#asciilifeform | 2022-02-25

(trilema) 2015-10-31 asciilifeform: see, if symbolics corp. had won the war and the notion of 'character' was not braindamagedly bound to 'octet' - we would not be having this conversation.

01:29

(trilema) 2015-10-31 mircea_popescu: explain this to me. what the fuck is a "sane notion of text" ?

01:29

(trilema) 2015-10-31 asciilifeform: one that fits will into the architecture of the machine.

~ 1 hours 29 minutes ~

02:58

re backscatter. I get a massage every single time at the airport. I must be denser than average

03:04

does texas still have those pickup mounted emp cannons?

03:08

i'd love to read verisimilitude's pest spec commentry/latin translation

03:13

I'm nowhere near that level of fluency yet, vex.

03:13

I've noticed I can now sight-read earlier chapters of my book, however, which felt so good to discover.

03:14

I'm inspired to write an explaination of alfs noise machine in drunk aussie

03:25

fuck, goats!

~ 27 minutes ~

03:52

any thing interesting from .ua gals who `went to cornell' yet shinohai? ie camwhores

03:58

I suppose they'll cut the feed before 1girl1tank.mpeg

04:04

Aplogies, we just threw a track out the front of your house. I will take some hours to fix...

04:06

brr is cold

~ 18 minutes ~

04:25

if you thought qqq was funny sp. qyld writes covered calls

04:37

what's the actual entry to `qualified investor' in usa mats?

04:38

presumably enuff dollars that if you ruin your irs work, you're worth hunting

04:52

https://www.youtube.com/watch?v=vuK5UUokdFs

~ 2 hours 47 minutes ~

07:39

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-21#1080249 << presumably u mean its not clear that the parser errors with any non-roman input as it shud

07:39

Logged on 2022-02-21 20:56:54 verisimilitude: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-19#1079981 This works, I tested, but it's not clear to me it always works; I quite like the little trick I used for mine, using the correctness of one to prove the other.

07:40

idk either, cuz dunno commonlisp, dont have a complete pic in hed of its behavior. it does return 0 on "", add (if (equal x "") (error "NOT ROMAN NUMBER")). otherwise intended algo clearly vork

07:40

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-21#1080251 << good

07:40

Logged on 2022-02-21 20:56:57 verisimilitude: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-19#1079982 I intend to have one at some point, jonsykkel.

07:41

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-21#1080256 << im dont undersand wat new code: 29 is here, a message?

07:41

Logged on 2022-02-21 21:00:10 verisimilitude: However, consider a system where those two words and their trailing punctuation fit in two thirty-two bit words; consider that the system wouldn't parse them, but simply know of the message ahead of time, and compare verbatim templates for the context until it determines which one matches; consider that new code of 29 isn't a number to be parsed, but inlined in the message as a known number.

07:47

Yes and yes.

07:47

The message is ``new code: 29'', but it's represented like that, and needs no parsing.

07:50

Isn't that neat?

07:51

maybe but i still dont undersand

07:51

how do they fit in two 32bit words

~ 18 minutes ~

08:09

Read this.

~ 1 hours 33 minutes ~

09:43

verisimilitude: aight, readed. but it creates larger # of q's

09:43

sounds like you would need 30k lines code and 1gig of cpu cache to make sense of ur compressed text in reasonable time

↖ ↖

09:43

you wont be able to encode this article using your system, since u made use of words "Seperating" and "inling" which was not in dict. but i guess your spellchecking feature would catch it

09:43

u mention updating dictionary with new words, who decides which words have this privilege, the uniword consortium? and who will do teh work of updating all softs taht use "text" feature. if reorganize dict, how to communicate with the 80% of your fellow text users that havent updated yet

09:43

dictionary would have to be append only or smth, unless you want to live in dynamically linked text hell

09:44

let me propose alternative "sykkel" encoding to be enforced on all planet inhabitants - identical to windows-1252. if u need to minimize bits u lz4 it, and if u insist on using weird chars that dont exist, mathspeak or similar, u send photo of the text

~ 5 hours 13 minutes ~

14:57

... cement testbed at 458k.

14:58

*

asciilifeform agrees w/ jonsykkel , has 0 interest in 'encode all text as pointers to oxford dict' scheme or any variant thereof

~ 1 hours 14 minutes ~

16:13

finally getting back to working on the crawler, i've implemented geolocation (ty for the recommendation punkman) and time series data collection, for charting

16:13

bitbot

Logged on 2021-08-11 09:21:57 punkman: here's link https://www.maxmind.com/en/geolite2/signup?lang=en

16:13

here are the active TRB nodes with their locations: http://paste.deedbot.org/?id=iAge

16:16

also unjammed my trbexplorer instance by rebuilding with jfw's dumpblock patch, it's been chugging along since yesterday and now at 639k

16:16

bitbot

Logged on 2022-02-22 23:55:06 billymg: whaack: at block 619140 now, it hit the orphan loop at 619141 since my trb isn't built with jfw's dumpblock patch

16:19

billymg: afaik that one, whaack's, and asciilifeform's patch were equivalent (tho not bitwise, lol)

16:19

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080707 << nifty.

