354000+ entries in 0.23s
gabriel_laddel: non-CL lispers by-and-large, fail
to understand why
they're using lisp at all.
mircea_popescu: i can readily understand why
the ~performance~ issue is a
true quagmire.
gabriel_laddel: mircea_popescu: After SMBX died CL had
to be re-implemented.
mircea_popescu: (note
that i'm not discussing performance. i'm not going hey, why doesn't lisp-gl push out 2x
the fps of directx in 2001. just, it should be abel
to do it, at all.)
mircea_popescu: but what's
the problem, anyway ? seems like
this is an exact match for
the argued strengths of lisp.
phf: mircea_popescu: oh
they do,
that's a common argument
mircea_popescu: and nobody seems
to rub
the lispers' face in
this, either. what,
there's some sort of chivalrous convention i don't know about going on ?
☟︎ mircea_popescu: meanwhile...
there is an implementation of particularly gnarly in c, but not in lisp.
☟︎ mircea_popescu: yeah well...
this logic is broken. so lisp is good because, and i quote, "Because Lisp, as a
tool, is
to
the mind as
the lever is
to
the arm. It amplifies your power and enables you
to embark on projects beyond
the scope of lesser languages like C. Writing in C is like building a mosaic out of lentils using a
tweezer and glue. Lisp is like wielding an air gun with power and precision. It opens out whole kingdoms
phf: mircea_popescu:
the spirit is willing, but
the flesh is weak?
phf: everything
talks
to xlib, win32, carbon/cocoa and on
top of
that
to opengl, directx.
there are parts of
that stack implemented in lisp (either as clients or as hosts).
there's no opengl in lisp,
there's opengl client in lisp.
there's both xlib client and server in lisp, etc.
there's not enough lisp code
to run
the whole graphics stack from
the lisp machine
to
the video card end
to end
mircea_popescu: what, all
through 1995 - 2005 everyone "in lisp" sat around on
their ass passing back and forth stories of how cool moon was in
the 70s and how great gabriel's hairdo looks ?
☟︎ phf: mircea_popescu: oh, no,
that was
the answer
to "there is no native lisp gfx engine?"
mircea_popescu: the lisp power rangers failed
to lead
the wave in implementing opengl correctly
twenty years ago
☟︎ mircea_popescu: incidentally if
things are
this fucked up,
the correct avenue would probably be
to define a full gfx stack
☟︎ phf: mircea_popescu: i
think
that
there's a bunch of people
trying
to chuck moore
their way out of "modern computing" here, as such
there's not native anything
phf: gabriel_laddel: i'm saying
that it's only worthwhile
to bring in zen if you want
to ultimately do x11 support for foreign code. right now your problem is clim+xlib, you want your problem
to be clim+xlib+zen+cl-opengl+glx. if
the ultimate goal is drop x,
then might as well
try and retarget clim
to framebuffer.
the
two (adding zen and framebuffer clim) or at 100% identical in
the amount of effort
mircea_popescu: why not back in 1999 when opengl was not even
taking over yet ?
gabriel_laddel: mircea_popescu: have you
tried reading
the OpenGL spec?
gabriel_laddel: It seems like what you're getting at is
that
the correct order of affairs is
to
tackle zen first, and
then CLIM?
gabriel_laddel: phf: yes, debugging CLIM + zen at
the same
time is going
to suck, but debugging CLIM + xlib sucks, and I have not yet compared
the
two
to find out which is worse.
gabriel_laddel: phf: "because can still attach c program outputs" + "zen doesn't support most of
the modern x11 extensions,
that you might want for running something like firefox". We have CLIM irc, a listener (repl), an editor, dired + other. I don't particularly want
to run c programs. You'd just end up reinventing CLIM anyways after you notice
that "gee, it would be nice
to have draggable crossbars in my GUI and geometry".
mats: and
then he disappeared forever (I didn't render any payment,
though)
mats: we were almost finalized on a small project where he'd design a widget and send me 3d prints of
the prototype
phf: i know you do clim, but running clim
through zen is essentially simultaneously
trying
to debug
two 70% speed/coverage projects. even simple handwritten xlib proggies don't work out of
the box.
phf: gabriel_laddel: gnuplot, xfoil kind of output, i.e. batch graphic, or anything
that requires input with a mouse
gabriel_laddel: "i've been building my own
tools
to be essentially a combination of cmucl repl,
that spawn xlib windows on demand" << xlib windows
that display what exactly? Large GUIs? Renderings of novel structures?
