02:25 |
crtdaydreams |
verisimilitude: wouldn't you just want to have a low level impl. of s-exp wherin you remove the abstraction between symbols and strings? |
02:26 |
crtdaydreams |
i.e. 'mystring == "mystring" |
02:27 |
crtdaydreams |
well, that's a poor example, rather '('my 'string) == "my string" or something along those lines |
02:27 |
crtdaydreams |
In that sense if it were to run on a lisp machine the implementation would in fact be the lowest level. |
02:28 |
verisimilitude |
Lisp isn't the ultimate abstraction, crtdaydreams. |
02:28 |
verisimilitude |
However, yes, Elision does resemble symbol interning in some ways. |
02:28 |
verisimilitude |
I'm going much further, though. |
02:29 |
verisimilitude |
I'll wait a little bit longer before revealing my solution to that problem, asciilifeform. |
02:30 |
crtdaydreams |
By no means to I assume that, also I was thinking it's more akin to a hardware impl. of a hash map. |
02:31 |
crtdaydreams |
But symbol interning makes more sense |
02:33 |
verisimilitude |
A hash table has O(N) worst time performance, so no. |
02:33 |
crtdaydreams |
Ofc then you have to deal with conflicting standards like all the different forms of UTF and custom fonts, dictionary extenstions and god-knows what else. I think a fundamentally extensible computer (i.e. a lisp machine with FPGA programmable in macros) would be requisite. |
02:33 |
verisimilitude |
Elision is based around O(1) time operations. |
02:34 |
crtdaydreams |
Ah ok. |
02:34 |
verisimilitude |
Unicode is irrelevant to me. |
02:34 |
verisimilitude |
Furthermore, one goal is to actually define text, and not treat it as a grab-bag of whatever shit the Unicode Consortium adds to it. |
02:35 |
verisimilitude |
A later goal is to attack the notion of a fixed character set for all text. |
02:35 |
verisimilitude |
Why shouldn't I be able to add this symbol to our conversation: http://verisimilitudes.net/crest.png |
02:36 |
verisimilitude |
Now, it would be one thing if the symbol masters were good, but they're not; observe the vile symbols added for homosexuals and associated illnesses. |
02:37 |
crtdaydreams |
i would not wish to taint my understand of a "symbol" with that of an emoji charset |
02:37 |
shinohai |
But verisimilitude we got to be inclusive to all those pregnant men out there. |
02:38 |
verisimilitude |
Why, we'd precisely the same vile symbol in-mind, shinohai. |
02:38 |
shinohai |
kek |
02:38 |
crtdaydreams |
not for long, they'll get an abortion at 24 weeks. |
02:39 |
shinohai |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-24#1080678 <<< so far imma stick with the Latinas and not imprt girls from the Alabama of Eastern Europe. |
02:39 |
dulapbot |
Logged on 2022-02-24 22:52:12 vex: any thing interesting from .ua gals who `went to cornell' yet shinohai? ie camwhores |
02:39 |
verisimilitude |
These ``emoji'' aren't text, crtdaydreams. Text can be written well by human hands, I'd argue it's monochromatic at the symbol-level, and it can be translated to systems such as braille. |
02:40 |
verisimilitude |
A blind man should at least be able to feel a textual symbol, but these ``emoji'' fail there as well. |
02:41 |
crtdaydreams |
hm ok, I get where you're at. I need to broaden my understanding of the current text processing faculties if I were to contribute. |
02:42 |
crtdaydreams |
i.e. how regex works and similar string manipulation tools. also what exactly constitutes as a string in a low level impl. |
02:45 |
verisimilitude |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-25#1080748 Try to solve this well, crtdaydreams. |
02:45 |
dulapbot |
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. |
02:45 |
verisimilitude |
This is an Elision problem, note. |
02:53 |
crtdaydreams |
hmm |
02:57 |
crtdaydreams |
I haven't done the maths for it, but my `solution' would be to have for a textual structure (be it a character, a word, a phrase etc.) that the mem location of each of these structures is stored in a sort of pointer-box. How these structures are identified is another problem. But To find that, the inputs would be the symbol for say a sentence with a search for the characters and there is a marker for what type it is, say a |
02:57 |
crtdaydreams |
3-bit marker before each structure. Each structure down to a character is a series of pointers to the lower structures... eh. i'm dumb |
02:58 |
verisimilitude |
My solution is so simple, but I'm waiting to see if asciilifeform wants to play. |
03:00 |
crtdaydreams |
can you pm me the solution? |
03:04 |
verisimilitude |
No; wait for it. |
03:05 |
verisimilitude |
I'll help, crtdaydreams. So, all words are represented as indices, in a holding structure such as an array, and it's needed to find the first from some point that begins with certain characters. Think about it. |
03:11 |
