: or GCJ
[21:20] <toad_> I believe GCJ runs reasonably on Windows
[21:20] <toad_> so we need to look at that too
[21:20] <jay> toad_: i know.. something is triggering it im saying.. which is *real* tough to search out i realize
[21:20] <KenMan> we should have a 386-compatible GMP, it should still be a decent speed up for many platforms...
[21:20] <toad_> need to get together with Robilad
[21:20] <toad_> jay: we can't
[21:21] <toad_> simply can't
[21:21] <toad_> I'm not going to even try to debug sun's code, it's totally futile
[21:21] <toad_> just try a few different JVMs
[21:21] <jay> toad_: it's been done.. im not saying you should do it
[21:21] <KenMan> i think scott mcnealy should do it himself. show them dumb engineers a trick or two
[21:21] <toad_> jay: it's illegal to modify and recompile iirc
[21:22] <jay> toad_: i mean there's java bytecode that goes down the codepath that causes a sig 11 somehow
[21:22] <toad_> even if you have the source, which apparently is available somewhere, as the BSDs compile their own Sun VM...
[21:22] <toad_> well, it's NOT OUR PROBLEM
[21:22] <KenMan> exactly
[21:22] <jay> toad_: well it's a problem.. didn't say whose it was
[21:22] <toad_> if necessary we will get it working on Kaffe/GCJ, and then it will be our problem
[21:22] <iip_i2p> <Newsbyte> well, I'm gone, really. If I stay any longer, I could aswell fogo with sleep. But then I'd be almost like a coder ;-)
[21:22] <jay> toad_: where can i get the source?
[21:22] <toad_> but in the meantime, please find which JVMs it happens on
[21:23] <jay> toad_: happens with my Sun 1.4.2_05
[21:23] <KenMan> count 1.4.2_03 in that group
[21:23] <iip_i2p> <Newsbyte> see you laterz (will be idling)
[21:23] <toad_> jay: try 1.5.0-b2
[21:23] <toad_> also try 1.4.1
[21:23] <KenMan> night NB
[21:23] <jay> toad_: can do
[21:23] <toad_> okay
[21:23] <jay> Newsbyte: lates
[21:23] <toad_> jay: please report to devl when done
[21:23] <jay> toad_: i did a couple of days ago
[21:23] <KenMan> toad, what did your little bro do this time ?
[21:24] <toad_> KenMan: tomorrow I'll want to know whether you have any idea what causes the simulation not to go to 100%
[21:24] <toad_> well sometime soon
[21:24] <KenMan> he simply must stop bringing the cops around for dinner
[21:24] <toad_> KenMan: not his fault now
[21:24] * toad_ just doesn't like going to bed and feels hot
[21:24] <toad_> but bbl zzz
[21:24] <KenMan> play some new games then ! Get a demo somewhere.
[21:25] <toad_> heh
[21:25] <KenMan> Night ribbit toad ...
[21:25] <jay> which reminds me i have to reinstall nascar 4 and superbike 2001 later
[21:25] <jay> or '02 i think
[21:25] <KenMan> the early nighttime here in baltimore is beautiful. Quiet, the crickets are going, even a few birds are still rustling after dark.
[21:26] <KenMan> no screaming kids or anything ... seems like a good time for a beer or three.
[21:26] <KenMan> temperature is perfect. There is a slight breeze to keep it nice...
[21:27] <KenMan> toad needs a nice big fan in his bedroom, that always helps me to sleep.
[21:28] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) has joined #freenet
[21:28] <KenMan> i have been thinking about buying a dehumidifier - they are almost like having a small A/C unit !!
[21:28] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) Quit ("Leaving")
[21:28] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) has joined #FREENET
[21:28] * thelema|aikido is now known as thelema
[21:28] <thelema> grr, I just missed toad
[21:29] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) has left #freenet
[21:29] <KenMan> but i have also been thinking about getting a job to pay for things like zalman coolers and dehumidifiers...
[21:29] <toad_> :)
[21:29] <toad_> thelema: hi
[21:29] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) has joined #FREENET
[21:29] <toad_> thelema: speak tomorrow ok?
[21:29] <toad_> a bit earlier
[21:29] <toad_> like 5 hours ;)
[21:29] <toad_> or more :)
[21:29] <thelema> I was around 5 hours ago...
[21:30] <thelema> but I'll see what I can do to see you here in 18 hours.
[21:30] <KenMan> don't worry thelema - i'll probably keep him busy... if i can.
[21:31] <jay> 20 route not founds in a row
[21:31] <jay> i think my node froze before
[21:31] <KenMan> i'll raise you another 15 RNFs then...
[21:32] <KenMan> and toss in a DNF, just to keep it interesting.
[21:32] <jay> they're all being "backed off"
[21:33] <KenMan> yeah, my 5086 is for sheets. Backed off at about 100% more than 50% of the time :(
[21:33] <jay> im running unstable
[21:33] <jay> 60170
[21:34] <KenMan> so we can share the joy, ?
[21:34] <thelema> jay: lots of local requests?
