<vaninwagen>
Candle: you could make it iso though through recover Rand(...) end
<vaninwagen>
And Xoroshiro128Plus has predictable first bits given random seeds? o_O
<SeanTAllen>
I believe the idea vaninwagen is that he wants to share the random number generator across acrots
<SeanTAllen>
also, good morning vaninwagen !
<vaninwagen>
Guten Morgen SeanTAllen
<vaninwagen>
Yeah that wouldnt work properly with multiple actors
_whitelogger has joined #ponylang
_whitelogger has joined #ponylang
endformationage has quit [Quit: WeeChat 1.9.1]
<vaninwagen>
Ive seen some persistent random number generator in a book about functional programming in scala, it should be feasible to build such a thing, that returns thr new state and the result on .next()
_whitelogger has joined #ponylang
<SeanTAllen>
Not sure for that helps here though
<SeanTAllen>
They'd end up using different generators
<vaninwagen>
True dat
<vaninwagen>
Ha, what if you just send the seeds, generated by 1 Rand instance
<vaninwagen>
B)
<vaninwagen>
Then each actor gets its own Rand
<vaninwagen>
BBB)
<SeanTAllen>
well the issue there would be that what i think candle wants is if for seed X, you get 1,4,7,10 as the numbers
<SeanTAllen>
that no matter how you call across actors you want them to get 1 then 4 etc
<SeanTAllen>
whereas sharing the seed
<SeanTAllen>
each would do 1,4,7,10
<SeanTAllen>
instead of picking from the sequence in a round robin fashion
<SeanTAllen>
two solutions i can see are promises (slower) or different generators in each actor, probably with different seeds and accept they are different
_whitelogger has joined #ponylang
<vaninwagen>
But if each actors Rand is based upon 1 seed from 1 and the same Rand, everys actors Rand is still transitively based on that first Rand and thus on the seed of that first Rand, so the output still deterministically depends on this 1 seed fed to the initializing Rand
<SeanTAllen>
right but
<SeanTAllen>
vaninwagen:
<SeanTAllen>
the goal isnt to get each actor to do 1,4,7,10 in order
<SeanTAllen>
its to get actor A to do 1 then actor B to do 4 then actor A when it calls again to do 7 etc
<SeanTAllen>
under your scenario it would be
<SeanTAllen>
actor A 1
<SeanTAllen>
actor B 1
<SeanTAllen>
actor A 4
<SeanTAllen>
i'm not aware of a scenario where that is behavior you would want
OtakuSenpai has joined #ponylang
OtakuSenpai has quit [Ping timeout: 244 seconds]
_whitelogger has joined #ponylang
endformationage has joined #ponylang
OtakuSenpai has joined #ponylang
<vaninwagen>
That is not the behavior i tried to describe - give each actor a different seed from 1 seeded Rand - so they have different random numbers
<SeanTAllen>
That's what I said to do
<SeanTAllen>
Not sure why you would give that as a counterpoint
<SeanTAllen>
Perhaps I wasn't clear
<vaninwagen>
Shit yeah, and i didnt read your answers correctly
<vaninwagen>
I blame sunday afternoon
<SeanTAllen>
that's a good blame
OtakuSenpai has quit [Quit: Leaving]
travis-ci has joined #ponylang
<travis-ci>
ponylang/ponyc#5426 (master - a57b4ff : [Main]): The build was fixed.