#pest | 2023-02-10

phf: was pretty good demo, imho, of 'uncatchable joe'(tm)

01:22

( that, and/or the fact that erryone knew that very little could be achieved by hijacking a bouncer. )

~ 9 hours 43 minutes ~

11:06

of possible interest re yest'd thrd -- asciilifeform recalls, when testing ada udp, receiving coupla oddball 'martians' with no obvious structure

11:06

bitbot

Logged on 2023-02-09 22:25:36 asciilifeform[jonsykkel|awt|deedbot]: i.e. hypothetically, pest letsya treat the net as a (sloow) broadcast medium, if necessary.

11:07

i.e. entirely possible that sumbody already was doing precisely that.

11:07

( if had to guess who -- botnet client , to avoid hardcoding control box addr )

11:10

*

asciilifeform vaguely recalls sumbody running a 'ipv4 radiotelescope' and picking these up en masse, but not recalls any interesting conclusions

~ 2 hours 10 minutes ~

13:21

yeah, back when i got my first root box on a fast connection, i had a tcpdump running on the external device. the result was, as you would say, snooze

13:26

phf: in the cited case, not quite snooze; but (unsurprisingly) entropic turds, impossible to make sense of w/out knowing what they came out of (and, prolly, the key)

13:26

*

asciilifeform fwiw strongly suspects that variously pest-like nets have existed for many yrs

~ 1 hours 1 minutes ~

14:28

huh, all kinds of weird behaviors if you run same pest key from multiple stations :>

14:30

i guess that's to be expected, but there probably should be some awareness in the system of what to do when you start getting packets from different ips

14:30

per spec, oughta simply update AT each time

14:31

you'll get, no doubt, chain oddities

~ 16 minutes ~

14:47

( per spec, stations are distinguished by key strictly, i.e. absolutely no diff b/w a properly-config'd station using >1 ip and 'coarse error of pilotage' where 1 key on >1 box, other than the inevitable chain fuckery )

14:53

the only expectation is that a station is presumed to be able to receive packets at the addr $peer last saw it transmit from.

14:57

if seeing selfchain conflicts in a direct peer, pretty good heuristic for '1 key on >1 box'. wat-do in this case is outside scope of spec.

14:58

( it'll be visible in a half-decent gui pestron, and then wainot ask $peer 'wat'. maybe his key -- compromised. )

15:00

the upside of stateless protocol here is that both genuine peer and impostor can speak. so, naturally, pgp him a new key and throw out the old.

~ 1 hours 46 minutes ~

16:46

http://logs.bitdash.io/pest/2023-02-08#1022544 << it has been a while I'm trying to do this, but you actually need a sane pool of paper, not just anything that iacr, arxiv and other public paper archives get. Thus for now my little attempt is deferred until I get a good enough list of books, papers and so on. Currently I'm

16:46

bitbot

Logged on 2023-02-08 09:59:20 signpost: how about same over scientific papers, assuming for the moment that a corpus of real one could be had (the lack of which isn't the fault of this particular tech)

16:47

hosting it here, but the link to the code is simply wrong. https://unpx.net/corpus/

~ 20 minutes ~

17:07

unpx: loox like only has search box ?

17:07

( where's the list of warez ? )

17:11

*

asciilifeform unenthusiastic re 'scihub'-type warez for simple reason that problem is that ~nobody is writing ~anyffin actually worth reading, rather than 'ohnoez, can't find $paper'

17:11

I think you may be able to list everything by searching for a very common word (e.g. 'a'), I don't recall limiting the results nor paging it, so there you have your list. But this isn't the good list, you can see it as a download dir.

17:12

*

asciilifeform in all cases where actually needed $paper, found trivially

17:13

asciilifeform, too many papers are hidden behind paywalls. i have scihub on speeddial

17:14

and not just luby, but foundational stuff

17:17

but kind of goes back to many a conversation and threads both in b-a, t and a, of the need for "rebuild civ from first principles" archive

17:19

perhaps once we have running luby

~ 27 minutes ~

17:46

> ~anyffin actually worth reading <<< the only way I found is reading what other people suggest and slowly compile my own list of suggested readings.

17:47

There was a time I wanted to create a protocol for this, but the point isn't in the protocol, rather in the network and its primary seeders: readers.

~ 17 minutes ~

18:05

but then your papers should be in some kind of context? like come with "who recommended" "relevant project" "relevant goal" ,etc.

↖

~ 1 hours 29 minutes ~

19:35

phf, unpx: fwiw all of the papers asciilifeform 'used' , are mirrored on his www, in the format seen in earlier link. imho this is the way to do it.

↖

19:35

the jpgs naturally aint searchable, but to date this also kept the copyrasty bots out.

19:36