16:19

Logged on 2022-02-25 11:13:09 billymg: finally getting back to working on the crawler, i've implemented geolocation (ty for the recommendation punkman) and time series data collection, for charting

16:21

asciilifeform: i thought whaack was using jfw's, though perhaps just assumed

16:26

also feel like an idiot because i discovered my motherboard has an m.2 slot for NVMe drives and i've just been using SATA this whole time. never thought to expect it since it's an old AM3+ board

16:27

discovered when looking at the specs to see what kind of PCI m.2 adapter i should buy, then saw has built-in m.2 slot

16:35

asciilifeform: ah, this one, now i remember the history

16:35

http://logs.nosuchlabs.com/log/asciilifeform/2021-06-10#1038702

16:35

Logged on 2021-06-10 19:12:46 asciilifeform: whaack: ok, baked, tested: asciilifeform_dumpblocks_force_mainchain.kv.vpatch ; asciilifeform_dumpblocks_force_mainchain.kv.vpatch.asciilifeform.sig .

16:42

cgra: i like your vpatch code viewer. i assume you rolled your own. is there a way to browse patches on your www, or only accessible if you have a direct link?

~ 26 minutes ~

17:08

billymg: make sure it's an nvme and not 'm2 sata'

17:08

they look rather similar, but incompat.

17:09

e.g. apu1/2 have 'm2 sata' slots which give 0 speed advantage (tho moar compact drive)

~ 22 minutes ~

17:31

asciilifeform: damn, glad you mentioned it. it is in fact only sata 3

17:32

hrm, actually maybe not. conflicting reports from the amazon reviews, and not specifically stated in the specs

17:33

specs say the m.2 slot is on a PCI express 2.0 x4 bus, which is quite a bit faster than sata 3

17:33

billymg: look at pic, the notch is in diff place

17:33

(m2-sata vs nvme)

17:48

was able to fit a spare nvme drive into the slot, so i guess nvme?

17:48

if fit w/out hammer, lol, then yes

17:48

nice, nb for AM3+ board then

17:49

though the PCI adapters are cheap enough, so not a huge deal

~ 29 minutes ~

18:18

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080700 This is incorrect. It could be done very easily with merely megabytes of memory and minimal computation.

18:18

Logged on 2022-02-25 04:43:30 jonsykkel: sounds like you would need 30k lines code and 1gig of cpu cache to make sense of ur compressed text in reasonable time

18:19

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080701 Unfortunately, English really would require a million or so entries, for all such variations.

18:19

Logged on 2022-02-25 04:43:34 jonsykkel: you wont be able to encode this article using your system, since u made use of words "Seperating" and "inling" which was not in dict. but i guess your spellchecking feature would catch it

18:20

Spellchecking comes for free in this system, yes, and words not in the primary dictionary are placed in an auxiliary dictionary.

18:20

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080702 I'd do it, at least for a while. It's actually very easy to create a dictionary difference, to enable different versions with minimal extra storage.

18:20

Logged on 2022-02-25 04:43:40 jonsykkel: u mention updating dictionary with new words, who decides which words have this privilege, the uniword consortium? and who will do teh work of updating all softs taht use "text" feature. if reorganize dict, how to communicate with the 80% of your fellow text users that havent updated yet

18:21

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080703 We agree software should be finished. A dictionary certainly can be finished.

18:21

Logged on 2022-02-25 04:43:45 jonsykkel: dictionary would have to be append only or smth, unless you want to live in dynamically linked text hell

18:22

Once the system knows of multiple languages, it can allow structured intermingling amongst them, and it can go even further, jonsykkel.

18:23

Consider Latin: Rather than store every word declension and conjugation, take advantage of the regularity and store a code representing this instead, for the system to take and mutate; irregular words pose no issue whatsoever.

18:25

So, the word TONITRUS is stored as its base code, along with the nominative singular code; the word TONITRUUM would be stored the same way, but with the plural genetive code.

18:26

This system makes all text operations I've considered so far more efficient, and easily made parallel.

18:26

All words use the same amount of space in the text.

18:26

Consider the problem of searching from one point to another, forwards or backwards, for the first word beginning with some specified letters. Let me know when to share my solution.

↖ ↖ ↖ ↖ ↖

18:27

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080700 Also, making sense of the text doesn't require knowing its letters, as a hint.

18:27

Logged on 2022-02-25 04:43:30 jonsykkel: sounds like you would need 30k lines code and 1gig of cpu cache to make sense of ur compressed text in reasonable time

18:35

thimbronion

What about abbreviations and neologisms?

18:35

thimbronion

acronyms

18:35

thimbronion

SPQR

18:36

There would be an abbreviation system. One disadvantage is it would require people to actually know what they mean, for those not already known.

18:41

Try to solve the problem, thimbronion.

18:52

procrustean solution.

18:57

cgra

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080721 << it's python 'pygments' for colors, spiced up with my own dog-food to add per-line links. in the url, "p" stands for "patch", and "r" for revision; most of the trb items are there, only if you give the correct url. i'm planning to add index pages, too, so can eventually also browse it.

