log☇︎
463900+ entries in 0.306s
decimation: asciilifeform: he predicts the roomba by a few decades :)
assbot: BtcAlpha.com F.MPIF Tracker estimated NAV per share: 0.00021372 B (Total: 467.36 B). Delta: 0.18 B. Last trade for F.MPIF on MPEX was at 0.000207 BTC [+]
gribble: Bitfinex BTCUSD ticker | Best bid: 224.52, Best ask: 224.55, Bid-ask spread: 0.03000, Last trade: 224.51, 24 hour volume: 17679.8669452, 24 hour low: 222.65, 24 hour high: 229.23, 24 hour vwap: None
asciilifeform: http://www.burnsideinstitute.com/robot_book/robots.html << with english transl.
Bingo_bar: Fuck the god dampened pony track
The20YearIRCloud: Granted, i guess if you wanted a $10k+ threshhold it'd make sense to up it to 50btc due to all the recent slides
The20YearIRCloud: Exchanges with less regulation an requirements (mostly) than even OTC
The20YearIRCloud: Yeah , crowdexchanges here in the US now that the SEC has finally got off their butt and allowed small companies to list on new exchanges.
The20YearIRCloud: Quite a bit more than the title 4 exchanges I've been talking to in the US
The20YearIRCloud: So I miss anything interesting or fun with mpex in the past 2-3 months?
WolfGoethe: some fool locking up my coins with bogus trades. what fool wants to deposit cash to the bank... at midnight???
punkman: !up The20YearIRCloud
punkman: "Justus Ranvier has one of the more interesting opinions on the bitcoin blockchain that doesn’t seem to get much attention. In his view, there is no need for the block size limit to exist at all."
cazalla: who's that at #4 heh
asciilifeform: (williamdunne? was there a scoopbot thing under testing?)
referredbyloper: yes, drunken bitching about bad marriage and things
cazalla: referredbyloper, you think i'm funny, i amuse you?
trinque: and how long does it take to register anywya
referredbyloper: trading most of the day, an find this channel realy amusing
referredbyloper: just don't want to go through reg and things
trinque: yep guy's clearly trying to make it look so
cazalla: ah ok, i assumed it might be someone coming here from asciilifeform's site in the same manner qntra directs people here
trinque: cazalla: pretty sure that's a bot which only spams one link
trinque: it would not blow my mind if there are session hijacking hax to be found within, so on
trinque: WS just bolt another huge attack surface to the thing
trinque: williamdunne: and to elaborate on the reason why, the browser was a nightmarish wad of complexity before WS
trinque: that said I've used websockets; the whole browser's shit, subset of which WS are shit
trinque: williamdunne: I don't even mean that, just that polling might not look much different load-wise from the WS doing its thing
trinque: "keep-alive" is the thing to look into there
trinque: so the ajax vs ws thing may be more similar than it appears
assbot: HTTP pipelining - Wikipedia, the free encyclopedia ... ( http://bit.ly/1eU5Kap )
trinque: browsers can pipeline requests through the same connection
trinque: there are other things to consider
williamdunne: Very true, but anything where the data is somewhat frequent I reckon would be better off in WS
trinque: and can hand them off to some backend nicely
trinque: I think nginx does pretty well handling a bunch of idle websocket connections
trinque: so you would have to actually try both and see based on your own situation which is more expensive
trinque: on the ajax side, you may have tons of unnecessary polling
trinque: on the one hand, anyone with the site open in a tab is going to be a connection your server will have open
trinque: no need to bet.
williamdunne: Overhead vs the AJAX method? I'd bet otherwise
williamdunne: But if there were best happening more than every second, sure that'd be a great feature to see
trinque: so adding a bunch of persistent connections is going to increase overhead
trinque: oh, you wanted bitbet to provide a websocket?
williamdunne: Streaming basic pricing data, and submitting/receiving bets are the two I have used it for
trinque: williamdunne: what do you want to do with it?
williamdunne: Would there be any particular benefit to that in its intended usage? JSON/MsgPack/whatever all work fine with anything I can imagine being necessary on the web ☟︎
trinque: apparently in some newer version of the standard it has a binary streaming thing
trinque: I mean to be specific it's a wrapper around a socket with certain assumptions on how it'll be used; it doesn't just give you a raw unix socket and say "go at it"
trinque: (maybe I'm wrong? don't think so)
trinque: first off the thing's not a socket; iirc it throws you wads o' data piecemeal
trinque: and on the other there's considering the entire stack that got you to that point, and whether that made any sense
trinque: williamdunne: on the one hand there's a superficial "did I make it do something, and did it 'work'" ☟︎
williamdunne: While with that description it does sound a bit tarded, in practise I've found them to work very well
decimation: the shortcomings of turning a fundamentally broadcast technology into a 'virtual channel'
williamdunne: What is particularly tarded about WS?
