#asciilifeform | 2020-07-03

wild stuff, even here in liberal texistan lite the fence posts are steel tube in concrete

00:01

yeah i have this annoying problem that the woman insists on shitting out valuable modifications no matter whether i rent or own

00:02

bonne soiree ascilifeform

00:02

nighty ben_vulpes

00:07

one last thing. in re http://verisimilitudes.net/2018-06-06 verisimilitude i'm not going to be bound by any fixed set of words you propose i limit myself to

~ 5 hours 26 minutes ~

05:33

feedbot

http://mvdstandard.net/2020/07/epsteins-pimp-ghislaine-maxwell-captured-by-fbi-indictment-full-text/ << The Montevideo Standard -- Epstein's Pimp Ghislaine Maxwell Captured By FBI: Indictment Full Text

~ 4 hours 15 minutes ~

09:48

shinohai

http://logs.nosuchlabs.com/log/asciilifeform/2020-07-02#1015671 <<< lel, #cucked

09:48

Logged on 2020-07-02 23:26:33 ben_vulpes: http://logs.nosuchlabs.com/log/asciilifeform/2020-06-28#1015442 << what has the scawwy mawwidge word have to do with either being boring or monogamous

~ 1 hours 10 minutes ~

10:59

!w poll

10:59

Polling 12 nodes...

10:59

205.134.172.6:8333 : (172-6.core.ai.net) Alive: (0.082s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=637409

10:59

205.134.172.26:8333 : Alive: (0.082s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=637515

10:59

205.134.172.4:8333 : (172-4.core.ai.net) Alive: (0.144s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=637515

10:59

205.134.172.27:8333 : Alive: (0.136s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=637515 (Operator: asciilifeform)

10:59

108.31.170.3:8333 : (pool-108-31-170-3.washdc.fios.verizon.net) Alive: (0.142s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=637515 (Operator: asciilifeform)

10:59

192.151.158.26:8333 : Alive: (0.146s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=637515

10:59

208.94.240.42:8333 : Alive: (0.232s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=637515

10:59

143.202.160.10:8333 : Alive: (0.286s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=637515

10:59

213.109.238.156:8333 : Alive: (0.273s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=637515

10:59

176.9.59.199:8333 : (static.199.59.9.176.clients.your-server.de) Alive: (0.426s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=189706 (Operator: jurov)

10:59

188.121.168.69:8333 : (rev-188-121-168-69.radiolan.sk) Alive: (0.379s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=637515

~ 49 minutes ~

11:48

http://logs.nosuchlabs.com/log/asciilifeform/2020-07-02#1015656 I agree, gregorynyssa. I claim that, after OpenSSL, it should be obvious to all that only fools make this suggestion, and that those who supposedly should clearly aren't qualified; it creates a perverse selection, where the least qualified do the work.

11:48

Logged on 2020-07-02 20:27:24 gregorynyssa: few programmers have had the courage to call out "never roll your own crypto" as propaganda.

11:50

http://logs.nosuchlabs.com/log/asciilifeform/2020-07-02#1015725 I agree that's poor, ben_vulpes, which is why the second system to be layered on top of that bedrock is a system for providing exactly that, an alternative dictionary specified by the user, containing words not in the primary.

11:50

Logged on 2020-07-03 00:07:14 ben_vulpes: one last thing. in re http://verisimilitudes.net/2018-06-06 verisimilitude i'm not going to be bound by any fixed set of words you propose i limit myself to

~ 31 minutes ~

12:21

oughta go study latin; "word" is not a fundamental in any sane language.

12:22

sounds like you've invented chinese here, single-symbol-per-concept

12:24

(and this is even an oversimplification of chinese; one can see the synthesis of fundamentals into higher concepts in traditional chinese)

12:29

trinque: even moar illustrative than the latins : in asia, concept of 'word' is quite loose (see e.g. turkish or japanese)

12:29

( 'agglutinative' langs being the term of art )

12:31

*

asciilifeform not aware of any earthling langs that lack ~any~ agglutination; not even conlangs. nowhere is '1 fixed word per concept' afaik.

12:39

http://logs.nosuchlabs.com/log/asciilifeform/2020-07-02#1015747 Amusingly, I've been told this before in my primary venue, trinque.

12:39

Logged on 2020-07-03 12:22:06 trinque: sounds like you've invented chinese here, single-symbol-per-concept

12:39

I've studied Latin and Chinese; I've been considering the issues moving to other languages would cause with my system, but ``will cross that bridge when I reach it''.

12:42

My system can also constitute a domain-specific compression scheme; it's likely favorable results could be achieved by having the dictionary simply contain every word used in the work and then packaging it with, but the gains grow as the dictionary becomes more comprehensive and more used. In any case, I describe this in my articles, although that oldest is due for a rewrite.

