00:04:07 | bramm: | bramm is now known as bramc |
00:11:00 | Eliel_: | I'd guess you'd at least want enough of the redundancy geographically close to be able download the entire file from reasonably local nodes. |
00:11:59 | op_null: | close nodes means failure by tsunami is very likely. you really want them to be as far away as possible. |
00:14:28 | Eliel_: | if you go for over 200% redundancy, you can pick local nodes for ~100% of the storage and then distribute the rest maximizing the geographical distribution. |
00:15:38 | Eliel_: | that will still withstand a tsunami that wipes out the local nodes. |
00:17:14 | Eliel_: | ... in fact, in most cases the network isn't needed for that since a copy of the file is stored locally on the user's computer :P |
04:18:14 | bramm: | bramm is now known as bramc |
06:21:22 | gmaxwell: | I'm a bad forum dweller and I bumped this old thread for an I told you so, https://bitcointalk.org/index.php?topic=144471.msg9764161#msg9764161 |
06:44:59 | gwillen: | gmaxwell: well, this is a channel for wizardry, and ... thread necromancy is clearly some kind of magic |
06:47:02 | bramc: | gmaxwell, This might be related to Jed not being terribly keen on sending me his technical docs |
07:15:33 | Taek: | In defense of storj, it's not so bad to repair your files if you require the users to log on every few months or something. Depending on the average reliability of a node, using 200% or 300% redundancy could last a year or more. They just need to pop in once to restore the original 300% redundancy. |
07:33:36 | fanquake: | andytoshi There’s way to much interesting stuff on your site. |
08:56:50 | op_null: | Taek: I'm not sure that's what users expect from such a thing. I can image the node churn rate with such a thing being quite high. |
09:00:44 | op_null: | guess to me it feels like a really hacked together system with low real world utility. as far as storage and retention goes a centralised service is better on almost every level. |
09:03:46 | op_null: | if your stored files are at all critical you're going to want to keep a good copy of them for Justin Case. might as well just throw an encrypted copy on a few cloud storage hosts and call it a day, the end result is the same if not better. |
09:05:19 | cameron.freenode.net: | topic is: This channel is not about short-term Bitcoin development | http://bitcoin.ninja/ | This channel is logged. | For logs and more information, visit http://bitcoin.ninja |
09:05:19 | cameron.freenode.net: | Users on #bitcoin-wizards: andy-logbot fanquake op_null Guyver2 go1111111 btcdrak coutts ahmed_ hguux_ koeppelmann paveljanik gavinandresen Shiftos waxwing TheSeven Dr-G wallet42 Graftec Hunger- rusty gues_ v3Rve zaan prepost nuke1989 NikolaiToryzin RoboTedd_ jb55 DougieBot5000 soundx nubbins` hashtag_ atgreen ryanxcharles Greed PRab Aquent warptangent devrandom Fistful_of_Coins jgarzik phantomcircuit wumpus heath koshii HM2 paperbot coiner Baz__ kefkius zibbo |
09:05:19 | cameron.freenode.net: | Users on #bitcoin-wizards: SubCreative samson_ maaku epscy OneFixt Transisto Starduster prodatalab lnovy EasyAt c0rw|zZz jaekwon_ Cory digitalmagus LarsLarsen Guest38445 coryfields iddo otoburb_ kinlo azariah4_ Guest6066 grandmaster HaltingState null_radix Luke-Jr fenn shesek bosma tdlfbx iambernie mr_burdell adlai tacotime mkarrer nsh_ yoleaux btc__ Graet Logicwax andytoshi pi07r Meeh starsoccer PaulCape_ DoctorBTC nsh mappum tromp K1773R nanotube dansmith_btc kanzure |
09:05:19 | cameron.freenode.net: | Users on #bitcoin-wizards: spinza Flyer33 arowser1 Anduck Nightwolf hollandais berndj lclc_bnc s1w Eliel_ bobke LaptopZZ sneak harrigan JonTitor sl01 Grishnakh sipa Alanius jbenet MRL-Relay fluffypony BlueMatt a5m0 gazab AdrianG gwillen livegnik BananaLotus @ChanServ burcin wizkid057 Dyaheon bbrittain tromp_ Krellan poggy comboy mmozeiko Taek optimator_ [\\\] Apocalyptic throughnothing petertodd crescendo cfields-away gmaxwell so [d__d] espes BigBitz jaromil helo Keefe |
09:05:19 | cameron.freenode.net: | Users on #bitcoin-wizards: Iriez huseby lechuga_ harrow roasbeef ryan-c TD-Linux catcow danneu amiller smooth gribble asoltys pigeons warren BrainOverfl0w midnightmagic phedny |
10:20:48 | BlueMatt: | anyone have some good examples off-hand that seem like great changes to bitcoin break things in very subtle ways? |
10:21:02 | BlueMatt: | preferably things related to consensus mechanism |
10:31:59 | BlueMatt: | or in overt ways...eg the "discourage newly-seen forks" |
10:34:35 | op_null: | the instant difficulty retargeting lots of altcoins has is a nice one. people got sick of difficulty changes getting pushed into weeks, so they made all sorts of systems where the retargeting happens every block or evey few blocks. |
