asciilifeform: hey trinque , what's the 'magic number' it oughta wait ?
asciilifeform: i'ma have to make a moar hightech rss mechanism, with queuing etc
asciilifeform realizes with some disappointment that most of today's pops will overflow the deedbot bucket; thing's eaten moar keyz in 20min than typically eats in 2days
asciilifeform: mircea_popescu: lemme know if you find anything other than the stats page ( still O(N) for nao ) that's >1s to load.
asciilifeform: ty to trinque , phf , mircea_popescu , the folx who gave asciilifeform sqltronics hints.
asciilifeform: ( i have nfi whether anyone was holding out for the cured db; but will for the record note : the time -- has come. )
asciilifeform: i will note, in case it wasn't obvious earlier : folx who previously were refraining from linking phuctor somewhere on account of 'how could it take the heat' are nao invited to open the throttle.☟︎
asciilifeform: the day aint far when the thing will be able to tell you , e.g., 'yer router is nsatronic', the minute you plug it in (i.e. realtime scan an' probe-against-db, no need to queue'em )
asciilifeform: used hashistic index followed by collision test
asciilifeform: mircea_popescu: meanwhile i made it do it ( not deployed to production yet, still in test.. )
asciilifeform: ^ mircea_popescu , ben_vulpes , other sqlists..? ^
asciilifeform: trinque: turns out, phuctor www end spends ~80% of cpu cycles in sequential scan for modulus, because index refused to be created. ( and no, 'hash them' is not solution, phuctor gotta be able to store ANY set of moduli, and not merely e.g. 512bits of possible ones )
asciilifeform: hey trinque , is it possible to cleanly remove the 8191-byte limitation for indexed fields in postgresql ?
asciilifeform: trinque: determination is orthogonal to correctness, elementarily; and so i suspect that most folx's experience with inventors, should they have any, resembles this one, 'fella worked on perpetuum mobile under meth for 20yrs , then went broke, drank'
asciilifeform: or are we talking about a d00d circa 1992 or somesuch
asciilifeform: surely had a notion that, e.g., spam, exists ?
asciilifeform: rright, i can picture this-much, but how did d00d's scheme get around the 'all-comers can make you spend 9000 cpu cycles whenever for the asking' problem.
asciilifeform: i admit to a little curiosity re what 'turn into distributed db' meant ( can query it ? who supplies the 10^9000 cpu cycles ..? )
asciilifeform: trinque: was it the sql master d00d from old thread ?
asciilifeform: ( iirc in old thread we discussed this: asciilifeform suspects that 'subjective lag' is not a function of absolute delay, but of ~variation~ in same )
asciilifeform: tho i wonder how he measured the bolix 3620; back when i had mine, it most certainly did not 'feel like' winblowz-levels of i/o lag
asciilifeform: mno, 'crystal' is what i arbitrarily titled the tar-like format .
asciilifeform: if you display the thing in phf's viewer, or mod6's , you will see a vertical 'pillar' of flow, rather than the familiar tree.
asciilifeform: mod6: the included illustration was to the effect of 'what if trb had been made using this type of vtron, from genesis to present day' .
asciilifeform: mod6: note, even tho the proggy ~could~ be used as-is, i doubt that anybody would want to; it is really demo of a potential frontend to be built into vtron .