log☇︎
44900+ entries in 0.345s
asciilifeform: out of fear of usg ? or from vanity ( he thinks it can be sold for millions ? to whom ? ) he is burning his life's work. i have nfi why.
asciilifeform: somebody shares ? i'll read, even if he also goes on re 'racism'
a111: Logged on 2018-01-17 20:24 fromloper: asciilifeform: I have noticed that some of the former Symbolics employees are fairly active on Twitter, like https://twitter.com/swmckay (one of the developers of VLM) and https://twitter.com/KalmanReti , do you think it is possible to pry any information out of them?
asciilifeform: http://btcbase.org/log/2018-01-17#1771995 << in these pages however all i see is claptrap re 'racism of trump', 0 about lispm ☝︎
fromloper: asciilifeform: very good, I'm glad I brought it to your attention
asciilifeform: somehow we cannot find common language, i have nfi.
fromloper: asciilifeform: I have noticed that some of the former Symbolics employees are fairly active on Twitter, like https://twitter.com/swmckay (one of the developers of VLM) and https://twitter.com/KalmanReti , do you think it is possible to pry any information out of them? ☟︎
fromloper: though it's still an early version of the architecture, revision 0 if I understand correctly
asciilifeform: fromloper: i'll confirm, this is a larger and moar detailed document than old one.
asciilifeform: apeloyee: look at the trb tree, and picture what the mass of the patches would have been, if this requirement had been in effect when i made it. ☟︎
apeloyee: I proposed being able to name arbitrary required antecedents << also probably needs a mechanism to declare "there are no other files in the tree" ☟︎
asciilifeform: i do not have, nor ever had, a working 'ivory' of any description.
asciilifeform: fromloper: currently however i do not even know where the power supply pins are, much less bus addressing, i/o, clock, etc
fromloper: asciilifeform: if I remember, you wanted to hook this intact chip to some emulation of Ivory's life support
asciilifeform: but i am speaking of ~use~.
apeloyee: "may be a 50kg sword" << doesn't seem to be. can be retrofitted into an existing design. as i said above "there needs to be a tree hash in the _leaf_ patch. and it MUST match the resulting tree"
asciilifeform: so i'ma save mine for proper commercial lab. but that means potentially forever. maybe whoever takes it off my corpse, can get it photo'd.
asciilifeform: fromloper: i only have two 'ivory' chips, and ideally would like to leave one intact , for active test . iirc phf also has 2 to use.
fromloper: I see, hopefully phf will have more luck with Zeptobars
asciilifeform: apeloyee: observe , i do not actually disagree .
asciilifeform: fromloper: that's where it stopped. i do not have 25k usd to use on ivory die photo.
asciilifeform: fromloper: i shopped around in commercial labs; the best bid was in the neighbourhood of 25,000 usd.
asciilifeform: fromloper: i'ma look. ty
asciilifeform: fromloper: pretty sure i've seen this before
fromloper: asciilifeform: someone uploaded "I-Machine Architecture Specification" to Bitsavers three days ago, I thought you might find it interesting: http://www.bitsavers.org/pdf/symbolics/I_Machine/I-Machine%20Architecture%20Specification.pdf ☟︎
trinque: then we are closer than it appeared in the long thread. I proposed being able to name arbitrary required antecedents in a vpatch's header, and this appears equivalent in effect to copying the file in whole.
asciilifeform: as i did in trb.
trinque: asciilifeform: to see if I can restate your opinion back to you, if I edit (as single author) both readme.txt and doesallthework.adb in separate vpatches, your view is I combine those into a single vpatch, if I want to build atop both in a new vpatch?
asciilifeform: i don't have a good counter-argument to this.
trinque: pretty obvious I'm saying signed hash, i.e. hash is in the signed vpatch
asciilifeform: i was speaking of brazil
asciilifeform: as i described in the linked thread, forcing the entire program under the antecedent hasher is not free
asciilifeform: hey it's the only kind i have nao!11
asciilifeform: apeloyee: trinque and mircea_popescu would like to put more of it on the machine. i haven't with what to dissuade them, it is a philosophical q, not even technical.
