| 01:23 |
adlai |
asciilifeform: my original idea is better expressible as, possibly, 'broadcaster pencil', or perhaps location or profile, although the latter two seem easier to misinterpret. |
| 01:24 |
adlai |
the idea was that, while target information should be left out of such an encoding, any listeners who have a reason to listen still can be expected to identify the source in which they are interested |
| |
↖ |
| 01:26 |
adlai |
this idea works equally well for sha3(ipv6-address), sha256d(blocks-mined-to-same-coinbase), pidigits-starting-at(frequency-and-location), etc |
| 01:27 |
* |
adlai was leaning towards a 'hash function' along the lines of a pi approximation solely for the purpose of having an _infinite_ code |
| 01:31 |
adlai |
s/hash function/prng/, there are quite likely already 'best' (and worst) practices about generating this kind of oracle. |
| 01:32 |
* |
adlai wonders to what degree the 'rateless' property is the interesting one |
| 01:45 |
adlai |
in case my jargon is too obtuse: "broadcaster pencil" is inspired by the term from classical optics, where 'pencil' is the word describing a circle section taken to infinite radius from a centre emitting rays of light; |
| 01:47 |
adlai |
and its equivalent, in the IP world, is arguably more specific than an IPv4 address; it's closer to the "Reply-To:" field, although even closer to a concatenation of LAN IP source, and gateway. |
| |
↖ |
| 01:48 |
adlai |
e.g., 127.0.0.1,127.0.0.255 ~= 'pencil' of a candle surrounded by circular mirror |
| 01:50 |
adlai |
the comparison to the "Reply-To:" field only emerges when considering why it's information worth treating as known before listening |
| 01:51 |
* |
adlai is not familiar enough with 21st century email relaying protocols to make a precise analogy |
| 01:56 |
adlai |
http://logs.nosuchlabs.com/log/asciilifeform/2021-03-09#1033094 << have you ever used dictation programs? they're about as good at understanding speakers, as arbitrarily chosen musicians at improvising to an undefined chord progression; i.e., "ymmv" |
| 01:56 |
snsabot |
Logged on 2021-03-09 22:11:27 verisimilitude: Dictating to a machine with speech would be better. |
| 01:59 |
adlai |
in case this somehow never got explicitly logged: I spent a non-trivial amount of man-hours tweaking programs for dimensionality reduction of recorded human speech, so although I don't have any certification from that miserable field, I actually do have a slight understanding of jargon, and much less slightly-formed opinions. |
| 01:59 |
* |
adlai did get paid, though, so at least something! |
| |
~ 19 minutes ~ |
| 02:19 |
adlai |
fwiw, the more I mull over the parallels between the various strengths and shortcomings of the different input methods, and those of prototypes that have had less commercialised progeny, the less I find the name "Engelbart's Violin" appropriate. |
| |
↖ |
| 02:19 |
* |
adlai wonders whether asciilifeform's hate of strings extends outside post-quantum physics! |
| 02:20 |
adlai |
e.g., is the 5-black-7-white piano keyboard the worst of all possible worlds, due to being both a keyboard, and based on strings |
| 02:31 |
verisimilitude |
I've not, adlai. |
| 02:33 |
adlai |
another input method you may have not encountered yet is 'dasher', although I believe that only exists as a smartphone keyboard that takes an interactive approach to autocorrect, instead of "let the user mash the keys, and then we guess". |
| |
↖ |
| 02:34 |
verisimilitude |
Still, I can see myself speaking to the machine and using the chorder to correct errors. |
| 02:35 |
adlai |
in 'dasher', the user uses one axis of the touchscreen to navigate a tree of accumulated letters, while the orthogonal axis allows for acceleration in the case of an unambiguous [i.e., suffix-free]word, or deceleration in the case of mistakes. |
| 02:35 |
* |
adlai used dasher for a while, when living in a "Google Play" device |
| 02:35 |
verisimilitude |
That seems interesting. |
| 02:36 |
adlai |
... and could not find it in any of the other app stores, after deciding to stop using that store. |
| 02:38 |
adlai |
it's a model that depends fundamentally on synchronous input, which is a problem for use in non-ideal conditions; e.g., most touchscreens that I've used are actually still usable when filthy (or worse: damp!), although the frequency of repeated false positives, when they do occur, increases. |
