asciilifeform: lol i suppose nao i gotta actually read what they were sayin'..
asciilifeform: it's approx as solid as 'linux dunwork'
asciilifeform: mircea_popescu: i dun even know that the traditional orc intuition of 'rich ~= receives payola from miami' is such terrible heuristic
asciilifeform: it's a 2edged blade, as it would be a great temptation to 'light client' idjits to parasitize on trb. but would make for easy litmus.
asciilifeform: in-block tx, that is, rather than mempool
asciilifeform: btw trb could in principle be made entirely distinguishable, if we were to permit asking for a ~tx~ in 'inv' command ( currently prohibited , because prb nuked indexing , but could be brought into trb with no ill effect aside from some cpu cost )
asciilifeform: i strongly suspect that a very large % of the current set of 'nodes' won't pass.
asciilifeform: the 1 'litmus test' i was able to think of , is the 'i pick a block hash and you gimme the block in <1s' algo.☟︎
asciilifeform: ( granted such a survey won't distinguish 'pseudos' or outright shams, the only item that can do it is an actual noad )
asciilifeform: re upstack -- as it happens, 'trb-compat' is pretty easy to distinguish mechanically -- anybody who has 'services' field != 1, aint trb-compat.
asciilifeform: if yer nuking the manifest, good form would be to regenesis then imho. but yes
asciilifeform: manifest is a piece of handy docs, not a straightjacket, lol
asciilifeform: diana_coman: no? you can branch on manifest like on anyffing else
asciilifeform: further in pipeline, will be a thing that actually asks'em for a randomly-picked block, and sees if it keccaks to what i've got.
asciilifeform: individual display only of trb-compat folx.
asciilifeform: mircea_popescu: btw i think you'll like this, by default thing does opposite of the heathen one, doesn't display prb nodes individually, but simply as a collective 'sad people' stat # .
asciilifeform: speaking of trb, i dusted off an old conveyor item, phuctor-like www proggy that maps out noadez ( you give it ips, and it connects to'em, then issues 'version' and 'getaddr', and builds a graph )
asciilifeform: lol, could even describe e.g. trb this way
asciilifeform: ( recall, early steam engine actually had ~this )
asciilifeform: i'd rather have addressable sram , and explicit movement to/from a slower/cheaper brick
asciilifeform: personally i fucking hate branch prediction and other hacks around the slow main memory that for some reason there's no buyable alternative to ( cray is dead )
asciilifeform: it has a legit, sorta, use -- long loops ( where branch 'likely', given as it is taken for all but final shot ) but still ugh
asciilifeform: mircea_popescu: i find myself wondering how many physics folx sit and bite their lips, 'monte carlo dun work so good...', cuz they bought $50k swiss 'quantum rng' that whitens with sha.
asciilifeform: in related lulz, and iirc we had a thread ( tho can't seem to unearth it ), urandom is sufficiently sad that when asciilifeform was running genetic algos / monte carlo / etc. crapolade , yrs ago, from it... would regularly fail to converge☟︎
asciilifeform: ( they also make a few high-end DACs, etc., but this prolly would not keep'em alive by itself )
asciilifeform: mircea_popescu: phunphakt : the only reason, afaik, why ti still not bankrupt , is that they somehow managed to get their bottom line calcs to be ~mandated~ in schools in ~whole usa
asciilifeform: they had a line of pocket calculators, this was ( and as i understand, remains ) their flagship product. but! they somehow managed to lose their upstream ic vendors (they dun make much with own hands any moar) and nao reduced to selling same calc with ~ipnoje+emulator internals, and multi-second boot(yes) times
asciilifeform: pissed away pretty much all market , despite at one time heavy muscle