asciilifeform: apeloyee: see the quite 'flammable' log from that thread. i put the burden of correct operation ~100% on the human operator. ☟︎
asciilifeform: i must've missed something
Covale`: http://btcbase.org/log/2018-01-17#1771610 - no, not mine, I don´t have one. Neither do I have that cup, but it´s a pretty decent Lenin nonetheless ☝︎
asciilifeform: if gcc actually puts in the 'jump-likely' . currently i have nfi whether it does.
asciilifeform: i can't picture how without a megatonne of conditionaljumps. but maybe just failure of my imagination.
freetlas: Wow, first time I see too many people in here :o
asciilifeform: ( i'd rather not. but there is nothing fundamentally unusable about it )
asciilifeform: ( i will omit the rest of the mechanism, i think it is pretty obvious )
asciilifeform: nao that i think about it, it doesn't even have to introduce 'access types' ( pointers in ada ) , can use ordinary integers
asciilifeform: whether this is cleaner than the existing item, i will leave up to the readers, incl. apeloyee .
asciilifeform: however this introduces explicit pointerism. ( though, i will add, NOT pointer-arithmetism ) ☟︎
asciilifeform: or for another example, take the ugliness and 'pointericity' of the traditional 'pivoting' form of karatsuba. which i killed by forcing all FZ bitnesses to be powers of 2.
asciilifeform: ftr several different items in ffa seemed to me to be 'five-angled heptagons' (starting with how to compute the asm-less addition carries) until i solved'em
asciilifeform: or maybe not. i say it is an open problem.
asciilifeform: i don't see any reason why the thing should ~rely~ on adatronicn bounds checks for correctness.
asciilifeform: ( i.e. equal to the arity of the highest-arity stack op )
asciilifeform: i was about to add :
a111: Logged on 2016-01-21 13:29 asciilifeform: 'if i make it what i think is the right size, it crashes!111'
asciilifeform: i.e. if you see an array reference, you ~know~ that it is valid, because of the type of the index.
apeloyee: well it'd turn a pop from empty stack into a range-dipping eggog instantly << the reason I even suggested that 3 weeks ago
apeloyee: SP _is_ a cursor, I'm merely suggesting to treat it as such
asciilifeform: hmm, what if one were to model SP as in apeloyee's analogy, the text editor cursor. i.e. the pos of the ~next~ valid stack cell, rather than 'current'.
asciilifeform: one way i considered doing this, is to do away with all such things as 'SP - 1', 'SP - 2', etc. and instead to have e.g. Get_Stack_First, Get_Stack_Second, etc., each of which individually would ensure that the desired element exists. but these would have to return 'access type' (pointers) which thus far i've avoided using .
asciilifeform: i'd like to resolve the apparent contradiction between 'all possible SP positions are valid dereferences' and 'stack can be empty'.
asciilifeform: apeloyee: well yes, i get it. but i DON'T LIKE it.
asciilifeform: i.e. in all cases where the array has a custom indexing type as the index, the array exists over the entire range of said type.
asciilifeform: apeloyee: i dunget why absurd ? my other arrays behave this way.
apeloyee: can you think of a way to have the range of SP and of Stack be the same << this is plainly absurd. a N-sized stack has (modulo contents) N+1 possible states: "0 elements", "1 element", ..., "N elements". i.e.the ranges MUST differ, by exactly one.
shinohai: Gotta sell all my Bitcorn before the Missus finds out I spent all our monies on the Bitconnect ponzi.
asciilifeform: i'm still curious re the SP tho
asciilifeform: this is the #1 type of stylistic fix that i look for -- 'can this be made more obviously-correct-when-used-as-prescribed' and 'can use-as-prescribed be made more obvious'
asciilifeform: possibly one can make a SPARKistic proof of non-aliasing , for proggy taken as a whole. i'ma look into it.
asciilifeform: ffa itself is a sort of tightrope walk, an attempt to 'и рыбку съесть и нахуй сесть' . it isn't actually possible to make all of the routines able to take every conceivable kind of compile-time abuse ( which i'd argue overlap of arguments, is ) without conditional jumps.
apeloyee: I think everything up to ch.5 is safe as long as arguments are either equal or don't overlap
asciilifeform: the 1 caveat re buffers, is that there is a practical limit as to what can be made anticlobbering: if i were to do it to items that occur in inner loops of O(n^3)istic items, e.g. mux, proggy will end up 'geological' 4ever.