decimation: williamdunne: websockets are 'tardation on top of tardation
decimation: it would be nice to click on '3' and call up all the moduli which are factored by '3'
decimation: asciilifeform: yeah that sounds like a good idea
asciilifeform wanders off to contemplate this
asciilifeform: likewise, each unknown is stamped with a date of last gcd-ing, as well as sha512 of the mass product at the time this was done.
asciilifeform: the mechanics of a correct phuctor are, roughly: thing keeps track of keys, rsa moduli within, as existing phuctor does; but instead of flagging 'phuctored' moduli, we keep table of known factors (associated with respective unknowns, which, in turn, may be yet-unphuctored moduli -or- fragments previously created by successful phuctorings)
asciilifeform: as it is, the gigantic turds yielded by the current gcd finds will not break apart
asciilifeform just woke up to the fact that he was a fool when originally wrote phuctor. a phuctored modulus must fragment into factors, each of which is in turn subject to all future attempts at phuctoring
williamdunne: mircea_popescu: Oh wait I take it back, thats me derping. Message comes up blank
williamdunne: mircea_popescu: Thanks MP
williamdunne: cazalla: foreplay is half the fun
cazalla: williamdunne cut the foreplay and just ask
williamdunne: Isn't that the point of WebSockets, getting rid of derpy polling?
asciilifeform: all else follows from this.
asciilifeform: williamdunne: fundamentally - the idiot polling
williamdunne: Anyhow, what in particular is wrong with it? While there is a host of derpy technologies I'm quite fond of some things - like WebSockets
asciilifeform: as in, we take everything post-1990 or so and shoot it. then start over.
williamdunne: Re: polling, what would you suggest as the alternative? Websockets?
williamdunne: With Scoop I worked out a structure that fixes some of the issues
asciilifeform: (and this is entirely aside from the idiocy of how servers get polled continuously, etc)
asciilifeform: was just pointing out the braindamage of the whole concept of rss
mircea_popescu: <asciilifeform> thing is, it is actually possible for the same phuctor url to be new twice. << it'll be fine im sure.
assbot: luke-jr comments on Call me inconsiderate, but THIS is why I don't take Luke-Jr seriously: He lost the majority of his coins on Mt. Gox when there were signs FOR MONTHS to NOT keep coins there. He lacks the ability to properly gauge situations and look into the future. ... ( http://bit.ly/1Jxr94n )
williamdunne: Fair enough, think he'll do fine. Fits in quite well with his existing logic
asciilifeform: wouldn't want scoopbot to choke on it
asciilifeform: it doesn't matter. just pointing out that the thing follows literally none of the prescribed assumptions
williamdunne: Well then why does this matter? Scoop shares it once and you see both
asciilifeform: it is entirely possible for same key to be 'new' twice in succession.
williamdunne: This is not a problem, so long as something changes between first seen and the update
williamdunne: In this case I may have to make a custom set of rules for Phuctor
asciilifeform: rss provided no sane way to handle this.
asciilifeform: thing is, it is actually possible for the same phuctor url to be new twice.
williamdunne: Whether or not the URL is in the map
williamdunne: Uses a persisted Map of URLs to objects. Messages when he sees a new one
asciilifeform: i recall that mircea_popescu asked for a real-time phuctor ticker.
asciilifeform: williamdunne: the dates are entirely unusable. see thread
kakobrekla: as far as i can tell dates are incorrect but the order is correct in your feed ? thats workable. perhaps scoop can cover it. i can too.
jurov: asciilifeform: you mean bot reading rss feed? we have that.
asciilifeform: kakobrekla: can the rss thing be made to work as expected in the form of a #b-a bot ? or too b0rk3d
asciilifeform: mircea_popescu: earlier link to freshly phucked key
mircea_popescu: needless to say...
mircea_popescu: for the record, my ftp problem was that someone had the bright idea of allowing port 21, but not port 20.
cazalla: jurov problem with what i'm doing or the client? not sure what else i can do here now
mircea_popescu: <cazalla> crystalspace3d.org is slow as fuck to get cs-win32libs-2.1_003.exe from btw << link ? jurov can then mirror it along with everything else.