380100+ entries in 0.224s
ascii_field: or rather,
the concept+the working mechanism (ver 99)
assbot: Logged on 05-08-2015 13:54:55; mircea_popescu:
the stator saving of
the continuity feeling was
too narrow for my liking.
ben_vulpes: ascii_field: any "v", your "v" or "v" in
the abstract?
ascii_field: the author of an ffi pkg has no control over what c
turd is on your box
ben_vulpes: ascii_field: don't mind me, i'm just
trying
to learn
the violin.
ascii_field: ben_vulpes: when
they ~are~ common lisp packages
ben_vulpes fails
to find
the logline wherein asciilifeform claims
that common lisp packages "tend
to work unlike in other languages"
☟︎ ascii_field: i have a puzzle for ben_vulpes, who is awake, and mircea_popescu, who might soon awaken, and
the rest:
ascii_field: ben_vulpes:
the
thing is shaky beyond belief
ben_vulpes: phf: where's a sane place from which
to procure gpg-error? i've found some
text files scattered about
the web but nothing in
the gpgme lang/cl dir
PeterL: I'm off
to get
the kids from school, mostly I lurk by reading
the logs
assbot: Logged on 27-10-2015 15:31:19; asciilifeform:
this keeps in perfect fitting with
the 'computing industry'-as-a-protection racket business model of
the last 20+ yrs, where it 'solves' exclusively
those problems which it, itself, creates.
ascii_field: PeterL: one of
the
things
to understand is
that your computer is
the way it is because
the 'computing industry' consists of 1,001 people mutually sucking one another's cocks and profiting off
the braindamage
to every fella actually
trying
to get useful work done.
ascii_field: microshit's very existence has deformed
thought processes of a great many people.
☟︎ ascii_field: 'when mass markets develop, pluralism suffers
the most --
there is no longer a concept of healthy participants: people become concerned with
the individual "winner", and instead of people being good at whatever
they are doing and proud of
that,
they will want
to flock around
the winner
to share some of
the glory.'
☟︎ PeterL: trying
to fit
trinque's nuke-artillery metaphore onto computers, clearly failing
PeterL: so "real computer"
takes over when "winblows users" all die off?
trinque: quantity is not
the only consideration, nor is everyone on
the same side
trinque: if I have 10 nukes and you have 100 artillery in a fortress, and I nuke
the fortress, does a nuke silo spring up in its midst?
trinque: well. yes, but not
that way
trinque: sounds more like "will
this superior military weapon ever supercede
this other weaker weapon in
terms of market penetration?"
ascii_field: i do not know how
to make a cpu for 100 people.
trinque: PeterL:
this assumes
that
the same group of derps will continue
to exist forever
ascii_field: it'd have
to be a computer for 100 people.
ascii_field: PeterL: no likelier
than bitcoin displacing usd
ascii_field: so now you have
to constantly chase
the largely nonexistent docs for new southbridge, etc
PeterL: If you hypothetically could make your own silicon, and "real computer", do you
think it could ever displace
the entrenched systems everybody is currently using?
ascii_field: the basic fact is
that no
two x86 pc boxes can be relied upon
to have
the same iron
ascii_field: even if we ignore
the profound braindamage of
the x86 arch - instruction set, memory structure, clumsily retrofitted multiprocessor, etc. -
ascii_field: (the odds are very good
that your southbridge is running just one such as we speak.)
ascii_field: and all questions of os aside, you will need
to bake own silicon if you don't like running usg blobs.
PeterL: so
the only way
to get a better operating system is
to start with purpose built hardware?
ascii_field: nor any built-in provision for battery-backed
transactional ram
ascii_field: i am deliberately not even mentioning gpu, because it is possible
to have a useful computer with no vga board (e.g., you can speak x11 over
tcp)
ascii_field: AND
this is a device for which we HAVE
the specs !
ascii_field: i picked
this example out of a
tall hat with many others,
to demonstrate
the impedance mismatch between
the lisp way of doing
things and pc iron
ascii_field: PeterL: see 'cdr coding' for why
this does not have
to be
the case
ascii_field: suddenly you need a massive 'bureaucracy'
to convert between
the contiguous blobs your devices eat and shit, and your notion.