asciilifeform: this is a bit of a downer, i confess that i hoped apeloyee had found a more subtle, lethal boojum.
asciilifeform: i'ma buffer'em.
asciilifeform: but i can't help but agree with apeloyee re the Product in FZ_Mod_Exp , it gotta be buffered.
asciilifeform: because there isn't actually a limit as to what less-clueful people can break, regardless of what i do
esthlos: I will proceed one step at a time
asciilifeform: i never proclaimed, fwiw, that all ffa routines must be able to cleanly walk over own inputs.
asciilifeform: that part i get, it doesn't have anything to do with the clobbering of the modulus
asciilifeform: i'ma listen
apeloyee: Hear, hear! If asciilifeform would agree to take discussion of why 1+1=2 elsewhere, I'll do.
apeloyee: MustNotZero has 0 to do with Stack(0) << I know.
shinohai: You can play with pen and paper, I'm sure, if you find a method for so doing.
diana_coman: esthlos, certainly, not an issue; I think you are overthinking/overreading into this
esthlos: diana_coman I didn't mean to imply it would be "official" (tm) client
asciilifeform: and no, not 'enough', i'd like to find out wtf apeloyee was thinking
a111: Logged on 2018-01-17 11:28 apeloyee: http://btcbase.org/log/2018-01-16#1771004 << it's not me that does it at the expense of obvious correctness, but _you_. instead of having the compiler enforce no access to Stack(0), one is forced to rely solely on manual checking. But enough about that trivial matter; I'll just make the change locally.
a111: Logged on 2018-01-17 11:22 apeloyee: http://btcbase.org/log/2018-01-16#1771008 << I doubt you actually understood why (If you had, you would be a lot more worried, per http://btcbase.org/log/2018-01-16#1771010 ). here's a simpler version of the puzzle: http://p.bvulpes.com/pastes/UmpSH/?raw=true . has exactly the same bug.
asciilifeform: http://btcbase.org/log/2018-01-17#1771637 << your item clobbers the modulus as it runs. what i dunget, is why this would justify making the range of SP wider than Stack itself, as seen in your paste earlier. ☝︎
shinohai: I started some work on this over a year ago, simply too much work for me because of irl obligations unfortunately.
esthlos: shinohai: I've targeted this as my first real contribution to la serenissima
apeloyee: http://btcbase.org/log/2018-01-16#1771004 << it's not me that does it at the expense of obvious correctness, but _you_. instead of having the compiler enforce no access to Stack(0), one is forced to rely solely on manual checking. But enough about that trivial matter; I'll just make the change locally. ☝︎☟︎
a111: Logged on 2018-01-16 15:15 asciilifeform: if anyone finds so much as the smell of one -- i would like to hear about it, asap.
a111: Logged on 2018-01-16 15:15 asciilifeform: and yes his example snippet ~will~ barf. and no i won't spoil the puzzle by saying where and how. and no it is not hard to make 9000 similar examples.
apeloyee: http://btcbase.org/log/2018-01-16#1771008 << I doubt you actually understood why (If you had, you would be a lot more worried, per http://btcbase.org/log/2018-01-16#1771010 ). here's a simpler version of the puzzle: http://p.bvulpes.com/pastes/UmpSH/?raw=true . has exactly the same bug. ☝︎☝︎☟︎
a111: Logged on 2018-01-17 04:31 phf: oh, hah, i remember this "bytebeat" stuff, before it had a fancy name
mircea_popescu: http://btcbase.org/log/2018-01-17#1771626 << shit i used to make these! 1992 represent! ☝︎
phf: oh, hah, i remember this "bytebeat" stuff, before it had a fancy name ☟︎
ben_vulpes: yeah i know shinohai
a111: Logged on 2018-01-16 19:35 asciilifeform: today i gotta do this with own hands. because nobody else will, for any money.
asciilifeform: mircea_popescu: yeah, i reread and saw
mircea_popescu: i have nfi which is which, but anyway.
asciilifeform: 'seals', 'deltas', 'rangers', at least 3 others that i can't recall immediately from memory