log☇︎
547800+ entries in 0.381s
thestringpuller: asciilifeform: the private spy networks are probably much more frightening compared to say the CIA
mircea_popescu: or whatever else, all paths lead to the same spot.
ben_vulpes: <coderwill> there were promises of equity but nothing official that i saw (other than verbal) << how does this fly with you people - verbal contracts for a thing with nothing written. when you sign the paper that says "i will work for thus and such" do you not stink the place up for lack of options?
mircea_popescu: whether some 'enemies' were inept enough to do all this and just land it in her lap,
mircea_popescu: now, wehther the woman ordered them to do a job knowing they'll bothc it, or
mircea_popescu: that's the facts of the matter : spy agency is inept (practically) and missaligned (ideally). it has to get gutted.
mircea_popescu: nah, too much.
mircea_popescu: that's the thing, cristina wants to behead them.
mircea_popescu: mno, the current one is us-friendly.
mircea_popescu: whether the woman was involved or wasn't, the ss should get beheaded and well shook up after this.
mircea_popescu: and VERY inept one at that.
mircea_popescu: hard to know what exactly happened there. however, there's little doubt this was a ss job.
asciilifeform: 1) kill inconvenient whoever 2) prattle on about 'dastardly murderers who must be brought to justice' 3) ??? 4) profit
asciilifeform: not necessary for the analogy.
mircea_popescu: asciilifeform guy wasn't some sort of bright young thing
asciilifeform: can use time of last block on wire.
assbot: Argentine president seeks to dissolve spy agency after murky death of state prosecutor - World - Israel News | Haaretz ... ( http://bit.ly/1zR5HEE )
mircea_popescu: <asciilifeform> can't use time of last block then << that's where i caught
asciilifeform: and node operates using system time
asciilifeform: ^ above code will have to be changed to actually set the machine clock ? ☟︎
mod6: i kinda like the notion.
mod6: ya, kalvin has been discussed too
BingoBoingo: <mod6> <+asciilifeform> call it whatever name, but assigning a version equal to ancient phoundation release is foolish. << foolish because of nodes that wont talk with < 0.8, or foolish another reason? i've always detested version numbering because of "marketing" << Propper solution is counting down from 9000
mircea_popescu: and it DID have the correct time once. comes preloaded with the timestamp of say when we started this discussion. or w/e.
mircea_popescu: ben_vulpes never any disagreeement on that score :)
mircea_popescu: asciilifeform> it is then tempting to omit ntp entirely... << i have no idea why it'd have been included.
asciilifeform: (ideally, what runs on toy node will be quite the same thing as used on normal machines, but built for the embedded cpu)
asciilifeform: simply to lazily avoid making yet another change to bitcoind.
decimation: why does system need correct time between bootup and say two live observations of blocks?
asciilifeform: can't use time of last block then
asciilifeform: in fact, this is guaranteed to happen during transport
asciilifeform: and then reconnected
asciilifeform: it is conceivable that the node is cut off from the net for a spell
asciilifeform: at any rate, this does not necessarily work
decimation: asciilifeform: what stuff needs clock? filesystem timestamps? logging?
ben_vulpes: <mircea_popescu> it's the only way to do time. << i've been saying there's no such thing as time for a while now, only the length of the longest chain...
asciilifeform: timestamp on fs only works if we had the correct time once
decimation: yeah, you would probably want to average a few blocks & get system time
asciilifeform: it is then tempting to omit ntp entirely...
asciilifeform: but can lift from file system timestamp
asciilifeform: may not even need to patch bitcoind
mircea_popescu: all that's needed for this application.
mircea_popescu: it's the only way to do time.
asciilifeform: one more patch then.
decimation: heh that's a good ida
mircea_popescu: box thinks the time is "the previous block's time"
asciilifeform: without ntp, box has literally no notion of what the time is.
asciilifeform: decimation: tricky, this. ntp is quite likely to be the weakest link here. central point of failure.
decimation: re: ntp server < to do a good job, you really need a dedicated server to host
assbot: The froth of our days pe Trilema - Un blog de Mircea Popescu. ... ( http://bit.ly/1y11M2n )
mircea_popescu: phillipsjk basically sanity split from idiocy in bitcoin on the topic of porn, back in 2012.