PeterL: wouldn't
that make strings
take up
twice as much memory?
ascii_field: where each character is stored in half of a word, with
the other half being a pointer
to
the next char ?
ascii_field: but what if - per last night's
thread - you decided
that you DON'T WANT
TO REPRESENT STRINGS AS CONTIGUOUS BYTES ?
ascii_field: when a packet comes in over
the wire, it gets
thrown into a buffer pointed
to by a particular register in
the card's cpu, and
the latter reads
the addr of
the next buffer from a pre-baked list
assbot: Logged on 05-11-2015 01:07:56; asciilifeform: at any rate, i (and $maxint others) will port $whateverthefuckyouwant
to, e.g., cray II. for a fee.
ascii_field: your NIC, regardless of who built it, operates with ring buffers, roughly like
this
trinque: PeterL: are you arguing from
the wisdom of
the mass market customer?
PeterL: if unix is so horrible, why is
there no alternative
taking over
the world?
phf: i've been
trying very hard for
the past 15 minutes
to remember its name or anything about it, and even
though i spent 8 months with it, i'm completely failing
to
phf: of magnitude improvements and
that
the whole
thing is not even funny. he suggested
that i do an implementation as an indepedent study for another semester and
that's what i did
phf: thestringpuller: it wasn't actually forced. started as a final project in william arbaugh's (the guy who did
the smooth handover wireless implementation at umd) class, where we were writing a malloc for
the
teaching os as a final project, only
three mallocs came anywhere near
the spec, since mine was one of
them i had
the opportunity
to say
that
the whole
thing was a sham "shit code written by retards", etc. and
that you can have orders
lobbes: kakobreklaa: fuckin chanserv << I know what you mean. I had
to ghost my nick off earlier
ascii_field: and only available presently for
the heathen one ?
ascii_field: jurov: do i understand,
this needs a heavily patched client ?
phf: ascii_field:
that's
the problem with a lot of lisp conversations here, your interlocutors have some vague idea about how
things are done right now and in unix world. even unixisms are used not in
their "pure form", but in
their final gnu/linux state.
there's complete lack of semantic match between
that and historical lisp. ultimately debates are reduced
to "macros are bad because cpp is bad"
☟︎ thestringpuller: phf: ugh
they made you do
that shit
too? we had
to write malloc for
the game boy advance, or a game boy advance game for a final project for course in C
ascii_field: but realized
that my interlocutors did not know
the fundamentals involved (because scarcely anybody does, and not everyone is an archaeologist)
ascii_field: phf: i
tried
to explain some of
this in last night's
thread
phf: ascii_field: my undergrad "thesis" was putting a malloc on a one of
those
teaching os's, it was interesting in
that "wrestling with bare metal" sort of way, but
then i learned how lispm does it, and realized
that i was lied
to
PeterL: mircea_popescu:
there is no reason under
the sun
to process, manage, consider or in any way interact with subdollar
transactions. << you are going
to link min bet amounts
to
the USD/BTC price now? are you going
to have
the website show
the "estimated $ amount" like
the VC circus sites do?
☟︎ ascii_field: if you want
to use it as a graphical
terminal
ascii_field: phf:
THE SIMPLEST malloc() is 10x more complicated
than, e.g,
the gc engine in 'tinyscheme'
ascii_field: phf: i was
trying
to see if anyone comes
to
this conclusion independently of my saying it.
mircea_popescu: punkman kinda what i was proposing, a progressive scale. imo
the way i stated it is better pr, but whatevs.
phf: v6 is a beauty, but for
the amount of stuff
that's required
to be written why not lisp os?
mircea_popescu: there is no reason under
the sun
to process, manage, consider or in any way interact with subdollar
transactions.
kakobreklaa: so you wanna make poorfags even poorer
to a point where
they no longer send a satoshi, is
this
the idea?
mircea_popescu: seems it is saner
to be paid in money
than in
time,
to me.
mircea_popescu: i suppose if you prefer i can instead make it a
time rule, and be like "bets under 0.0001 are paid in a year. bets under 0.01 are paid in a week"
mircea_popescu: which will come back
to me, and i have
to sit on it, and create more dust
to payout.
mircea_popescu: think
that i have
to keep SOME dust over here while idiots get aggitated because "o noes it's been 25 hours" so im now creating more dust
ascii_field: mircea_popescu: i don't hate'em, but i did search in vain for signs
that somebody was ~thinking~ rather
than merely piling intel xeons
ten metres high
mircea_popescu: it is not worth my
time
to handle 10k satoshi
txn for less
than 10k satoshi.
this is a fact.
kakobreklaa: if you leave dust sitting long enough and at some point flush all
the dust in a single
tx
to a nondusty result ?
mircea_popescu: ascii_field so you hate
them,
that's fine, but you can't impugn
the whole subdiscipline just because idiots somewhere
mircea_popescu: currently
this CAN work as an extraction engine - it costs me more
to make
the payouts
than i
take in.
mircea_popescu: it's also a problem of cost. dust
txn impose a cost on us.