crtdaydreams |
whoa wait a minute |
03:12 |
crtdaydreams |
hmm |
03:14 |
crtdaydreams |
wait you're telling me that's not how they do it already? |
03:15 |
crtdaydreams |
so like, consider an array; sentence1 = array[word1, word2, word3] |
03:15 |
verisimilitude |
What passes for text is commonly stored character-by-character, crtdaydreams. |
03:15 |
crtdaydreams |
well yes. |
03:16 |
crtdaydreams |
Of course, but wouldn't you essentially end up with a string of structures. |
03:16 |
crtdaydreams |
s/string/tree/ |
03:16 |
verisimilitude |
There's some manner of misunderstanding here which I don't care to correct. |
03:16 |
verisimilitude |
I'll just reveal mine answer now. |
| |
↖ |
03:16 |
crtdaydreams |
ok. |
03:16 |
verisimilitude |
Get the index of the first possible such word, and of the last possible such word; now the searching is a bounds check on each word, needing no more dictionary access. Searching the auxiliary dictionary would mean just an additional bounds check. Isn't this neat? Certainly, turning a letter problem into an integer problem satisfies this. This is power, turning figurin |
| |
↖ |
03:16 |
dulapbot |
(trilema) 2015-10-31 asciilifeform: one that fits will into the architecture of the machine. |
03:20 |
verisimilitude |
This works because the dictionary is sorted, and even if it weren't, there could be a transformation table to make it so; thus, other problems can have such tables and use such simple methods. |
03:21 |
crtdaydreams |
I.. don't get it. Sorry. I'm dumb. |
03:22 |
crtdaydreams |
waot |
03:25 |
crtdaydreams |
keyboard cat. sorry. |
03:25 |
verisimilitude |
Read this. |
03:33 |
crtdaydreams |
what if you want to process random strings like say an RSA pub key? |
03:34 |
crtdaydreams |
would you solely be only able to transport such formats in binary? |
03:35 |
verisimilitude |
Be an RSA public key text or an RSA public key? |
03:37 |
verisimilitude |
Of course if people insist on using these unnecessary qualities of character-by-character ``text'' then they should store it character-by-character, crtdaydreams. |
03:37 |
verisimilitude |
Is most text ASCII line-noise or a sequence of words with punctuation? |
03:38 |
crtdaydreams |
there would be no distinction, no? |
03:39 |
verisimilitude |
What? |
03:39 |
crtdaydreams |
Between ASCII line-noise or a sequence of words with no punctuation |
03:40 |
crtdaydreams |
misread, sorry. but the point remains as above ^ |
03:40 |
verisimilitude |
Elision is designed to store text as a sequence of words, perhaps with punctuation. |
03:40 |
verisimilitude |
ASCII is designed to store text as a sequence of characters, perhaps with meaning. |
03:41 |
crtdaydreams |
soisthisawordorasequenceofowordsorasequenceofcharachters? |
03:41 |
verisimilitude |
That's unnecessary is what it is. |
03:42 |
crtdaydreams |
of course, but it poses issues when dealing with things like say urls. |
03:42 |
verisimilitude |
URLs are vile shit. |
03:42 |
verisimilitude |
When interacting with such things in Elision, a trap door is needed, certainly. |
03:42 |
verisimilitude |
Most text wouldn't need it. |
03:43 |
verisimilitude |
Importantly, this trap door would only exist at a message-formatting layer. |
03:44 |
verisimilitude |
Now sure, crtdaydreams, I can't easily express line-noise in Elision, but ASCII and Unicode can't express concepts to which they haven't explicitly given a symbol; which one is the real issue for humanity? |
03:45 |
crtdaydreams |
why must you need symbols as apart of a character standard? you could have a seperate symbol standard specifically for that purpose? |
03:46 |
verisimilitude |
Why does addition deserve + but some new fundamental operation only gets a name? This should be set in stone forever? |
03:47 |
verisimilitude |
We should have the ability to add symbols ourselves, crtdaydreams. |
03:47 |
verisimilitude |
Anything less is unacceptable to me. |
03:47 |
crtdaydreams |
of course, so why not remove those symbols from a charachter standard and put them in a malleable symbol standard |
03:47 |
crtdaydreams |
s/charachter/character/ |
03:47 |
crtdaydreams |
I'm not myself today. Eugh. |
03:48 |
verisimilitude |
That's never going to happen with Unicode. |
03:48 |
crtdaydreams |
Who said anything about using unicode's shitty standard. |
03:48 |
crtdaydreams |
s/'s/s/ I can't even english today. Great apologies. |
03:49 |
verisimilitude |
No, that's right. |
03:49 |
crtdaydreams |
is it? x_x |
03:49 |
verisimilitude |
I've a secret, crtdaydreams; I think before I hit enter. |
03:50 |
crtdaydreams |
Yes, I'm in a rather unthoughtful state of mind. |
03:51 |
verisimilitude |
Unfortunately, this is the current state-of-the-art regarding Elision. |
03:53 |
crtdaydreams |