| 02:40 |
adlai |
this reminds me of one specific bit of hardware that the IDF used quite ubiquitously, when I was there; one day someone discovered that this was identical to a certain consumer-grade laptop, and simply booted into the military program, on top of XP; |
| 02:40 |
verisimilitude |
I'm glad to not use phones. |
| 02:40 |
adlai |
everyone who actually had to use the damn things for their tasks, immediately began singing the praises of this laptop, and how they wanted to buy one for their personal use after they are discharged. |
| 02:41 |
verisimilitude |
How amusing. |
| 02:41 |
* |
adlai somehow never saw one of these, ever again! |
| 02:42 |
adlai |
they probably would have been affordable at the time, due to the fact that they didn't need a lot of computing power; the reasons for them to be expensive were only the durability of construction |
| 02:43 |
adlai |
the things got dropped, tossed, etc, a few times each day; weighed as much as a modern "gaming laptop", and were ~smaller~ than the "executive laptops"; would probably be branded as "jungle notebook" or something similar. |
| |
↖ |
| 02:43 |
adlai |
er, 'netbook' |
| 02:45 |
adlai |
they even had a standard builtin wireless card, and those docks where you could add external wireless cards. |
| 02:48 |
* |
adlai wonders whether he'll hit this same wall that verisimilitude has, after a few more years of using standard keyboards at burst-rpm |
| 02:49 |
adlai |
my meagre coping, for nearly a decade already, has been to occasionally use it one-handed; although I don't switch to one-handed layouts. |
| 02:49 |
verisimilitude |
I write a great deal, and mine effort is preventative in nature. |
| 02:51 |
verisimilitude |
It's nicely supplemented with a mouse. |
| 02:55 |
verisimilitude |
Unfortunately, I realized no keyboard can replace the ideal of having another do the work. |
| 03:03 |
adlai |
regarding the alternate base glyphs, the grid's decomposition into line segments is not obvious to someone who's unfamiliar with that kind of mobile phone lock pattern |
| 03:03 |
snsabot |
Logged on 2021-03-10 13:59:18 verisimilitude: That's the one, adlai; this has some of my designs for numerical glyphs. |
| 03:05 |
* |
adlai for example is biased towards ridiculously 'finer' [in the topological sense] decompositions, due to seeing all intersections as endpoints when considering the first pattern without thinking of it as a 3x3 grid |
| 03:06 |
verisimilitude |
It's even less apparent when made circular. |
| 03:14 |
adlai |
whenever I dabble in hieroglyphics, I set myself the arguably-unreasonable goal of the various glyphs being maximally unambiguous |
| |
↖ |
| 03:14 |
verisimilitude |
Yes. |
| 03:15 |
adlai |
i.e., I want a scribbled set of digits to survive both OCR, and chromatography! |
| 03:16 |
adlai |
the use-case being, scanning pages written while drunk, that may have absorbed significant amounts of the drunk drinks. |
| 03:16 |
verisimilitude |
Mine will be unique, despite any rotations. |
| 03:16 |
adlai |
reflections? |
| 03:17 |
verisimilitude |
Yes. |
| 03:18 |
verisimilitude |
Those as well will be unique. |
| 03:18 |
adlai |
the absence of inversion symmetry in the space group that you get by passing notes around doesn't mean that writing can't soak into adjacent pages, or that you wouldn't want to scan both side of a translucent paper simultaneously. |
| 03:18 |
adlai |
so there are tangible benefits to having rotations and reflections both unambiguous. |
| 03:18 |
verisimilitude |
Yes. |
| 03:20 |
adlai |
slightly upstack: http://logs.nosuchlabs.com/log/asciilifeform/2021-03-09#1033057 << any specific reason for preferring this over Dvorak? |
| 03:20 |
snsabot |
Logged on 2021-03-09 16:43:13 trinque: incidentally just switched this keyboard to colemak, seems like it will take some load off left hand |
| 03:21 |
* |
adlai does not recall what layout trinque was previously using; he may have provided the reason right there |
| 03:23 |
adlai |
from skimming search results, it seems that the fundamental difference is that Colemak was designed with an additional constraint to retain parts of QWERTY |
| |
~ 23 minutes ~ |
| 03:47 |
verisimilitude |
