punkman: https://pthree.org/2017/12/22/the-entropy-of-a-digital-camera-ccd-cmos-sensor/ "when putting the generated binary files through the Dieharder tests, it comes out pretty bad. I get 20 "PASSED", 13 "WEAK", and 81 "FAILED" results."
punkman: just googled the vault12 app, and of course it's a wallet
punkman: their blog post about the their rng https://medium.com/vault12/how-to-get-true-randomness-from-your-apple-device-with-particle-physics-and-thermal-entropy-a9d47ca80c9b
punkman: "TxProbe: Discovering Bitcoin’s Network Topology
punkman: Using Orphan Transactions" https://arxiv.org/pdf/1812.00942.pdf
adlai: where's the angry old "icanhazpdf|ocr.slut" spammer when you need him!?
adlai: ... which reminds me, I was gonna properly phrase some questions about OCR, for this place, since iirc asciilifeform [and maybe others?] have experience scanning books
adlai wonders how bad the quality of printed material has to get before the publishing industry thanks him personally for writing an 'unpub' tool
adlai: unlike bitcoin mining (i.e. a tool for censorship), 'unpub' is not a tool for ogilvisation; rather, it'd hope to someday be packaged as 'pdf2tex', with no delusions whatsoever of ever being 'pdf2latex'
adlai: the question that led to this dangling thread is quite boring: "have you ever OCRed sheet music?"
adlai: as for the old entropy question: working almost entirely off my readings of the past conversations on this, I'm surprised ~any~ peripheral built as an input device, in this case optical, is considered a good entropy source
adlai: perhaps a better metric for shootouts is (kb/s)/(J/s), a sorta 'miles per gallon' for noise generation
adlai: this metric would probably cause the camera approach to fail miserably, given how much extra computation would get spent making the frames coherent, and then debiasing... useless heat
adlai defaults to "musst nicht sprechen" re:1812.00942; after all, the only thing he understands worse than bitcoin, is topology
punkman: used tesseract-ocr many years ago, wasn't getting great results. they do have neural-something in newer version though.
asciilifeform: ocr does't work, adlai et al.
asciilifeform: ('99.999%' != 'works' when it comes to potentially gnarly maffs typo)
asciilifeform: it 'works' well enuff for fiction on ru warez sites, but this only because 'over 9000' readers and they send in corrections.
asciilifeform: fughet about a 'pdf2tex'.
asciilifeform: ai-complete(tm)(r)(c).
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-07-30#1049382 << 'a remarkable bear, but not a remarkable dancer'
dulapbot: Logged on 2021-07-30 08:25:03 adlai: as for the old entropy question: working almost entirely off my readings of the past conversations on this, I'm surprised ~any~ peripheral built as an input device, in this case optical, is considered a good entropy source
asciilifeform does not, ftr, rec the use of such camera ~optically~, but as a (poor) particle counter.
dulapbot: Logged on 2021-07-29 15:23:59 asciilifeform: but i've yet to see someone who is using camera as trng do this. (btw if you must do so, copper foil over the objective.)
asciilifeform: ~if you must~.
dulapbot: Logged on 2021-07-28 23:13:45 asciilifeform: whaack: imho it's a little like asking 'what household object makes the best hammer?' -- buy a fucking hammer.
shinohai: And here I've been using a microscope all this time.
whaack: !b 6
whaack: good morning
asciilifeform: wb whaack
whaack: asciilifeform: is there a way to attack a bias'd rng without knowing what the bias is?
asciilifeform: whaack: certainly. you find out what it is! lol
whaack is pondering
asciilifeform: whaack: note that bias in the usual sense aint the only possible rng failure. there's also, e.g., sequential correlation.
asciilifeform: whaack: not to mention 'debianization' where keyspace is 2^smallint instead of bigint, etc
whaack: asciilifeform: ah, i may have been grouping the terms bias and sequential correlation together. If the only 'bias' is that bit 1 appears more frequently, then since all rng input can be transposed to a bitstream, attacking an rng with this type of bias could be done by prioritizing the search of all N bit streams with arrays starting with all 1s (most individually likely outcome) , then all 1s
whaack: with one 0, etc.