This is interesting though, because it poses a problem with symbolic languages. For example Mandarin composed almost entirely of single-symbol words could be represented fluently with the ASCII character-strings. |
03:53 |
crtdaydreams |
Perhaps the problem lies not within the digital representation of language, but rather our mental representation? |
03:54 |
verisimilitude |
The digital representation makes a bounded problem seem infinite, so no. |
03:55 |
verisimilitude |
Elision recognizes that we use fewer than, say, sixteen million words in one language. |
03:55 |
verisimilitude |
We can bind the problem, and then certain issues are conquered immediately. |
03:55 |
crtdaydreams |
Of course this has it's quirks, there is no room for evolution (nor devolution) in a purely symbolic language, so you'll never have to change the standard as no new words will be added or removed. Of course this also means your thinking is now heavily limited by the number of words you have in your language. |
03:56 |
verisimilitude |
The auxiliary dictionary exists for this. |
03:57 |
verisimilitude |
However, is a document with more made-up words than standard words really written in that language? I think not so. |
03:57 |
crtdaydreams |
In a language like Mandarin would you have any made-up words as they are not built from letters but symbols? |
04:01 |
crtdaydreams |
The representation of text should not impede the users way of conveying their thoughts to a computer. This means if a user should want to type askjdnfajsd, why should they not be permitted to do so? |
04:03 |
verisimilitude |
This is already an issue with Chinese, crtdaydreams, because of the inflexibility. |
04:03 |
crtdaydreams |
Of course your system would permit it, but in time would it not be just another layer of added complexity? |
04:04 |
verisimilitude |
If someone wants to write on the wall with shit, he may, but I'll use ink and paper. |
04:04 |
verisimilitude |
I want tools that help how I write, crtdaydreams. |
04:04 |
crtdaydreams |
I know. The limitations imposed on thought by a language like Mandarin are not ignorable no. |
04:05 |
verisimilitude |
It exists for me before anyone else. |
04:05 |
verisimilitude |
I don't like how idiots warp English now. |
04:06 |
verisimilitude |
First we get a hashtag ``MeToo'' and then it ``becomes'' a verb and people get ``MeToo'd'' and it's sickening. |
04:06 |
crtdaydreams |
Would have said the same for Shakepeare in the 16th century? |
04:07 |
crtdaydreams |
s/Would/Would you/ |
04:08 |
crtdaydreams |
Don't worry I understand your frustration with the devolution of language, but constricting the evolution as a side-effect isn't worth it. |
04:08 |
crtdaydreams |
Besides, those who are devolving are simply a loud minority. |
04:09 |
verisimilitude |
How does Elision constrict it? |
04:09 |
verisimilitude |
It's been designed not to do this. |
04:09 |
verisimilitude |
The ability to make words is present. |
04:10 |
verisimilitude |
It adds structure. |
04:10 |
crtdaydreams |
I understand, but the need to update those words into an auxiliary dictionary and so-on and so forth in the name of optimization seems like over-engineering. |
04:11 |
verisimilitude |
Rather, it recognizes structure. |
04:11 |
verisimilitude |
No, Unicode is over-engineering. |
04:11 |
crtdaydreams |
^ I don't disagree. |
04:12 |
verisimilitude |
It's similar to how I don't like programming by entering text into a file. |
04:12 |
verisimilitude |
How do I write machine code, crtdaydreams? |
04:13 |
crtdaydreams |
I don't know, how do you? |
04:13 |
verisimilitude |
This is how I write machine code. |
04:14 |
verisimilitude |
I use this. |
04:14 |
verisimilitude |
Those who think an assembler necessary will never create such things. |
04:15 |
crtdaydreams |
ah the MMC one of the first articles I viewed on your site |
04:16 |
crtdaydreams |
I'd almost forgotten about it, It's been a long while since I read it. |
04:17 |
crtdaydreams |
Perhaps you're wasting your time talking to me then. It's evident you're far beyond a skill level I could hope to achieve. |
04:19 |
verisimilitude |
Make a website and make something of which to be proud, crtdaydreams. |
04:20 |
crtdaydreams |
Router firmware is being a cunt, I physically cannot port forward to my domain. |
04:21 |
verisimilitude |
Rent a server, as I do. |
04:22 |
crtdaydreams |
I would, besides I have a chronic fear of getting pwned. |
04:23 |
verisimilitude |
My first article was a masturbation joke; just write something decent, crtdaydreams. |
04:24 |
crtdaydreams |
I have naught the way with words. I'd just be making a bigger fool of myself. |
04:26 |
crtdaydreams |
Enough internet for one day. |
| |
~ 8 hours 50 minutes ~ |
13:16 |
* |
d4 is amused fun looking at cryptonomicon |
13:16 |
d4 |
s/fun// |
| |
~ 1 hours 25 minutes ~ |
14:41 |
mats |
see, buying the invasion would have been a solid play |
| |
↖ |
14:44 |
shinohai |
$vwap |