[21:34] <jay> thelema: im trying to insert a splitfile
[21:34] <jay> haven't requested anything
[21:35] <jay> retrieved even
[21:36] <KenMan> I'm not doing ANY local requests. This is 5086 after being up for about 24 hours : http://mywebpages.comcast.net/jkcorson/FreeNet/5086c.png
[21:37] <KenMan> the query count is just jumping all over the place :(
[21:37] <KenMan> but I don't know if my outbound MRI's are encouraging it, or it is due to some external event(s) ...
[21:38] <thelema> jay: lots of simultaneous inserts are the same.
[21:38] <KenMan> for comparison - http://mywebpages.comcast.net/jkcorson/FreeNet/60167c.png fluctuates a good deal too...
[21:39] <toad_> KenMan: there's diagnostics for outbound MRIs
[21:39] <toad_> they're reasonably well averaged
[21:39] <toad_> please show them on the graphs
[21:39] <KenMan> this is a snapshot of better times... http://mywebpages.comcast.net/jkcorson/FreeNet/60139c.png
[21:39] <jay> thelema: it's not a multi-threaded insert.. one block at a time
[21:40] <KenMan> oh hell. Now I need to figger out how to log MRIs again, without using my own log messages. I had bad luck last time, so maybe I'm due for some good luck. I can practice with a testnet node first, before recycling the stable node.
[21:40] * toad_ (toad@82-32-16-91.cable.ubr03.azte.blueyonder.co.uk) Quit ("using sirc version 2.211+KSIRC/1.3.10")
[21:40] * toad_ (toad@82-32-16-91.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[21:41] * toad_ grrr, just lost all my log... oh well
[21:41] <toad_> only because I was impatient
[21:41] <toad_> anyway KenMan: there's a diagnostic var
[21:41] <toad_> just grab that
[21:41] * jay (~jay@198.38.10.200) Quit ("bbl")
[21:41] <KenMan> really ? I think i saw it recently. Cool.
[21:41] <toad_> outgoingRequestInterval
[21:42] <KenMan> yeah, thanks! now go take a walk and then go to bed (mild exercise will help).
[21:42] <toad_> mine is reasonably stable... hrrm, and disturbingly high
[21:42] <toad_> http://127.0.0.1:7888/servlet/nodestatus/diagnostics/outgoingRequestInterval/minute
[21:42] <toad_> 599442.3626992428
[21:42] <toad_> 599497.2150970852
[21:42] <toad_> 599258.7471951856
[21:42] <toad_> etc etc
[21:43] <toad_> grrr
[21:43] <toad_> it shouldn't be sending 10 minute intervals any more
[21:43] <toad_> oh well, 1001 little bugs in the code... squash one bug, compile it again, 25,002 bugs in the code...
[21:43] <toad_> bbl zzz
[21:44] <KenMan> that diag draws a nasty graph for 60 minutes worth. Now I log it "long-time" (spoken with an asian accent).
[21:44] <KenMan> heehee - 25002 is about on target.
[21:45] <lostlogic> toad_: are stable and unstable still on different routing systems ATM?
[21:45] * KenMan assumes that was sung to the melody of "21 bottles of beer on the wall"
[21:46] <toad_> yup
[21:46] <KenMan> i think they are about the same code for routing now. but i'm not sure. and toad is preparing himself to sleep.
[21:46] <toad_> 1001 bugs in the code, 1002 bugs in the code... squash one bug, compile it again, 25,003 bugs in the code!
[21:46] <toad_> :)
[21:46] <KenMan> GO TO SLEEP
[21:46] <lostlogic> hehe :-D
[21:46] <toad_> bye ppl talk to you tomorrow
[21:46] <KenMan> :)
[21:46] <lostlogic> night, toad
[21:47] * lostlogic decides to do his running unstable freenet on stable network thing again.
[21:47] <KenMan> sorry, i meant GO FOR WALK. then sleep
[21:47] <KenMan> so what's new, B? have you replaced that brunette yet ?
[21:48] <KenMan> HelenMinor or something ?
[21:48] <lostlogic> hehe, no... pathetically enough, I'm calling her right now
[21:48] <lostlogic> Heather...
[21:48] <lostlogic> you have an impressive memory
[21:48] <lostlogic> hmm... nobody home
[21:48] <KenMan> stop. don't do that. you'll only hurt her and you.
[21:48] <lostlogic> hehe
[21:48] <KenMan> just send her a nice card. It is easier on both of you.
[21:49] <lostlogic> we've actually had 3 or 4 consecutive non-hurtful conversations recently...
[21:49] <KenMan> well, communication is good for helping it all to wind down.
[21:49] <lostlogic> *shrug* I'd rather communcation was good for us ending up back together
[21:50] <KenMan> It's not gonna happen. I had a friend who lived with a guy for 10 years, and then she just up and left him.
[21:50] <KenMan> Well, actually, she wouldn't leave him. And she wouldn't talk to him about it. She just brought other guys home and slept with them in the other room.
[21:50] <lostlogic> wtf.