phf: also zen doesn't support most of
the modern x11 extensions,
that you might want for running something like firefox. in fact last
time i checked it didn't support xrander, which is an old image composition extension
that might as well be standard
phf: main problem with zen right now is
that it does rendering, but it uses opengl (glx specifically) as
the framebuffer. i've not looked at
the implementation bits, but i assume
that it goes beyond simple blit, so you might be able
to retarget it
to other opengl, but not necessarily straight
to framebuffer, without implementing your own blit level rendering primitives
phf: how xserver
talks
to hardware is entirely outside
the scope of protocol, since server accepts
the xproto packets and
then does rendering
phf: gabriel_laddel: i was drunk when answering you last
time. i like
the idea of a self-contained lisp system with ux built on
top of xproto/xlib/zen, because can still attach c program outputs, and stream own output
to unix systems proper. i've been building my own
tools
to be essentially a combination of cmucl repl,
that spawn xlib windows on demand. not sure if you know but x is essentially a binary network protocol (described in
mircea_popescu: "oh, i'm not like
those suckers working at pretending haskell is a
thing ... i'm better
than
them... i'ma do
the same
thing by myself!!!!1" herp.
assbot: Logged on 12-01-2016 23:08:29;
trinque:
took something
that fits
to sexps just fine, made HOON
mircea_popescu braces himself for "number
theory is not really analysis and you're not doing REAL math, only we multiplying matrixes while polishing lenses are doing
the real maths!"
mircea_popescu: see
the discussion with cads on
the
topic
two years ago.
the for ~ANY~
thing is what's important
there.
mircea_popescu: because nobody fucking sane actually
thinks
THAT is how you note down
that
thing.
mircea_popescu: actually... when it came
time for
the cantor proof,
the "work hard" galz
tried
to copy
the whole fucking matrix off
the
table, and
this made a meme at
the
time
assbot: Logged on 12-01-2016 22:59:46; jurov: also, i can't imagine
taking notes on math analysis lectures using "plaintext" only
mircea_popescu: in fact,
the eulora experience is very much in line with a lengthy personal history whereby very smart, well intentioned, patient people
trying
to cater
to my shockingly insane whims and whines end up with exceptionally well crafted items
that stand proudly on
their own.
mircea_popescu: and don't
tell me "oh i can not keep maps in head o noes"
mircea_popescu: and im not saying
the gfx add-on is not helpful or w/e. but also not fucking mandatory.
mircea_popescu: not by any means all my doing. but
the design is sane and yes, it does. she's currently making a map of resources in practically
this.
mircea_popescu: looky : run diana_coman's bot,
turn of
the gfx altogether, it's 100% bash
mircea_popescu: you don't have
to run
the graphical client! for all it cares you can
talk
to it by hand.
ascii_butugychag: jurov: if
this were actually so, and he were consistent about 'in wot', would have
to avoid, e.g., aircraft with computerized autopilot
ascii_butugychag: actually english speakers are so habituated
to
this
that
they have problems with ANY attempt at 'english-parsing machine'
jurov: ascii_butugychag: you see, quaternions are not in his wot and it's problem of
the lusers who use it
mircea_popescu: but be all
this as it may, for as long as gossipd.pop | print yields ((1,2),(2,3)) for me while gossipd.pop | mathgraph yields whatever ascii wants it
to yield, we're happy.
jurov: even with best affort of
the author, ast in
the recipients varies
mircea_popescu: jurov
there's a difference between a proper existing ast, such as we have with words, unless we're idiots ; and a random gcc barf product.
ascii_butugychag: but somehow
this
turned into 'let's do maths as euclid did, with WORDZ!111'
jurov: well if I have
to explore context what a word mean, well,
then it is C++
too
mircea_popescu: so
then we basically are happy with an alphabet of ~100 ish characters and
that's
that.
ascii_butugychag: mircea_popescu: i did say many
times
that a canonical representation of whatever is, rightfully, sexpr
mircea_popescu: jurov
the later is actually part of
the graph. i can not meaningfully search for a motherfucking dot.
mircea_popescu: if you expect
to have a [
that scales
to lines, we got a problem.
there's not going
to be any 244/245/179 bs
jurov: how is adding symbols different from you
taking a word and redefining it?
mircea_popescu: as long as what we do is we
take ((1,2),(2,3)) and either print it as such or else draw it for your benefit as seen
there, if you run a plug in everyone's happy.
mircea_popescu: ~= equivalent
to all
the rest of
the us pulp, call it hubbard and send it home.
mircea_popescu: the complete meaning of star warsa can be conveyed in
two lines.
mircea_popescu: that i can see
the complete meaning in plain
text whenever i opt
to not run whatever program.
mircea_popescu: which is what started
this entire discussion : pdf does not degrade gracefully
mircea_popescu: nothing, as long as you don't expect me
to run
the program
mircea_popescu: do you understand
that your computer is broken by design for as long as you allow
the conceptual possibility of adding symbols ?
mircea_popescu: not even. it's just
that i'm so allergic
to "dynamically linked" bullshit
that i barf at
the very notion of a computing system
that might need "expansions" because people don't want ot learn
to read and write