10:36:13 | op_null: | while not subtle as a cryptographic thing, it caused all sorts of issues and ruins the expectation that somebody can't isolate you and feed you low level blocks. |
10:42:07 | BlueMatt: | * BlueMatt goes to read the ethereum blog to find some good ones...quite a lot here |
10:47:05 | gmaxwell: | BlueMatt: I'd caution that you should abstract enough to not trigger anyone's defences. |
10:47:17 | gmaxwell: | BlueMatt: maybe use the response on CLTV on bitcoin-development recently. |
10:47:20 | BlueMatt: | gmaxwell: I'm trying to pick things that were never implemented :) |
10:47:46 | op_null: | oh. |
10:48:21 | gmaxwell: | Petertodd and I so lovingly crafted that design to _not_ be a fork magnet, and a presumably smart earnest person responded on the list with a bunch of "oh why not do X, Y, Z" ... which were mostly things that would turn it into a triggerable hardfork. |
10:48:40 | BlueMatt: | gmaxwell: yea, I have a separate section of the talk on forking stuff |
10:48:52 | BlueMatt: | but thats a good example for that part, thanks |
10:48:53 | gmaxwell: | BlueMatt: oh I see. |
10:49:31 | BlueMatt: | * BlueMatt had a reasonable conversation with someone a few weeks ago (the guy who was on the ml saying "we need to hardfork regularly") challenging that declaration |
10:49:45 | BlueMatt: | only to find out after 20 minutes that he wasnt aware of the definition of soft/hard fork |
10:49:49 | BlueMatt: | these things need explained :) |
10:49:58 | gmaxwell: | ouch, yea, seen that happen before. |
10:52:03 | gmaxwell: | BlueMatt: an older often one is "why not just refuse to reorg at 6 blocks deep". You get lovely effects like attackers who just intentionally mine / continue two forks at depth 6 and then concurrently announce them to the whole network and the network is forever partitioned. |
10:52:33 | BlueMatt: | yea, I was gonna do the recent one where you discourage newly-seen forks |
10:52:40 | BlueMatt: | pretty much the same thing, in result at least |
10:52:47 | gmaxwell: | Now you might say, "well thats really hard to compute surely you don't need to worry about it" ... but apparently reorgs of that depth were enough of a concern that you wanted to "fix" them. |
10:53:00 | BlueMatt: | heh |
10:53:39 | gmaxwell: | Plus new nodes entering the network would automatically be inconsistent with existing ones if there were a long fork. |
10:53:53 | BlueMatt: | yea |
10:54:42 | gmaxwell: | I guess the notion of tradeoffs is important to express. Some things are possible, but seldom for free. Not sure what would make the best example of that. |
20:21:40 | wolfe.freenode.net: | topic is: This channel is not about short-term Bitcoin development | http://bitcoin.ninja/ | This channel is logged. | For logs and more information, visit http://bitcoin.ninja |
20:21:40 | wolfe.freenode.net: | Users on #bitcoin-wizards: andy-logbot koeppelmann askmike RoboTeddy nullbyte Shiftos Guyver2 lechuga_ gues__ zwischenzug2 Starduster waxwing Emcy bobke nuke1989 coiner btcdrak coutts CryptOprah artifexd instagibbs Muis hguux_ michagogo ahmed_ kumavis Flyer9933 Greed belcher [\\\] nickler MoALTz austinhill1 op_null go1111111 paveljanik gavinandresen TheSeven Dr-G Graftec Hunger- v3Rve prepost NikolaiToryzin nubbins` hashtag_ atgreen ryanxcharles PRab Aquent warptangent |
20:21:40 | wolfe.freenode.net: | Users on #bitcoin-wizards: devrandom Fistful_of_Coins jgarzik phantomcircuit wumpus heath koshii HM2 paperbot kefkius zibbo SubCreative samson_ maaku epscy OneFixt Transisto prodatalab lnovy EasyAt c0rw|zZz jaekwon_ Cory digitalmagus LarsLarsen Guest38445 coryfields iddo otoburb_ kinlo azariah4_ Guest6066 grandmaster HaltingState null_radix Luke-Jr fenn shesek bosma tdlfbx iambernie mr_burdell adlai tacotime mkarrer nsh_ yoleaux btc__ Graet Logicwax pi07r Meeh starsoccer |
20:21:40 | wolfe.freenode.net: | Users on #bitcoin-wizards: PaulCape_ DoctorBTC nsh mappum tromp K1773R nanotube dansmith_btc kanzure spinza arowser1 Anduck Nightwolf hollandais berndj lclc_bnc s1w Eliel_ LaptopZZ sneak harrigan JonTitor sl01 Grishnakh sipa Alanius jbenet MRL-Relay fluffypony BlueMatt a5m0 gazab AdrianG gwillen livegnik BananaLotus @ChanServ burcin wizkid057 Dyaheon bbrittain tromp_ Krellan poggy comboy mmozeiko Taek optimator_ Apocalyptic throughnothing petertodd crescendo cfields-away |
20:21:40 | wolfe.freenode.net: | Users on #bitcoin-wizards: gmaxwell so [d__d] espes BigBitz jaromil helo Keefe Iriez huseby phedny midnightmagic BrainOverfl0w warren harrow roasbeef ryan-c TD-Linux catcow danneu amiller smooth gribble asoltys pigeons |
21:16:24 | c0rw|zZz: | c0rw|zZz is now known as c0rw1n |