asciilifeform: it's shovel work, and necessary shovel work, just like maintaining trb.
asciilifeform: but would rather not suffer illusions re 'this is a step to designing gossipd'
asciilifeform: having said this, i'm all for setting up a not-fleanode . and will put relay on dulap as soon as the matter of 'which ircd' is settled.
asciilifeform: there's a severe narrowing of perspective that almost always comes from crawling around inside mine
asciilifeform: so ftr i will have to disagree with hanbot , in that imho coal miner is ~not~ the best geologist.
asciilifeform: ( and it took asciilifeform, in turn, 10+yrs , measured from time of 1st thinking about subj, to get there )
asciilifeform: or , say, take tcp. mircea_popescu aint even a programmer, and is just about as 'clean' as a fella can get in re programming radiation damage and still have worked with comp. but it took asciilifeform 3+yrs to get him to see that tcp is -- by design -- garbage
asciilifeform: the ~damage~ ( and imho is just about as factual as radiation damage ) from being a programmer, is that familiarity with crud , leads to swallowing, and eventually you learn to live with things that by all rights no one oughta live with. witness how many people think that it is acceptable for a program to sometimes crash, for instance.
asciilifeform: ( and if there's 1 thing there is no shortage of in the sad monkey house that passes for kompyooting, it's 'antipattern' -- there's enuff that 'you can eat it with arse' )
asciilifeform: sadly troo. otherwise microshit escapees would make for brilliant teachers of the art.
asciilifeform: obj i'm ill-equipped to say for sure, hanbot
asciilifeform: an asciilifeform who had never touched microshit, i suspect, would be a better man than asciilifeform who did..☟︎
asciilifeform: i think aristotle actually had , sadly, correct notion, where doing some kinds of work actually makes you dumber☟︎
asciilifeform: hanbot: this gets complicated. e.g. i put high value on mircea_popescu's pov re programming, even tho he (afaik) moar or less never programs
asciilifeform: fwiw i dun feel like my many yrs of setting up multi-MB balls of c ??? liquishit on various net-facing boxen, taught me ~anyffin that'd be applicable on hypothetical sane comp
asciilifeform: hanbot: but was curious whether you had moar specific notion, or simply gut feeling (nuffin wrong with gut, imho)
asciilifeform: hanbot: this is fair enuff, and mircea_popescu wrote at length on the subj.
asciilifeform: ( and specifically, why not also 'i'd trust operator who also had in the past set up winblowz 2012' then )
asciilifeform: hanbot: wouldja be willing to expand re why ?
asciilifeform: ( tbf, this was after his 1st 20yrs -- 100% wheelbarrow. during his 2nd 20y, he learned to stick thermometers in arses, became 'medic' , hence lived )
asciilifeform: ( he went in for trotskyism. and policy was, trotskyists only wheelbarrow, no cutting wood, no nuffin else )
asciilifeform: and he answered honestly, 'to roll wheelbarrow fulla rocks on narrow planks. is all i learned.'
asciilifeform: fella did 40+yrs in gulag. then got out, and was often asked by curious folx, 'what did you learn to do'
asciilifeform: diana_coman: ever read any varlaam shalamov ?
asciilifeform: for instance, in saeculum i've set up microshit boxen of various sorts. and could not honestly say that i learned from the experience anything applicable on sane-people planet.
asciilifeform: diana_coman: i hesitate to say 'no useful'. but possibly 'very little'.
asciilifeform: it isn't even that i dun think it's worth doing ( fleanode only ever gets ~moar~ rotten, and never less ) . but imho the cost is prolly not recoverable ( in the engineering sense ) .
asciilifeform: ( e.g. the problems of mitigating tcp ddos are irrelevant to proper udpistic gossipd. and ditto authentication of handles. )
asciilifeform: indeed will learn what the troubles are in practice. only i suspect that 0 of it will have any bearing on the gossipd side.
asciilifeform: diana_coman: plox to expand here : what in particular is to be learned ?
asciilifeform: diana_coman: i'll admit that it isn't clear to me how effort put into baking glue for oddball nonstandard ircisms helps in re gossipd . irc as i see it is an entirely dead-end tech ( rides on tcp, and 0 notion of crypto , and cannot be retrofitted really )
asciilifeform: folx without lymphocytes ~will~ succumb to fungi & errything else under the sun, it dun matter what they do
asciilifeform: rather than sumthing particular to irc.
asciilifeform: ~those~ succumb to fart in the wind, not even speaking of 'attack'
asciilifeform: BingoBoingo: 2014. and imho sounds like less of 'attack on irc' and moar of 'attack' on wotless heathens / allcomerists
asciilifeform: diana_coman et al : i'm also not 100% convinced that the proposed multi-network thing is actually less of a nut to crack than adult gossipd. ( consider, how wouldja do authentication ? the extant ircd's support either '~nuffin, anyone can hijack session' or ssltardism . )
asciilifeform: last time this came up , asciilifeform went on a dig of 'what ircd', and found (surprise!1111) MB of c liquishit
asciilifeform: really oughta spread across a minimum of 3 geographically separate racks, before troo move , imho
asciilifeform: mircea_popescu ( in linked thrd ) dug up a coupla possible 'wheres'. but imho oughta begin with gluing together small net ( with boxen in piz ~and elsewhere~ ) and ~then~ bridge it to public nets.
asciilifeform: this is because a n-bit gcd (as appears in ch15 ffa) is simply 2n n-bit subtractions, 4n n-bit shifts, 2n n-bit muxes, plus some small change.
asciilifeform: err, 4096b m-r (i.e. price of 2 modexps)
asciilifeform: if the upstack note was unclear, btw -- you can do ~2500 4096b gcd's for the price of 1 4096b modexp.☟︎
asciilifeform: bvt: ty for the link tho, it confirms the (already yrs-long) suspicion that d00d's head is fulla maggots nao.
asciilifeform: so presently i cannot think of a scenario where i'd want to reopen the case of gcd.☟︎
asciilifeform: constanttimeized stein's o(n^2) gcd ( http://www.loper-os.org/?p=2963 ) is not only imho fast enuff (even a magical 100fold speedup in it, would not affect speed of rsa key gen measurably , consider above ) but fits-in-head and has no error terms.
asciilifeform: ( modular exponentiation is , in all operations which feature it )
asciilifeform: that being said, gcd is not the bottleneck of anyffin in ffa.
asciilifeform: in re single-shot gcd, you either ~look at the bits~ , and that's lehmer's (not constant-time-able) method, or you do not, and that's stein's or the division-powered one, and i suspect it can be proven that yer stuck with quadratic runtime if you dun look at the bits.
asciilifeform expected to find that linked item is re ~mass~ gcd, in the sense of 'bernsteinization' used in phuctor. but apparently djb has nothing to add re ~that~
asciilifeform: funny how easy to tell 'the new djb' from 'the old'
asciilifeform: and he's doing what appears to be the classic dividing variant of gcd, + newton's method . so idea is to... introduce floatism!? ( and the assoc. eggogs )
asciilifeform: bvt: from surface look, loox like lulzy piece, e.g. 'The algorithm is not constant-time as shown but can be made constant-time with low overhead'
asciilifeform: aside from that, protocol per se also has well-known problems (the reliance on 'trackers', and general ease of 'poisoning')
asciilifeform: re 'torrent', consider fact that orig author is an unrepentant lulzcow and recidivist shitcoinist iirc☟︎
