<eligrey>
arrow just ended and GoT's next ep isnt out yet
<eligrey>
vigs: not into the stupid shows
<vigs>
lol
<purr>
lol
<eligrey>
like reality tv
<eligrey>
anything else is fine as long as it's interesting to ELLIOTTCABLE as well
<vigs>
I dunno, I haven't had time to watch anything recently
<eligrey>
what about interesting youtube channels?
<vigs>
idk I only ever tune into sxephil if I remember in the evening or minutephysics/veritasium, vlogbrothers, other stuff…but again, I don't really seek out youtube channels much anymore
Rusky has quit [Read error: Connection reset by peer]
Rusky1 is now known as Rusky
<ELLIOTTCABLE>
vigs: wat
<ELLIOTTCABLE>
vigs: re: 3DS and wireless
<ELLIOTTCABLE>
“the stupid shows”
<ELLIOTTCABLE>
there's *so* much shit you haven't seen.
<ELLIOTTCABLE>
reality TV is, by the way, fucking dumb. agreed 'bout that.
<ELLIOTTCABLE>
Have you watched: Walking Dead? Breaking Bad? Castle? Dexter? The Americans?House of Cards? White Collar? The Sopranos? Downton Abbey? Mad Men? Sherlock?
<ELLIOTTCABLE>
those are all absolute must-watch-every-episode-ever-produced.
<ELLIOTTCABLE>
and if you're looking for something lighter, and enjoy medium-production-value comedy gold,
<ELLIOTTCABLE>
How I Met Your Mother, Archer, Community, Happy Endings, Psych, Enlisted …
<vigs>
ELLIOTTCABLE: hi?
<vigs>
oh
<ELLIOTTCABLE>
hi
<purr>
ELLIOTTCABLE: hi!
<ELLIOTTCABLE>
playing New Vegas for the first time
<vigs>
yeah that's because my apartment place uses the same network as the college so
<vigs>
I have to authenticate with my university ID and password, and that's not supported on the 3DS
<ELLIOTTCABLE>
ah
<vigs>
yup
<vigs>
It'll only be a problem for like…one more month, though
<ELLIOTTCABLE>
lolk
<purr>
lolk
sharkbot has quit [Remote host closed the connection]
sharkbot has joined #elliottcable
colbyl has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rckts has joined #elliottcable
oldskirt_ has joined #elliottcable
oldskirt has quit [Ping timeout: 252 seconds]
rckts is now known as colbyl
<devyn>
okay so good news for Paws.rs
<devyn>
turns out the functionality I needed
<devyn>
actually is there
<devyn>
I just missed it
glowcoil has joined #elliottcable
colbyl has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
eligrey has quit [Quit: Leaving]
Rusky has quit [Ping timeout: 240 seconds]
oldskirt_ is now known as oldskirt
<devyn>
ELLIOTTCABLE: potential problem for the future: I can't think of a good way to do multiple Paws reactors. it seems like they'd be blocking each other grabbing locks for internal structures all the time.
<devyn>
ELLIOTTCABLE: so effectively, you might as well not have multiple Reactors...
<devyn>
ELLIOTTCABLE: but maybe there's a better way
glowcoil has quit [Quit: Connection closed for inactivity]
yorick has joined #elliottcable
<SwooshyCueb>
lol
<purr>
lol
<SwooshyCueb>
<ELLIOTTCABLE> focus for now, because I am drunk, dumb, and easily sidetracked like a rabbit but not as fluffy
<SwooshyCueb>
<ELLIOTTCABLE> GOD I would love to be that fluffy
<SwooshyCueb>
<ELLIOTTCABLE> SwooshyCueb: TEACH ME TO EMBRACE MY INNER FURRY
<SwooshyCueb>
<ELLIOTTCABLE> I'm an incipient furry, I just know it
<ELLIOTTCABLE>
devyn: what internal structures would they share?
<ELLIOTTCABLE>
Oh, yeah, gotcha. that's intentional, afaict. the Table and Queue are global, and of course they're going to be locked-on.
<ELLIOTTCABLE>
not sure why that's a problem.
<devyn>
well
<devyn>
due to the rate of new combinations being generated
<devyn>
and the relatively small amount of computation per average combination
<devyn>
I think that would add up to a lot
<devyn>
of locking
<devyn>
ELLIOTTCABLE: I am here, by the way; it's unusual I know but hi
<ELLIOTTCABLE>
Yes, agreed.
<ELLIOTTCABLE>
But there's a related problem there.
<ELLIOTTCABLE>
paws is just slow. Multi-reactor or no.
<devyn>
the problem is more specific than that I think, in this case
<devyn>
there is no benefit really to multi-reactor Paws
<devyn>
even theoretically
<devyn>
because of this
<ELLIOTTCABLE>
A big problem is that the overhead of looking for do-able work (so, next() algorithm) happens **every combo** right now.
<ELLIOTTCABLE>
On my books is a way to refine the ordering rules so that multiple adjacent combinations are allowed to occur, on the stack, in series.
<ELLIOTTCABLE>
Once that's the case, then there's the chance of entire executions being processed in one go (which is really the eventual goal: we want the lowest level of abstraction to be as fast as possible).
<devyn>
alright
<devyn>
until then though, Paws.rs is going to be 1-reactor
<devyn>
even though Rust is very good at parallelism
prophile has joined #elliottcable
<devyn>
because it's not worth trying to implement parallelism for no benefit right now
<devyn>
haha
<devyn>
oh by the way
<devyn>
I just wrote advance() and I'm in the process of testing it to make sure it actually works
<ELLIOTTCABLE>
Excellent!