phillipsjk: I take it the female population in this channel is low then.
asciilifeform: mod6: if we're prepared for the eventuality that only our nodes answer our nodes - then yes.
phillipsjk was asking about the pics
mod6: so it seems to me, the most logical thing to call this is v0.5.3.1
asciilifeform: mod6: version number is actually used in the proggy
mod6: <+asciilifeform> call it whatever name, but assigning a version equal to ancient phoundation release is foolish. << foolish because of nodes that wont talk with < 0.8, or foolish another reason? i've always detested version numbering because of "marketing"
mircea_popescu: !up phillipsjk nah, just tradition.
mircea_popescu: nothing terribly.
ben_vulpes: anything interesting in the logs today?
ben_vulpes: <asciilifeform> mircea_popescu, ben_vulpes, et al: missing ingredient - armv5 build of bitcoind 0.5.3 with necessary patches - that is, 100% therealbitcoin-controlled seeds. << pretty neat
ben_vulpes: mkay, well i published a patch that bumps to 0.5.3.1, don't know if you saw that
asciilifeform: call it whatever name, but assigning a version equal to ancient phoundation release is foolish.
mircea_popescu: bitcoin is bitcoin. no need to give bitcoin a special name just because some derps are derping.
asciilifeform: mircea_popescu: splitting the net
asciilifeform: why make it easy for the bastards.
asciilifeform: ben_vulpes et al: this has more than merely ceremonial effect, as, iirc, there are phoundation nodes that won't talk to < 0.8
asciilifeform: ben_vulpes et al: consider mutilating the version string. perhaps 'Serenissima 1.5.3.x' ?
asciilifeform: or if cannot reach the net.
asciilifeform: i'm having it turn red once the drive fills.
asciilifeform: incidentally, there is an output device: bicolour led.
asciilifeform: iirc mircea_popescu wants to hand these out at parties.
phillipsjk: Should be possible then.
asciilifeform: but these, afaik, do not exist either
phillipsjk: Hmm Freicoin is apparently using only like 36MB of RAM at the moment.
BingoBoingo: <asciilifeform> phillipsjk: neither fanless, nor pocket-sized, nor still in production. << The trifecta of want
phillipsjk: I recommend 2GB or RAM for running a node. sounds like you are trying to get that number down.
phillipsjk: yeah, had to move it yesterday. revised the weight estimate up from 50lbs to 60lbs to 70lbs.
asciilifeform: phillipsjk: original thread concerned experiments with machines that will run bitcoind that fit these characteristics.
asciilifeform: and doesn't have dedicated cores running weird, evil closed-sourced blob at all times. or a defective i/o bridge.
asciilifeform: 'pogo' - costs less than half of rpi, incidentally.
asciilifeform: in two dozen different ways.
trinque: in my experience the rpi skews clock as some function of cpu load
trinque: asciilifeform: ah I glossed over "why more than once"
phillipsjk: Bitcoin tolerates time-stamps off by an hour anyway since a certain popular OS did not support hardware UTC until recently.
asciilifeform: phillipsjk: we'd like some that are not under nato control.
phillipsjk: I suggest pool.ntp.org for NTP servers that work reasonably often.
BingoBoingo: <asciilifeform> (why ever run it more than once? leap seconds?) << One of those coming up soon
asciilifeform: (until somebody plugs the leaks)
asciilifeform: so i'm not averse to simply having it reboot weekly or so
phillipsjk: running NTP more than once lets you compute clock drift.
trinque: ntpd can apparently be told to always update, never give up when skew is huge, but *shrug* never worked for me
trinque: I've found ntpd to be a pain on such devices; I just cron a call to rdate instead
asciilifeform: (why ever run it more than once? leap seconds?)
asciilifeform: and then ntp client can be omitted from the rootfs itself.
asciilifeform: it turns out that ntp can happen straight from u-boot (bootloader)
mod6: that shouldn't be a huge problem. bbiab -- we can talk about that part a bit more.
asciilifeform: preferably including one or two controlled by wot folks
asciilifeform: also needed is a small list of ntp servers that can be trusted to work reasonably often
asciilifeform: http://cfile2.uf.tistory.com/image/1809383A517761BB1B27C9 << where to put