[21:50] * greycat (rfc1413@wooledge.org) Quit ("Lick Bush in '04")
[21:50] <KenMan> Ouch !!! I really was not on her side by then... I told her, "talk with him about it. Otherwise you will both have questions later on..."
[21:51] <KenMan> but she refused to confront the matter properly, and after a few months this guy got the message. I was completely on his side after that.
[21:51] <KenMan> I didn't speak to her for 2 years... then when *I* started dating her, I wished I hadn't.
[21:51] <lostlogic> hehe, yeah, she sounds like one not to keep
[21:52] <KenMan> she was a fairly decent friend for 15 years ... but I wouldn't wish her on my middle-most enemy. Worst perhaps, but not middle-most.
[21:53] <lostlogic> heh
[21:53] <lostlogic> ok, my node should _look_ like a stable node, but is running unstable code, lets see what I can break
[21:54] <thelema> lostlogic: whee, hacked nodes.
[21:54] <lostlogic> but yeah, Heather and I are (obviously) still close friends, and I definitely still have more than just friends feelings for her, and from the way she talks she does too, but she just isn't ready to say so or something...
[21:54] <lostlogic> thelema: whee, indeed :)
[21:55] <thelema> I don't share toad and ian's dislike for heterogenous networks.
[21:55] <lostlogic> thelema: I just got sick of the high computer usage of stable, and the lack of contenton unstable... I don't have a political position about it
[22:08] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) Quit (Remote closed the connection)
[22:21] * thelema is now known as thelema|away
[22:35] <lostlogic> Jul 26, 2004 9:51:59 PM (freenet.node.rt.OptimizingSlidingBucketsKeyspaceEstimator, YThread-24, ERROR): Interpolated crazy value: -5073.157098066234 in guess(aa3de1cfe79c50c18c5653c92497e9accdf3c468140302 -> aa3de1cfe79c50c18c5653c92497e9accdf3c468140302) on freenet.node.rt.OptimizingSlidingBucketsKeyspaceEstimator@12f41a5: time: keyBefore=98292288be47df576da9ff4fb454a4f52c86698400ee62, keyAfter=a8d2bb999ce4ec93b4585af5518ae62c568
[22:35] <lostlogic> 21228ed732e, key=16305917956063206651044441142514345496222363242472932098, beforeValue=2643.5272727272727, afterValue=3183.96
[22:36] <lostlogic> I wonder if that is caused by running unstable code on stable network...
[22:36] <lostlogic> might be
[23:08] <KenMan> Wait, I haven't had any enemies for several years. Time to don my tin face-mask and go out into the night. "Pie Man is here!"
[23:08] <KenMan> dangitall... somehow my clock is just plain screwed up. Even without running ntp (checking... Nope!). Fred suddenly has per-minute stats for an hour into the future .
[23:08] <KenMan> this only happens when the system has mild arrythmia :(
[23:08] <KenMan> I wonder if it could be a java thing, or if a 'C' gettimeofday program would also see it ... hmmm. Not good in any case. Time to run fledora.
[23:08] <KenMan> Besides , this C library wasn't certified to run with this kernel, or that browser, or any of the other programs I semi-randomly threw togther.
[23:08] <KenMan> Brandon, you better not be talkin' to that girl right now !!! It will be more productive to exchange letters. Actual physical ones. And, if you want to, dope yours with pheremones.
[23:08] <KenMan> She isn't gonna move from AL to St Louis . And you're not gonna move the other way. Right ? Is it any more complicated than that ?
[23:08] <KenMan> Perhaps she would go for some plane sex, but you would have to fly to her. Or else purchase her ticket(s). Tell her you got a return one, and buy it if you can't produce marriage. You do want her around forever, right ? not just for sex a little while longer ?
[23:08] <KenMan> because if you just want more of the same, then it is past time for you to move on.
[23:08] <KenMan> Propose to her by hiring a sky-writing pilot in her area. Call her and tell her to read your message in the sky.
[23:08] <KenMan> Although the US may have outlawed skywriting by now...
[23:08] <KenMan> okay, if you really ARE serious about her, go to her and propose. See what she says. Then you will know what to do. You move, she moves, nobody moves... whatever !
[23:08] <KenMan> I would never move for a woman, let alone a job. But that is just me. And I will move if the bank takes my house away. But people move for other people all the time.
[23:08] <KenMan> shoot. Time to restart my node. It doesn't like it when the system spuriously reports the time as being an hour into the future.
[23:08] <KenMan> Ah Hah! My testnet node has fallen into 100% CPU loop somehow. No wonder the inserts are not working any longer... According to top, it is 10% user and 90% system. Odd...
[23:08] <KenMan> when inserting at HTL=0 (and no peers) can fred still desire to forward inserts with 15% randomness ? I doubt that is it. Maybe I should blame 1.5.0_b2 . Or native biggint.
[23:08] * KenMan doesn't know how to diagnose. No controlling tty, can't get fproxy response... what's left ?? attach strace somehow ? gdb maybe ?
[23:29] * Ash-Fox (Hal-9000@aaq112.neoplus.adsl.tpnet.pl) has joined #FreeNET
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.