<ELLIOTTCABLE>
So. That concern is related to another thing that's been on my mind, for a while.
<ELLIOTTCABLE>
Discarding ordering.
<ELLIOTTCABLE>
At the moment, two operations that are not ordered relative to each other *by ownership*, are still ordered relatively to each other:
<ELLIOTTCABLE>
Not in a clear, defined, easily-manipulable way (without an unprecedented day-to-day understanding of every little action of the Paws machine; even *I* can't well-predict the ordering of two things),
<ELLIOTTCABLE>
but they're still deterministically ordered.
<ELLIOTTCABLE>
two conforming implementations, or the same implementation running the same code twice, will order things in the same way.
<ELLIOTTCABLE>
I'm not convinced this is a useful property, as batshit as that statement may sound.
<ELLIOTTCABLE>
It's possible to *remove* that restriction on ordering, with *potentially* huge performance consequences.
<ELLIOTTCABLE>
But again: "potentially."
<ELLIOTTCABLE>
Afraid to remove ordering in favour of *possible* (even if they're likely) performance gains; even if they're very likely.
<ELLIOTTCABLE>
whitequark taught me that >,>
<ELLIOTTCABLE>
Doing so is simple, though, if you'd like to write your implementation that way, see how it works out:
<ELLIOTTCABLE>
instead of the reactor being, basically,
<devyn>
ELLIOTTCABLE: what happens if a pristine Execution is advance()d? it's not in the spec
<devyn>
only advance() out of those four is described in the spec, at least explicitly
<ELLIOTTCABLE>
sorry, phone died
<ELLIOTTCABLE>
trying to rush and deal with a grumpy bff while having this conversation >:
<devyn>
:( then don't worry, I can just go to sleep :p
<ELLIOTTCABLE>
those are approximate names, not actually shit as named in the spec. sorry, in a hurry, so summarizing.
<ELLIOTTCABLE>
basically,
<ELLIOTTCABLE>
ugh nope I screwed it up anyway
<ELLIOTTCABLE>
re: pristine, that should definitely be in the spec
<devyn>
it's not though heh
<devyn>
it is, but, it's not described how to deal with it
<devyn>
if it's true
<ELLIOTTCABLE>
confused.
<ELLIOTTCABLE>
oh
<ELLIOTTCABLE>
ahhah
<ELLIOTTCABLE>
I *think*, not sure, have to go, but am pretty sure,
<ELLIOTTCABLE>
that first line should read:
<ELLIOTTCABLE>
If the pc is non-existent, the stack is empty, **then** the Execution is no longer pristine
<ELLIOTTCABLE>
hrm, nope
<ELLIOTTCABLE>
fuck me
<devyn>
anyway don't worry about it for now
<devyn>
I'll bug you later
<ELLIOTTCABLE>
I'm pretty sure the rest of the algorithm handles both pristine and non-pristine. But I'll double-check when I get back.
<ELLIOTTCABLE>
or, you can for me. If you implement it as it's spec'd (which is disgusting and horrible ugh) then you should be able to feed it pristine exec's
<ELLIOTTCABLE>
more later. talk soon much love.
<devyn>
<3
colbyl has joined #elliottcable
colbyl has joined #elliottcable
colbyl has quit [Changing host]
colbyl has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
colbyl has joined #elliottcable
colbyl has joined #elliottcable
colbyl has quit [Changing host]
prophile has quit [Quit: The Game]
manveru has quit [Ping timeout: 240 seconds]
colbyl has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
manveru has joined #elliottcable
colbyl has joined #elliottcable
colbyl has joined #elliottcable
colbyl has quit [Changing host]
colbyl has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Rusky has joined #elliottcable
glowcoil has joined #elliottcable
eligrey has joined #elliottcable
alexgordon has joined #elliottcable
vil has quit [Ping timeout: 265 seconds]
vil has joined #elliottcable
PragCypher has joined #elliottcable
prophile has joined #elliottcable
colbyl has joined #elliottcable
colbyl has joined #elliottcable
colbyl has quit [Changing host]
vil has quit [*.net *.split]
sharkbot has quit [*.net *.split]
nuck has quit [*.net *.split]
Sgeo has quit [*.net *.split]
ohhmaar has quit [*.net *.split]
erynofwales has quit [*.net *.split]
joelteon has quit [*.net *.split]
prophile has quit [*.net *.split]
eligrey has quit [*.net *.split]
vigs has quit [*.net *.split]
inimino has quit [*.net *.split]
ec has quit [*.net *.split]
jeannicolas has quit [*.net *.split]
colbyl has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
prophile has joined #elliottcable
joelteon has joined #elliottcable
erynofwales has joined #elliottcable
sharkbot has joined #elliottcable
vil has joined #elliottcable
ec has joined #elliottcable
jeannicolas has joined #elliottcable
ohhmaar has joined #elliottcable
vigs has joined #elliottcable
Sgeo has joined #elliottcable
nuck has joined #elliottcable
inimino has joined #elliottcable
eligrey has joined #elliottcable
ec_ has joined #elliottcable
ec_ has quit [Max SendQ exceeded]
colbyl has joined #elliottcable
colbyl has joined #elliottcable
colbyl has quit [Changing host]
colbyl has quit [Client Quit]
ec has quit [Write error: Connection reset by peer]