12:45

compressing text is a pretty solved problem neh?

12:47

shinohai

In other word things: https://twitter.com/TwitterEng/status/1278733305190342656

12:49

shinohai: i fully expect 'words are nazi, replace errything with emojihieroglyphs' by this time next year..

12:49

shinohai

^

~ 44 minutes ~

13:33

It could be compressed better, trinque. The fundamental of my idea is there's no good reason not historical for things to be as they're currently, so I want to start anew and pursue something different I think may be better.

13:34

It fits in well with my mindset regarding use of language, also.

13:35

If only they'd use a form of Newspeak, rather than continue to abuse English; why not base Newspeak on a nice African language, which can't express complex ideas already; most of the work is done for them, there.

~ 42 minutes ~

14:17

verisimilitude: that was a great article about textual processing. reminded me of Forth as well as Common Lisp's interning of symbols.

14:20

excellent observation at the beginning as well: support for rich text in 2020 is also alarmingly bad.

14:20

proficient Linux users are forced to choose between a bulky, implementation-defined GTK/Qt program which supports rich text,

14:21

or a general-purpose text-editor which doesn't. despite that even VT100 has rich-text escape-codes.

14:22

the situation is akin to how webpages before 2010 had trouble displaying rounded corners.

14:22

rich text was already a solved problem on the Genera, the Canon Cat, and, arguably, Windows 9x

~ 1 hours 32 minutes ~

15:54

verisimilitude: http://logs.nosuchlabs.com/log/trinque/2020-06-15#1000678 << sure, couldn't agree more with questioning precedent.

15:54

(trinque) 2020-06-15 trinque: jfw: I am disinclined to take precedent for anything here, but am satisfied that "supporting empty files" is pointless

15:55

not that I find your approach here sensible, but go forth and do weird shit!

15:56

I don't actually find fault with your building a dictionary, so much as the apparent chosen data-structure.

15:56

curious if you've seen the "whitaker's words" latin parser

~ 2 hours 16 minutes ~

18:13

trinque: I believe that was written in Ada.

18:17

re. your point above, Chinese (monosyllabic) words were never composed through a process of synthesis

18:20

according to DeFrancis, re. every character, the spoken word existed first; then some strokes were assigned to represent it.

~ 18 minutes ~

18:38

btw. I liked your postings about the Republican OS.

18:41

I'm glad that article of mine was enjoyed, gregorynyssa; I'm in the process of rewriting all of my older articles, and that one could be improved.

18:42

I'm vaguely familiar with it, trinque, and have taken a look now.

18:43

I'd appreciate an elaboration on the issues with my dictionary idea.

18:44

As with other languages doing this, there's a subset of Chinese Hanzi used solely for their sounds, perhaps akin to the Japanese katakana, by my understanding, gregorynyssa.

18:48

the situation is rather erratic, sadly. there is no canonical set of sound-only characters.

18:48

non-Chinese names are transliterated into Chinese on a case-by-case basis.

18:49

the Mainland, Hong Kong, and Taiwan maintain separate lists of official transliterations of every known name.

~ 32 minutes ~

19:22

http://logs.nosuchlabs.com/log/asciilifeform/2020-07-03#1015764 << i've a pertinent tale : 10y ago, asciilifeform tried to write a (naively, thought...) very simple proggy :

↖ ↖ ↖ ↖ ↖

19:22

Logged on 2020-07-03 14:20:01 gregorynyssa: excellent observation at the beginning as well: support for rich text in 2020 is also alarmingly bad.

19:24

... was to be very basic text editor where edits db entries, each one titled w/ arbitrary string. principle being, 1st pair of [[ ... ]] in the text is title of given entry; and any subsequent [[ ... ]] become clickable links to other entries. with tab-completion and floating selector while typed. ( plan included other knobs, but this alone already gives very useful item imho )

↖ ↖

19:25

went through 3 diff gui libs : native 'gtk' ; 'wxwidgets' ; finally 'qt' . none of'em could do it!!

19:26

i.e. could write the proggy, but turned out 100% impossible to make the editor behave as specified, with e.g. links being clickable ~while text box is editable~ (no modality liquishit pleez) , with completor/selector box, etc

↖ ↖

19:26

sawed on it for almost whole year and gave up in disgust.

19:27

apparently such proggy can only be written if write 100% of graphics stack (incl. font renderer) from 0. as in e.g. emacs (before anyone asks, i ~specifically~ did not and still do not want this kinda thing as elisp progggy -- it'll 100% guaranteed choke on 200MB+ of indexable db.. )

↖ ↖ ↖ ↖

19:30

( i wanted it specifically to recognize, while type-completing , ~any~ substring from ~any~ existing entry title )

~ 21 minutes ~

19:51

font-renderer from scratch is a project I've been interested in. should probably focus on METAFONT rather than TrueType.