14:44 |
busybot |
The 24-Hour VWAP for BTC is $ 39142.77 USD |
| |
~ 50 minutes ~ |
15:34 |
asciilifeform |
... cement testbed at 472+k. |
| |
~ 48 minutes ~ |
16:23 |
shinohai |
thimbronion: 9982 pressed fine, am fiddling with knobs and will try to connect shortly. |
| |
~ 17 minutes ~ |
16:41 |
billymg |
cgra: very cool. i think something like that would be useful if built-in to mp-wp (not for you perhaps, since you've already rolled your own, but for others in the future). something that one could point to a directory containing patches organized by vtree and have it auto-generate more or less what you have on your www |
| |
↖ |
16:42 |
billymg |
cgra: btw, how do you feel about the less-enormous pile of '???' after using it for a few months? |
| |
↖ |
16:44 |
billymg |
part of me still would like to make something out of it, but it's hard to motivate myself to work on something no one really cares about |
16:45 |
billymg |
seems that most who use mp-wp, myself included even, do so because in the mp era it was declared the standard and because it works |
16:47 |
billymg |
looking at the php internals is so off putting i can see why no one *likes* using it |
| |
~ 1 hours 37 minutes ~ |
18:24 |
verisimilitude |
Besides comments, why use it at all, billymg? |
18:38 |
billymg |
verisimilitude: because i don't know what else to use instead, any suggestions? (yes, comment system would be a requirement) |
18:41 |
verisimilitude |
I just write everything by hand, and don't yet have an automatic comments system, so no. |
18:42 |
verisimilitude |
Someone less allergic to other software could probably just use a comments system from elsewhere for that, though. |
18:42 |
cgra |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1080979 << can't say i'm too familiar with the php stack so thought i'd use that name :) but overall, i fell back to the tinymce approach (knitted together a localhost "wysiwyg" editor hack), because couldn't stand hand-cranking that many trb links and code tags as i did. another thing is not sure yet why couldn't i access the admin page through ssh-forwarded port, intending to give it |
18:42 |
dulapbot |
Logged on 2022-02-26 11:42:34 billymg: cgra: btw, how do you feel about the less-enormous pile of '???' after using it for a few months? |
18:42 |
cgra |
an another look sometime |
18:43 |
verisimilitude |
This would be easy for a website such as mine, in which the comments are on other pages. |
18:58 |
cgra |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1080978 << as these are static pages, they don't need a server-side generator. perhaps only something for highlight-linking, say, multi-line highlights at minimum. |
18:58 |
dulapbot |
Logged on 2022-02-26 11:40:51 billymg: cgra: very cool. i think something like that would be useful if built-in to mp-wp (not for you perhaps, since you've already rolled your own, but for others in the future). something that one could point to a directory containing patches organized by vtree and have it auto-generate more or less what you have on your www |
18:58 |
billymg |
verisimilitude: i like the sidebar that updates based on content as well, and that archive and category pages are generated automatically, and that it's all searchable |
19:01 |
billymg |
cgra: i hate using wysiwyg editors because they tend to spit out very ugly html with inline styles everywhere, but i have come to appreciate the select text -> ctrl-k scheme for generating links used in some editors |
| |
↖ |
19:01 |
verisimilitude |
Well, that would require much more work, yes. |
19:01 |
cgra |
billymg: oh, and one minor detail: personally i would find yearly archive grouping more suitable myself than monthly. and anyone else with 0-2 articles/month :) |
19:02 |
verisimilitude |
Internet speeds are so great, I put everything on one page, not that it's large anyway. |
19:02 |
billymg |
cgra: i agree it's a pain hand-editing though |
19:02 |
billymg |
cgra: heh, good point |
19:03 |
billymg |
how would you feel about a box that ate markdown instead of html? |
| |
↖ |
19:04 |
billymg |
that could at least be configured to spit out clean html, which you then style manually with css, and doesn't require a JS textbox |
19:06 |
cgra |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1080994 << i had the same feeling from the earlier usage, but see for example my latest, look for 'storycontent', and it's pretty alright wysiwyg output |
19:06 |
dulapbot |
Logged on 2022-02-26 14:01:02 billymg: cgra: i hate using wysiwyg editors because they tend to spit out very ugly html with inline styles everywhere, but i have come to appreciate the select text -> ctrl-k scheme for generating links used in some editors |
19:07 |
billymg |
cgra: ah, yeah, very clean |
19:11 |