A nicety of chorders is no more layout nonsense. |
| 03:50 |
adlai |
nonsense, ergonomics still exists in the shape of the device, and roles and locations of keys |
| 03:51 |
adlai |
e.g., "I like the microwriter, except half of my head was screwed on backwards shortly after gastrulation, and I want the roles of the four finger buttons barrel-rotated one finger thumbwards" |
| 03:53 |
verisimilitude |
I merely mean the layouts are more-or-less ``take-it-or-leave-it''. |
| 03:55 |
verisimilitude |
There's less need and demand for alternations. |
| 04:01 |
verisimilitude |
I agree with asciilifeform; individual key adjustment would be nice. |
| |
↖ |
| 04:05 |
* |
adlai wonders to what degree the approach of alf's mwemu.js can be copied in an OS driver for a regular keyboard |
| |
↖ |
| 04:06 |
adlai |
before y'all scoff: 'to what degree' is a monodromy-aware question; i.e., certain systems can detect key state at higher granularity than pressed / released |
| 04:07 |
adlai |
and while we're on mathematics, I owe trinque an apology |
| 04:07 |
verisimilitude |
It would be bad. |
| 04:07 |
adlai |
his intuition about integral domains was correct, and mine was wrong. |
| 04:08 |
adlai |
re: this thread |
| 04:08 |
snsabot |
Logged on 2021-02-04 11:24:38 trinque: right, "I can do arithmetic in this rule system" eh? |
| 04:09 |
* |
adlai often oversimplifies |
| 04:09 |
snsabot |
Logged on 2021-02-04 11:25:33 adlai: you can do arithmetic, although certain solution methods that are taught in grade school, become impossible, because you can't progress past a zero product in a deterministic way. |
| 04:09 |
adlai |
although my oversimplification was not incorrect, trinque's intuition was entirely correct. |
| 04:10 |
adlai |
specifically, there should only be one zero divisor in an integral domain, and that is the zero element; exactly as in regular arithmetic. |
| 04:16 |
* |
adlai recently read himself to sleep with the topology textbook, and ended up revisiting the old "so what is the zeroth power of zero?" argument |
| 04:17 |
adlai |
it is a remarkably dull one, once you've learned enough; dull enough to make you realise that you're arguing with yourself, in a dream, and wake up and go do something less dull. |
| 04:19 |
adlai |
tl;dr: 'power' is an ambiguous abuse of notation; disambiguate by explicitly stating whether you are exponentiating, by repeated multiplication, or tetrating, by repeated repetition of lesser operations. |
| 04:20 |
adlai |
amusingly enough, the topology textbook basically assumes you've finished studying group theory; there's a parallel thread of "topological group" problems, although they are separated so you can read the text while skipping all these |
| 04:22 |
* |
adlai wonders where this thread goes; specifically, anywhere other than 'elision' ? |
| 04:22 |
snsabot |
Logged on 2021-03-04 16:18:50 verisimilitude: Say, does anyone know offhand how many words Latin has, sans declensions and conjugations? |
| 04:24 |
verisimilitude |
No. |
| |
~ 14 hours 49 minutes ~ |
| 19:14 |
Apocalyptic |
adlai: are you reading Munkres ? |
| |
~ 1 hours 29 minutes ~ |
| 20:43 |
feedbot |
http://mvdstandard.net/2021/03/on-this-day-in-2020/ << The Montevideo Standard -- On This Day In 2020 |
| |
~ 2 hours 39 minutes ~ |
| 23:22 |
adlai |
Apocalyptic: 2nd edition |
| 23:23 |
adlai |
as always, garbage copy from phree shelf at engineering library |
| 23:23 |
adlai |
now, 'garbage' is perhaps uncharitable, same as "toilet lappy" might be downright insulting to some who'd kill to get off the touchscreen continent |
| 23:23 |
adlai |
however, my copy of Munkres has several "many others like it, but this one is mine" that are quite hilarious |
| 23:24 |
adlai |
e.g., leaves with pages printed in reverse order |
| 23:24 |
adlai |
it goes well with the frequent "can be skipped without breaking continuity" that the bastard slips in all over the place |
| 23:26 |
* |
adlai occasionally wonders whether the editors of mathematical textbook publishers actually go through the book as though they were students, instead of merely weighing it, estimating its relative beauty on various possible shelves, and farming out 'peer review' to their rolodex |
| 23:28 |
adlai |
so the charitable description is 'lovingly restored', rather than 'used; poor as compost' |