19:52

on grounds that the spec. of METAFONT is simpler.

19:57

gregorynyssa: problem was that asciilifeform dun have 20y to devote to writing a font renderer.

19:58

and at same time aint about to glue in a 500kloc piece of obfuscated c legacy, into what'd otherwise be a 2kloc proggy.

19:59

( using bitmap font also quickly ruled out, asciilifeform regularly inhabits screens w/ at least 6 variant pixel densities )

20:00

afaik this is what one might call a 'os-complete' problem.

20:01

i.e. a satisfactory pill demands baking sane os in its entirety.

20:01

your story reminds me of a friend back in college who decided to make an MS Paint clone in Haskell for his term-project. eventually gave up.

20:02

not worst possib. outcome : he could've e.g. also wanted cut&paste to work in it, and ended up eating pistol..

20:02

I ran into same dilemma actually. decided to go with bitmap and avoid screens of different pixel-densities.

20:07

wrote several tools for personal use; all using the stilted X11 fonts.

20:08

*

asciilifeform avoided guism entirely ever since that episode

20:08

"also wanted cut&paste to work in it, and ended up eating pistol.." haha.. =D

20:12

How ``amusing'', as this mirrors my reflection on working on my MMC.

20:12

It's unnecessarily arduous to write a user interface under current operating systems.

20:12

asciilifeform: under Linux, TUIs are hardly better. do you just try never to leave Emacs?

20:13

I do eventually want to write my own operating system, and perhaps then have what I actually want. I must remind myself to not allow the system to gaslight me and make me forget about what I actually want.

20:13

I also considered using Emacs as the underlying system, but didn't want to cede any control, which Emacs requires.

20:14

http://logs.nosuchlabs.com/log/asciilifeform/2020-07-03#1015790 I intimately know this disgust.

20:14

Logged on 2020-07-03 19:26:12 asciilifeform: sawed on it for almost whole year and gave up in disgust.

20:16

gregorynyssa: believe or not, i only use emacs to edit text (proggies & human txt), i dun abuse it as wwwtron, or irc client etc

20:17

cuz elisp has the efficiency & reliability of a trs-80 soldered from dumpster scraps

↖ ↖

20:19

asciilifeform: I wonder why none of the numerous projects to re-write Emacs in CL were finished.

20:19

even for programs, it has detectable and infuriating lag on my systems ( i use 'flymake' ) and i use it while hating

20:19

gregorynyssa: search #t logs for clim

20:20

it's one of those quicksand traps, where no one has the 20y to do a proper feature-complete job, and each attempt produces yet more nauseating pile of shit

20:21

wow that explains a lot. have you tried Smalltalk for GUIs? does that get you further?

20:22

*

asciilifeform can't stand 'this could have been a lisp but i'm infix-retarded' abortions like smalltalk. so no, not tried

20:23

i don't expect that it's any less hobbled by x11 than the other vintage gui kits tho

20:23

my friend used Seaside Smalltalk for a web-project. once a visitor accessed a URL and the code ran into an exception,

20:23

lol!

20:24

thing has no condition handling ?!

20:24

he rewrote the method on the fly..

20:24

from the visitor's point of view, there was no error.. just a delay in loading the page.

20:24

very reminiscent of Lisp Machine.

20:25

this sounds great until becomes habit. iirc was why phf never published the orig logger, he found that it was not reducible to a linear src snapshot

20:25

instead was multi-year 'save-lisp-and-die' bin image..

20:29

I suspected there would be problems with image-based workflow. this is the first concrete example I have heard.

~ 1 hours 23 minutes ~

21:53

feedbot

http://mvdstandard.net/2020/07/dalit-sues-cisco-over-caste-discrimination-at-us-headquarters/ << The Montevideo Standard -- Dalit Sues Cisco Over Caste Discrimination At US Headquarters

~ 26 minutes ~

22:19

image-based workflow as seen in e.g. sbcl on pc, tends to nail you when you try to publish. and even the lispms did not have 100% proper realtime linkage between editable src and the running image.

22:20

( in proper implementation -- should not even be physically possible for the two to become uncoupled )

22:21

i.e. should not ~ever~ have to manually invoke compiler.

22:21

oddly enuff, of all people, terry davis got this right.

22:21

( afaik -- nobody else, to date )

22:34

*

shinohai switches to TempleOS

~ 35 minutes ~

23:10

Giving it more thought, that program should be simple enough to construct using the terminal as the interface, asciilifeform, since full control of the text qualities, such as underlining, are under control of the program, as the interface is mere facade. The mouse interfaces are terrible, but my ACUTE-TERMINAL-CONTROL has support for them. I suppose Common Lisp wouldn't be considered suitable for this program, or there's some other

23:10

issue, however.

23:11