cgra |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1081000 << atm the hack i have, works reasonably well, so it's now sunk somewhat low in priority to think of a better replacement |
19:11 |
dulapbot |
Logged on 2022-02-26 14:03:23 billymg: how would you feel about a box that ate markdown instead of html? |
19:13 |
cgra |
if i had to guess though, long link texts might still throw your concentration annoyingly off from actual writing |
19:14 |
cgra |
ah, but i'm already adding links as a last, separate step, so may not apply |
19:15 |
billymg |
cgra: re: the code viewer, i was thinking of server side vpressing, i think like what whaack describes here but only for what the blog owner puts there |
19:15 |
bitbot |
Logged on 2022-02-19 15:53:16 whaack: http://btcbase.org/patches <-- phf are you around? this is a great tool. i think that updating this site shoul be made automatic. there should be a way to submit a vpatch , and if it has a sig from the host's l1, then the vpatch is automatically submitted and the version graph is automatically updated |
19:17 |
billymg |
cgra: i also do that, i wrap the words i want to link in [] and come back to it at the end |
19:21 |
cgra |
i'm not sure if vpatch volume is high enough that it needed server-side automation, while phf's chart certainly is neat |
19:22 |
billymg |
also a fair point |
19:23 |
cgra |
ah and if blog owner puts it there anyway, not his l1, what's the difference if the html generator runs on his desktop? |
19:23 |
billymg |
yes, true |
19:29 |
billymg |
you're right, it could be a separate set of tools that one runs locally, then uploads to some directory on their server. doesn't need to be stored anywhere in the mp-wp db |
19:33 |
cgra |
billymg: gives you more choice in picking your favourite sewing needle. and even if pages are stored in the db, prolly still applies. if wanted multi-line code highlight, then you'd likely best off with a minimal mpwp tweak on top of that |
19:40 |
billymg |
i think i'm trying to solve problems for users that i imagine exist but don't actually. as in, turn v into something like github so that "anyone can use it". but who is this "anyone" that would select v over git in the first place, and what would they do with it? |
19:40 |
bitbot |
(trilema) 2016-01-18 ascii_butugychag: btw i hope everybody understands that life with 'v' is always going to resemble dark age blood sports like cvs, etc. far more than modern greased poles (e.g., 'git') |
19:41 |
verisimilitude |
I'll eventually choose V over git, but I wouldn't use vhub. |
19:44 |
billymg |
verisimilitude: exactly |
| |
~ 1 hours 24 minutes ~ |
21:08 |
signpost |
what's needed is a decentralized mechanism where I can say "give me the vpatch which yields the antecedent I need for $hash" |
21:08 |
signpost |
not a centralized item. |
21:09 |
signpost |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1080973 << every major market collapse has bull traps the whole way down. are you telling me you know how high they'll bounce? |
21:09 |
dulapbot |
Logged on 2022-02-26 09:41:32 mats: see, buying the invasion would have been a solid play |
21:10 |
signpost |
swing trade, sure, but you're speaking in finbro memethought |
21:11 |
verisimilitude |
That almost seems like Bittorrent, signpost. |
21:11 |
signpost |
couple of things wrong with bittorrent, but it's not a terrible analogy. |
21:12 |
signpost |
relies on centralized "bootstrap" nodes even for "trackerless" torrents, for example. |
21:13 |
signpost |
all peers being equal is another problem. |
21:14 |
verisimilitude |
If I were to remark on how asciilifeform has educated me in three ways, I'd note random number generation, flaws with the Internet, and a good attitude to have on computing in general. Be there truly no good way to send a large packet without fragmentation through the Internet, asciilifeform; it must require splitting it into chunks somehow? |
21:15 |
verisimilitude |
Doesn't the DHT mechanism only require knowing of any node, similarly to DNS, signpost? |
21:15 |
signpost |
there is a "don't fragment" flag in the IP header, but admittedly I am not an expert on how it's handled. |
| |
↖ |
21:16 |
signpost |
naively, I'd expect that DF header set, plus MTU discovery by increasing packet size, plus a signal back from the far end, would tell you how large you can go |
| |
↖ |
21:16 |
signpost |
perhaps this DF field is ignored somewhere along the way |
21:17 |
signpost |
verisimilitude: I believe the clients carry along their own bootstrap nodes hard-coded, though perhaps some of them allow configuration |
21:17 |
verisimilitude |
That flag can result in an ICMP error response, signpost. |
21:17 |
signpost |
right, that tells you the maximum transmission unit, when you get an error, right? |
21:17 |
verisimilitude |
I don't recall. |
21:18 |
verisimilitude |
The reason Pest uses its size is because that's the minimum transmission unit. |