asciilifeform: whaack: in practice, bias in rng comes usually in discrete ('erry 5th bit is a 1') rather than continuous ('weighted coin') form
asciilifeform: depending on what the victim is doing with his broken rng, different approach (e.g. if rsa or ecdsa -- solve lattice)
cgra: asciilifeform: now that we're on topic, i've been for a while wanting to ask a q. at the bottom of nosuchlabs.com front page, you say "We recommend at least 24 hours of entropy distillation (solely via XOR-in-place !) if generating mission-critical, long-term cryptographic keys."
cgra: is this is just because 'there's no 100%'? ie. humans make mistakes and only *mostly* notice
punkman: whaack: one of the papers I linked recently, found bias in btc tx nonce, and recovered private key due to that bias.
asciilifeform: cgra: per 'xor lemma', distillation via xor cannot subtract entropy under any circumstances (aside from feedback!) -- hence if you have the time, it doesn't hurt.
punkman: asciilifeform: what does "feedback" mean?
asciilifeform: punkman: arse-mouth connection. i.e. if the source of what you xor has nonzero info re what you're xoring with.
punkman: http://logs.nosuchlabs.com/log/asciilifeform/2021-07-30#1049394 << I've seen this in papers. One of them arrived at "put thumb over camera".
dulapbot: Logged on 2021-07-30 12:49:31 asciilifeform: does not, ftr, rec the use of such camera ~optically~, but as a (poor) particle counter.
asciilifeform: lol why thumb?!
punkman: always have thumb
punkman: actually, I think their scheme was thumb over camera and flash led, to have ~uniform "scene" with photons coming in
asciilifeform: punkman: sounds like a good way to make accidental pulse oximeter, lol
asciilifeform: instead of rng.
punkman: heh
punkman: asciilifeform: so with copper you get photons from photoelectric effect?
asciilifeform: punkman: afaik most of what you get is from thermal electrons, rather than energetic photons making their way through the foil.
punkman: asciilifeform: that's what I meant yeah
asciilifeform: notbad likbez re subj if anyone givesashit
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-07-30#1049376 << largely for noobs, will point out that trb tolerates neither tx nor block 'orphans'. they are intrinsically a denial of service vector whereby randos can eat arbitrary memory.
dulapbot: Logged on 2021-07-30 06:24:01 punkman: Using Orphan Transactions" https://arxiv.org/pdf/1812.00942.pdf
asciilifeform considers the removal of that garbage to be one of his very few -- perhaps the only -- nontrivial contributions to trb
asciilifeform: !w poll
watchglass: Polling 17 nodes...
watchglass: 185.85.38.54:8333 : Could not connect!
watchglass: 185.163.46.29:8333 : Could not connect!
watchglass: 205.134.172.6:8333 : (172-6.core.ai.net) Alive: (0.081s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=693427
watchglass: 205.134.172.4:8333 : (172-4.core.ai.net) Alive: (0.042s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=693450
watchglass: 205.134.172.26:8333 : Alive: (0.143s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=693444
watchglass: 54.39.156.171:8333 : (ns562940.ip-54-39-156.net) Alive: (0.111s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=693450
watchglass: 71.191.220.241:8333 : (pool-71-191-220-241.washdc.fios.verizon.net) Alive: (0.093s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=693450 (Operator: asciilifeform)
watchglass: 205.134.172.28:8333 : Alive: (0.083s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=693450 (Operator: whaack)
watchglass: 208.94.240.42:8333 : Alive: (0.160s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=693450
watchglass: 54.38.94.63:8333 : (ns3140226.ip-54-38-94.eu) Alive: (0.322s) V=88888 (/therealbitcoin.org:0.8.88.88/) Jumpers=0x1 (TRB-Compat.) Blocks=693450
watchglass: 143.202.160.10:8333 : Alive: (0.234s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=693450
watchglass: 213.109.238.156:8333 : Alive: (0.389s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=693450
watchglass: 205.134.172.27:8333 : Alive: (0.755s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=693450 (Operator: asciilifeform)
watchglass: 84.16.46.130:8333 : Could not connect!
watchglass: 176.9.59.199:8333 : Violated BTC Protocol: Bad header length! (Operator: jurov)
watchglass: 103.36.92.112:8333 : Violated BTC Protocol: Bad header length!
watchglass: 192.151.158.26:8333 : Busy? (No answer in 100 sec.)