18:57

Logged on 2022-02-25 11:41:51 billymg: cgra: i like your vpatch code viewer. i assume you rolled your own. is there a way to browse patches on your www, or only accessible if you have a direct link?

18:57

cgra

the item you sampled, was added by hand to the presented set of files, the others are bulk-generated from a vtree dir

18:57

What, asciilifeform?

18:58

cgra: asciilifeform still prefers lxr's clickable identifiers

18:58

verisimilitude: your encoding thing

18:58

Explain.

18:58

verisimilitude: yer essentially asking to uninvent the alphabet.

18:59

i.e. return to egyptian hieroglyphs, effectively

18:59

I'm not, no more than binary computers uninvent decimal.

18:59

cgra

asciilifeform: yeah, understandably.

18:59

Try to solve the problem, asciilifeform, or care to see my solution now?

19:00

verisimilitude: which problem, exactly?

19:00

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080748

19:00

Logged on 2022-02-25 13:26:38 verisimilitude: Consider the problem of searching from one point to another, forwards or backwards, for the first word beginning with some specified letters. Let me know when to share my solution.

~ 1 hours 1 minutes ~

20:01

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080734 << right, mbs of memory means either sacrifice 30% of precious cache for text decoding or constantly suffer 200cycles to decode single words

20:01

Logged on 2022-02-25 13:18:31 verisimilitude: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080700 This is incorrect. It could be done very easily with merely megabytes of memory and minimal computation.

20:02

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080741 << i dont see how, considering language evolves

20:02

Logged on 2022-02-25 13:21:17 verisimilitude: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080703 We agree software should be finished. A dictionary certainly can be finished.

20:03

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080749 << if all youre doing is searching for words, sure

20:03

Logged on 2022-02-25 13:27:35 verisimilitude: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080700 Also, making sense of the text doesn't require knowing its letters, as a hint.

20:05

if you need to substitute a non word aligned string in 3gig file enjoy cacherape

20:10

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080772 That's megabytes of memory for the entire system, and not all of it would be needed at once. Consider almost all text would be much smaller in this system.

20:10

Logged on 2022-02-25 15:01:43 jonsykkel: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080734 << right, mbs of memory means either sacrifice 30% of precious cache for text decoding or constantly suffer 200cycles to decode single words

20:11

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080748 << would guess optimal solution depends on how much text u are searching throuhg and how big "alfabet" is

20:11

Logged on 2022-02-25 13:26:38 verisimilitude: Consider the problem of searching from one point to another, forwards or backwards, for the first word beginning with some specified letters. Let me know when to share my solution.

20:11

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080774 Language devolves.

20:11

Logged on 2022-02-25 15:02:29 jonsykkel: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080741 << i dont see how, considering language evolves

20:11

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080778 This doesn't occur, due to the auxiliary dictionary.

20:11

Logged on 2022-02-25 15:05:03 jonsykkel: if you need to substitute a non word aligned string in 3gig file enjoy cacherape

20:12

http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080781 No, but I'll wait for asciilifeform to perhaps give his solution.

20:12

Logged on 2022-02-25 15:11:02 jonsykkel: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080748 << would guess optimal solution depends on how much text u are searching throuhg and how big "alfabet" is

20:13

In a specialized machine, the text could be handled by a separate and greatly-parallel computer, jonsykkel.

20:17

ill start making room for wordcard next to gfxcard

20:20

Think of something more like the Canon Cat, although I still need to read more about that.

20:21

dunno it

20:22

Elision can work with a typical machine, but opens the path to processing text in other ways.

20:29

http://logs.nosuchlabs.com/log/trilema/2015-10-31#1312905

20:29

(trilema) 2015-10-31 ascii_field: http://log.bitcoin-assets.com/?date=31-10-2015#1312876 << as symbolics lispm console was !!

20:29

http://log.bitcoin-assets.com/?date=31-10-2015#1312876

20:29

http://logs.nosuchlabs.com/log/trilema/2015-10-31#1312876

20:29

(trilema) 2015-10-31 mircea_popescu: http://log.bitcoin-assets.com/?date=31-10-2015#1312712 << the more i think abvout it the more i understand the "display has its own computer" cut of this gordian knot.

~ 1 hours 6 minutes ~

21:36

Hello, PeterL.

21:37

hi verisimilitude

21:39

PeterL: http://logs.bitdash.io/pest/2022-02-25#1003763

21:39

bitbot

(pest) 2022-02-25 billymg: PeterL: are you on latest blatta? for some reason my station keeps crashing now, starting from http://logs.bitdash.io/pest/2022-02-25#1003761

21:41

billymg: did it kill the log bot too?

21:42

are you on the latest or an older version?

21:42

PeterL: the bot is still running, presumably because not peered with you. though if you also crashed asciilifeform's station then the bot has no peers and is effectively dead

21:43

i'm still on 9983

21:43

bot is also

21:43

did you patch it to handle getdata requests?

21:43