21:18 |
verisimilitude |
Still, that's too small for many interesting protocols. |
21:18 |
signpost |
yeah, it puts a limit on the rateless code for example |
| |
↖ |
21:19 |
signpost |
gets much slower with smaller block sizes. |
21:19 |
signpost |
though there is likely lots of room for optimization in my implementation, and even without it was doing a few megabytes per second on one core |
21:21 |
verisimilitude |
For something in the future, I'd prefer to be able to use packets at least a few kibibytes in size. |
21:22 |
verisimilitude |
IP fragmentation was easy to ignore in the IETF RFC, but it's clearly broken after asciilifeform explained the issues. |
21:22 |
signpost |
I provided a knob in mine. |
21:22 |
signpost |
yeah curious to hear from asciilifeform why one can't use MTU discovery with don't-frag |
| |
↖ |
21:22 |
* |
signpost will be entirely unsurprised if there's a good reason |
21:24 |
verisimilitude |
I'd recognized by myself why the top secret security clearance IP header feature was stupid, but not fragmentation. |
21:26 |
verisimilitude |
Another thing about the Internet is how it's really not designed for anything but messages. It can't really provide a stream like the telephone network could, because the telephone network effectively established a physical line. A packet-switching network could do this, by reserving buffers and establishing a chain of reserved buffers, but this would need some way to restrict the reservation to avoid issues, such as identificati |
| |
↖ |
21:26 |
signpost |
reasoning through it, one reason higher MTU might be dangerous is allowing the enemy to cause you to rediscover MTU, say by changing routes on you, or whatever else. |
21:28 |
signpost |
I could still see reason to opportunistically blast bits at a peer as fast as one can in certain cases. |
21:29 |
signpost |
not with pest packets, no need, but for things which can degrade with less harm. |
21:30 |
signpost |
mats: what do you make of the increased s&p put volume since last year, for example? |
21:31 |
signpost |
"hedges only cost money bro!" |
21:31 |
signpost |
and I'm not beating you up here, just having the thread. |
21:31 |
signpost |
one place I've been looking is commodities. they appear to have done pretty well after the other recent large crashes. |
21:32 |
signpost |
I know next to nothing about investing in commidities, so I'm just saying I'm reading about it, not giving advice. |
21:33 |
verisimilitude |
Why not learn how to count cards and visit the casinos instead, signpost? |
21:35 |
signpost |
because casinos are for taxing the poor, and capital markets are for taxing the rich, of course. |
| |
↖ |
21:35 |
* |
signpost chuckles |
| |
~ 21 minutes ~ |
21:56 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1081033 << 'don't frag' simply means 'drop rather than frag' ftr. |
21:56 |
dulapbot |
Logged on 2022-02-26 16:15:33 signpost: there is a "don't fragment" flag in the IP header, but admittedly I am not an expert on how it's handled. |
21:57 |
asciilifeform |
i.e. will get dropped if somewhere along transit sumbody would've have normally fragged it |
21:58 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1081034 << this'd work if could assume that it'll always traverse same route. which often enuff aint the case (even if neither yer own station nor the other side's is on a moving train etc) |
21:58 |
dulapbot |
Logged on 2022-02-26 16:15:58 signpost: naively, I'd expect that DF header set, plus MTU discovery by increasing packet size, plus a signal back from the far end, would tell you how large you can go |
21:59 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1081042 << well if yer using luby, turbo, etc. can set it up so some % of the packets are 'fat', but will still converge if these dun make it to other side, cuz redundancy |
21:59 |
dulapbot |
Logged on 2022-02-26 16:18:31 signpost: yeah, it puts a limit on the rateless code for example |
21:59 |
asciilifeform |
... simply will take longer. |
21:59 |
signpost |
yep yep |
22:00 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1081048 << see above |
22:00 |
dulapbot |
Logged on 2022-02-26 16:22:18 signpost: yeah curious to hear from asciilifeform why one can't use MTU discovery with don't-frag |
22:00 |
signpost |
ended up reasoning out the "MTU ain't fixed for all time, stupid" matter slightly farther on |
22:01 |
asciilifeform |
if the dice fall a certain way, you can indeed reliably send/receive 'obese' packets. problem is, can't know apriori, and can't rely on the dice to fall thusly |
22:02 |
asciilifeform |
the other thing, afaik we dun currently have 'don't reassemble' turned on anywhere, so can't even reliably test (atm you'll see 'obese' packets go through as if they actually had, whereas what actually happens is that os reassembles, and this is a ddosable mechanism , cuz frags aint signed ) |