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-07-30#1049412 << imho scoopbot really oughta say the www name also
asciilifeform: y'know, like feedbot did
PeterL: you mean name of website?
PeterL: I can probably add that
asciilifeform: aha
signpost: yeah, when reading from the logs it's not so clear
PeterL: like "new post on Thimbron:" then the link as it is above, or just the link and title without the bracket notation?
asciilifeform: the former
PeterL: sure
asciilifeform: PeterL: ty
shinohai: thimbronion: Can has plaintext of alcuin code? I still am unable to get properly formatted code trying to copy from WP page.
whaack seconds shinohai's request
asciilifeform: vpatch plz.
asciilifeform observes that thimbronion made a py demo of perhaps 20% of asciilifeform's (yet unpublished!) spec
asciilifeform: which is notbad
asciilifeform: ( there are some nontrivial differences but essential idea is just about same )
dulapbot: Logged on 2021-07-28 21:00:58 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-07-28#1049144 << ftr the indeterminate-width fields of irc (and similar protocols) don't play well with udp. ( in asciilifeform's thing -- fixed-length fields. for absolutely everything, defined as it were hardware )
PeterL: asciilifeform: are you writing your thing in Ada?
PeterL: asciilifeform: do you have preference on whitespace around items in the bracket notation?
PeterL: how does this look for a format? "%s: [ %s ][ %s ]" % (blog_title, item['link'], item['title'])
shinohai: $ticker btc usd
busybot: Current BTC price in USD: $41220.38
PeterL: !s uptime
scoopbot: PeterL: time since my last reconnect : 0d 0h 0m
PeterL: It should be more clear now
thimbronion: shinohai, whaack: does this work for you? http://share.alethepedia.com/alcuin/genesis.vdiff
thimbronion: asciilifeform: will attempt to implement encryption this weekend. I can tighten up field lengths when I have more info.
shinohai: tyvm thimbronion looks proper.
shinohai: thimbronion: is specific version of python required to run ?
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-07-30#1049465 << regarding the wp code viewer, users unfortunately must manually escape html characters (see this footnote containing a sed script which handles all the offending characters)
dulapbot: Logged on 2021-07-30 18:14:24 shinohai: thimbronion: Can has plaintext of alcuin code? I still am unable to get properly formatted code trying to copy from WP page.
billymg: but otherwise i copy/pasted what thimbronion has in his post and diffed it against the genesis he added and there was only one line that got munged L318
thimbronion: shinohai: I'm using python 2.7 from inside pyenv.
thimbronion: billymg: I manually escaped those two right brackets - I guess the escaping doesn't get escaped!
billymg: thimbronion: i'm also using this extra bit of css to widen the code blocks, which i think improves legibility quite a bit, e.g. here
billymg: thimbronion: haha yeah, not that kind of escaping
billymg: needs to be the html kind, e.g. ]
thimbronion: ooooooh ok.
shinohai: thanx thimbronion will try from inside pyenv, got "ImportError: No module named lib.server" otherwise :/
thimbronion: shinohai: it appears that one particular file didn't make it into the diff. You need an empty __init__.py file inside the lib dir for imports to work.
thimbronion: I'm not sure why vdiff would have skipped that file.
thimbronion: shinohai: here's a tarball that has everything: http://share.alethepedia.com/alcuin/alcuin-9999.tar.gz
shinohai: Hmm when I did the genesis I got a `lib/` directory and it contains:
shinohai: channel.py client.py funcs.py infosec.py peer.py __pycache__ server.py
shinohai: so yeah missing the init
shinohai: aha thimbronion just `touch __init__.py` indeed fixed it thanx
shinohai: I'm guessing because diff ignores empty files is reason that didn't make it into genesis.