150500+ entries in 0.093s
a111: Logged on 2017-10-19 04:37 asciilifeform: i'll mass produce switches some
time after doing same for ordinary dc power supplies,
they rank ahead in seekritude
a111: Logged on 2017-10-19 08:05 mircea_popescu: yeah. same principle, but for gas rather
than electricity.
there's a special button by
the main derivation
that resets
the flow
a111: Logged on 2017-10-19 07:10 mircea_popescu: proper (ie, like power chord, either works or doesn't) switch very much on
the republican slate. no extant item satisfy
this requirement.,
a111: Logged on 2017-10-19 07:03 mircea_popescu:
http://btcbase.org/log/2017-10-19#1726558 <<
the expectation is
that it will allow nsa comms, such as exfiltration in response
to magic packets and
the very magic packets in question.
a111: Logged on 2017-10-19 06:30 mod6: ok, just ran 2 quick
tests on different box. results are very similar. different # of samples on new ones, however old|new: 25%|27.42%, 50%|53.57% respectively
ben_vulpes: vegetarians have
terrible internal flora, can confirm
☟︎ mircea_popescu: well for one
thing she had no idea how/where, i never saw
the item installed anywhere else in
that country.
mircea_popescu: yeah. same principle, but for gas rather
than electricity.
there's a special button by
the main derivation
that resets
the flow
☟︎ ben_vulpes: ah i see; i was
thinking about little glass amp-rated
tubes
mircea_popescu: this had
the stove and
the heater close
together and so it made sense, if you got a leak it's
there.
mircea_popescu: ben_vulpes no, see,
the super-safe way is
to have a detector by
the burners
that cuts
the gas line if stimulated.
mircea_popescu: mortified noob had
to endure
the humiliation of me having
to actually go outside reset
the gas line
mircea_popescu: did i recount
that event when nude girly at
the stove farted and
the gas detector immediately blew a fuse ?
ben_vulpes: however new residence has a "whole house fan", and so exhausting all interior gasses was a
trivial exercise
ben_vulpes just had an exciting; apparently closed
the flue a bit
too soon/didn't give
the fireplace enough
time
to self-extinguish and
the co detectors went off
a111: Logged on 2017-10-19 06:30 mod6: ok, just ran 2 quick
tests on different box. results are very similar. different # of samples on new ones, however old|new: 25%|27.42%, 50%|53.57% respectively
ben_vulpes: i
think we've had a few
threads in re creativity in
the eastern empire
a111: Logged on 2017-10-19 05:01 asciilifeform:
this is not unreasonable request, btw, and will filter out bozos
a111: Logged on 2017-10-19 04:54 mod6: gens, sure. do you want
to share power conditioning equip?
a111: Logged on 2017-10-19 04:52 asciilifeform: BingoBoingo: i got one like
that, it wanted 30bux/mo per amp (220v)
mircea_popescu:
http://btcbase.org/log/2017-10-19#1726602 << 19 cents
to
the KWh is not really space station price, considering a) electricity is 5
to 10 cents
to
the KWh pretty much everywhere and b)
this price includes HVAC, which is for obvious reasons just about half of
the cost of power.
☝︎☟︎ mircea_popescu: proper (ie, like power chord, either works or doesn't) switch very much on
the republican slate. no extant item satisfy
this requirement.,
☟︎ a111: Logged on 2017-10-12 21:35 asciilifeform:
trinque: will need, laugh if you must, 2nd inverter
a111: Logged on 2017-10-19 04:36 asciilifeform: i don't manufacture switches, BingoBoingo , and i have no intention of starting, it isn't a device
that handles seekrits, and very literally either works or does not -- like mains cord.
a111: Logged on 2017-10-19 04:25 davout: you're going
to have
the same issues upstream anyway neh?
mod6: originals were
taken with a much older version of ent
than
the new.
mod6: ok, just ran 2 quick
tests on different box. results are very similar. different # of samples on new ones, however old|new: 25%|27.42%, 50%|53.57% respectively
☟︎☟︎☟︎☟︎ mod6: (just
to see if I get a different value, or non-round value for X^2)
mod6: ima
try using an ent on a different box against some of
the runs
that I did previously...