22:03 |
signpost |
sounds like best move is to pack several check blocks of min-UDP size into the fatties, such that they can be used with the smaller ones. |
22:03 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1081061 << capital markets still for (indirectly, via pension scams etc) 'taxing the poor' lol |
22:03 |
dulapbot |
Logged on 2022-02-26 16:34:56 signpost: because casinos are for taxing the poor, and capital markets are for taxing the rich, of course. |
22:04 |
asciilifeform |
signpost: well several luby quanta at any rate |
22:04 |
signpost |
right |
22:04 |
signpost |
if they arrive, hooray, maybe they unlock a few xor chains of data for ya. if not, just wait longer for enough tinies. |
22:04 |
asciilifeform |
aaha |
22:05 |
signpost |
one'd have to reason about making sure you're not biasing parts of the distribution into fat dispatches too |
22:05 |
signpost |
such that you're harming the rate at which the tinies converge. |
22:06 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2022-02-26#1081051 << you dun have control over latency at any juncture, important to note. (in olden days existed e.g. 'frame relay' where in fact reserved timeslots. but we dun have this on general-purpose internet) |
22:06 |
dulapbot |
Logged on 2022-02-26 16:25:58 verisimilitude: Another thing about the Internet is how it's really not designed for anything but messages. It can't really provide a stream like the telephone network could, because the telephone network effectively established a physical line. A packet-switching network could do this, by reserving buffers and establishing a chain of reserved buffers, but this would need some way to restrict the reservation to avoid i |
22:07 |
asciilifeform |
i.e. you can't actually emulate circuit switching over a packet network, any moar than you can 'push a rope' |
22:08 |
* |
asciilifeform must bbl, very short on time, but hopefully answered basic q. |
22:08 |
signpost |
appreciate it, take care |
22:11 |
verisimilitude |
Well, what I described assumed the packet-switching network could make guarantees, but this requires each switcher to communicate with each other; it doesn't apply to the Internet, no. |
22:13 |
verisimilitude |
Here's a fun fact, asciilifeform, I came across ``a priori'' not long ago, and realized it was wrong; the Ā PRIŌRĪ is dative, whereas Ā requires an ablative; the correct form is actually Ā PRIŌRE. Isn't that neat? |
22:16 |
verisimilitude |
I don't believe the creators of the Internet would've cheered letting the telephone system literally rot to be replaced by it. |
| |
~ 38 minutes ~ |
22:55 |
mats |
signpost: people seeing frothiness and subsequently hedging, but also a general rise in options contracts vol due to retail shmucks |
22:57 |
mats |
hard to find think of anything novel or interesting to say about something so broad |
23:10 |
signpost |
so help me see the difference between your permabull comments and what was said before and during any previous major crash? |
23:13 |
signpost |
https://www.bls.gov/charts/employment-situation/civilian-labor-force-participation-rate.htm << seems like the demo with peloton subscription money is dwindling. |
23:15 |
signpost |
https://app.koyfin.com/cmty << gotta explain why the inputs to manufacturing and food skyrocketing are actually good, too. |
23:16 |
signpost |
at the very least, they compress profit margins, assuming companies can eat the higher input costs. so then you have to explain why P/E will increase forever, rather than mean-revert. |
23:17 |
* |
signpost is hearing quite a lot of "hopeless if true, so can't be so" from meatwot. reality doesn't care. |
23:20 |
billymg |
"if true" meaning if market crashes? |
23:20 |
signpost |
take it as broadly as you like. |
23:25 |
billymg |
well doesn't "crash" mean relative to USD? so if you're |
23:25 |
signpost |
s&p is already down about 10%; that's not froth. |
23:25 |
billymg |
sorry, ...if you're "bearish" you're actually bullish on USD |
23:26 |
billymg |
hard for me to be bullish on USD these days |
23:26 |
billymg |
seeing as how the collective west is trying to win a war against russia by #cancelling putin |
23:26 |
signpost |
billymg: I mean, the central bank definitely can yank enough liquidity out of the system to crash both assets and prices of whatever slop is left on the grocery shelves. |
23:29 |
verisimilitude |
Putin will buckle any day now; he can't just ignore criticism. |
23:31 |
signpost |
verisimilitude: you know, "unprecedented" may be the most american of the adjectives. |
| |
↖ |
23:31 |
verisimilitude |
I see countless people on Twitter, Reddit, and Hacker News looking to bring Putin to heel. |
23:31 |
verisimilitude |
He's just like Voldemort, as we know. |