mod6: yeah, not
too sure. i've also never looked at
the `ent` code.
ben_vulpes: statistics being
the
thing
that hoses intuition every
time.
ben_vulpes: leaning on intuition here; feels odd
to me
that
the "randomly would exceed" value is so round on so many of your
test runs, but it's gotta be weakness in my statistics intuition
mod6: There's a bunch of in-depth information in
this book, I highly suggest if examining
these results.
ben_vulpes: mod6: yeah i
think i'll just have
to read
this myself
mod6: if (by interpolation in
the
table) V lies between
the 5% and 10% entries, or
the 90% and 95% entries,
the numbers might be ``almost suspect.''" -- Knuth,
The Art of Computer Programming Vol. 2, Seminumerical Algorithms pp. 46,47
mod6: "If V is less
than
the 1% entry or greater
than
the 99% entry, we reject
the numbers as no sufficiently random. If V lies between
the 1% and 5% entries or between
the 95% and 99% entries,
the numbers are ``suspect'';
mod6: yah, i'll go grab
the book, standby.
ben_vulpes: mod6:
this is from "the art of computer programming" or... ?
mod6: might be in
the logs
mod6: ben_vulpes: seems fine. iirc, and i don't have
the book in hand at
the moment, but knuth outside certain boundaries is suspect.
ben_vulpes: mine
too; what did nickserv
tell you when you
typed exactly /nick raven_zero
ben_vulpes: Guest29647: once you register
the nick, you gotta also set it as your nick like /nick raven_zero
ben_vulpes: mod6: for contrast
this first run just came out with 80.89%
ben_vulpes: but
that one's also going
to be
taken, freenode is a very busy place
mod6: i'd have
to look at other's results, can't recall off
the
top of my head.
mod6: ben_vulpes: ah, i seem
to remember reading about
the X^2 percentiles in knuth. mine in particular came out pretty round ya. or do you mean "your" in
the general sense?
ben_vulpes: Guest29647:
type /msg nickserv help for details on
that
ben_vulpes: mod6: kinda interesting
that your `ent` chi-squared "randomly would exceed
this value XX percent of
the
time" always comes out
to something nice and round like 25.00 or 50.00
Guest29647: maybe check
the instructions in
the mail?
ben_vulpes: lol yeah
that name has good odds of already being in use
mod6: would figure
that would be something you would want in-cage.
mod6: gens, sure. do you want
to share power conditioning equip?
☟︎ mod6: <+BingoBoingo> <ben_vulpes> mhm; estimated as 42*.3kW << Of
those 42U, 1 or 2 go
to switch.
Then you gotta keep a sane heat envelope lest some alf-weird cook everything and piss off
the neighbors. 42U ends up
turning into something like 20 servers apparently. << minus space for UPSs
BingoBoingo: But when exploring datacenter options, biggest obstacle is encountering Jolo
the Pinoy in all nationalities. From "10 mbps
to rack"
to "Try
this managed colo where we buy you boxes and do everything, it's just like getting 7Kva of dedicated servers but with more upsells"
BingoBoingo: And
today fiber runs
to "meet me" room or other nonesuch,
then
to street or sewer ditch or wherever
BingoBoingo: Options for
that are either have a hardware on
the rack
that aggregates ISP links
to
the DC and handles
this shit, OR beat a jolo into sense.
BingoBoingo: Maybe customer wants super anal static routing BBQ for
their box?
BingoBoingo: Own in
the sense I manage
the settings on
the appliance instead of
trusting DC
to not assign
the
task
to
their jolo. And at desired connection speeds switch handles
the routing, unless you wanna go Cisco.
BingoBoingo: Point is not "weed out all annoyances", point is make
the one finite mitigation (own switch) which is less mitigation and more practice for scaling later.
a111: Logged on 2017-10-18 19:08 asciilifeform:
the implication also is
that
there is a particular kind of mistake, exemplified imho by
the zeroasset corp
thing in bbet,
that is easy
to make --
to insist
that
the energy be conserved at all
times