23:31 |
billymg |
verisimilitude: you had me there for a second, thought you were being serious |
23:31 |
signpost |
hahaha, me as well. |
23:35 |
signpost |
billymg: https://app.koyfin.com/share/4eb13db5fc << "this time is different" |
23:36 |
signpost |
then later they'll say "unprecedented drop in equities!" |
23:36 |
signpost |
"unprecedented inflation/supply-chain issues/shortages" |
23:36 |
billymg |
would be lulzy if equities drop and btc moons |
23:37 |
signpost |
btc seems pretty vulnerable to the money printer mid-cycle |
23:37 |
billymg |
so they can queue up the "holders are literally terrorists" headlines |
23:37 |
signpost |
but '24 is not too far away. |
23:37 |
signpost |
yeah, I'm glad I have no BTC. |
23:37 |
billymg |
yeah, same |
23:41 |
shinohai |
verisimilitude: Here's how the West will defeat totalitarians like Putin: http://btc.info.gf/uploads/TROONS1.png |
23:41 |
signpost |
oof |
23:43 |
billymg |
lol, are shinohai and i the only ones here who browse the shitposts on 4chan for entertainment? |
23:43 |
mats |
i wouldn't characterise what i said as permabull |
23:43 |
shinohai |
Only worx if you use 4chan-X |
23:44 |
billymg |
shinohai: what's that? |
23:44 |
verisimilitude |
Oh, I'm not the only imageboard user here? |
23:44 |
shinohai |
billymg: It's a userscript for tampermonkey that let's you filter bait posts, expand images, etc. |
23:45 |
billymg |
shinohai: ah, neat, will have to check it out |
23:45 |
mats |
moreover, plenty of people have been predicting crashes for years, while losing money for years, like einhorn |
23:46 |
mats |
its very edgy and drives clicks, but unless you're einhorn and putting money where your mouth is, it rings hollow |
23:46 |
verisimilitude |
So, be either of you users of 4chan's /pol/, billymg and shinohai? |
23:47 |
signpost |
mats: I gave an alternative to btfd in stocks. |
23:47 |
signpost |
will gladly eat my hat right here if it eats shit. |
23:48 |
shinohai |
billymg: https://www.4chan-x.net/builds/4chan-X.user.js |
23:48 |
shinohai |
verisimilitude: occasionally browse it when bored. |
23:48 |
verisimilitude |
I'm just wondering if either of you be ``election tourists''. |
23:49 |
mats |
i bought the bitcorns and don't invest in equities, except i guess the few low four digit index fund accounts over the years, but i don't think it'll be immune to crashes, they are risk-on assets |
23:49 |
signpost |
yeah, agreed. bitcoin will have a near-term crash if the rest crashes. |
23:49 |
mats |
there have been so many crashes i stopped counting |
23:49 |
billymg |
verisimilitude: yeah, i load the catalog now and then when i'm bored |
23:49 |
verisimilitude |
I no longer use 4chan much, but it's how I wound up here. I seem to recall seeing Loper-OS around a decade ago on 4chan's /g/, and reading it over the years. |
23:50 |
verisimilitude |
There are much better imageboards, however. |
23:50 |
mats |
pton's free cash flow numbers are very good, from what i remember in their last report |
23:50 |
shinohai |
/g/ is about the only sane board left. /b/ generally gets filtered to single page |
23:50 |
signpost |
mats: just looks like an exemplar of "luxury good" |
23:51 |
signpost |
maybe the techtard class is large enough for them to sustain it. |
23:51 |
verisimilitude |
That board was garbage back when I still used it, shinohai. |
23:51 |
mats |
same thing with resto hardware, its a veblen good and will resist declines during a prolonged market rout better than most |
23:51 |
shinohai |
Well to be fair, it's *always* been garbage. |
23:51 |
signpost |
mats: yeah, quite agree that e.g. Square is a good company. |
23:51 |
mats |
veblen makers are a decent inflation and rising inequality hedge |
23:52 |
verisimilitude |
It's been mentioned before, but care for a link to a decent imageboard? |
23:52 |
* |
signpost considered going there after last $megacorp job, but farting around with own projects for a bit first. |
23:52 |
signpost |
my last company was restaurant online ordering, started right after 08-09 crash, was timed well. |
23:52 |
signpost |
that industry churns over continuously. |
23:53 |
mats |
is anyone earning money yet |
23:53 |
mats |
i haven't watched that segment at all, but i noticed uber only recently turned a profit |
23:53 |
signpost |
we did fine, but we were tiny, and didn't touch fulfillment, just facilitated it with other providers |
23:53 |
mats |
after losing uncountable billions for, like, a solid decade |
23:54 |
verisimilitude |
If only I'd the opportunity to lose billions of US dollars. |
23:56 |
mats |
https://www.macrotrends.net/stocks/charts/UBER/uber-technologies/net-income-loss k e k |
23:57 |
signpost |
cash incinerator |
23:57 |
signpost |
and they can't stop, because the driver incentives are the only way to keep capacity |
23:57 |
signpost |
delivery sucks; I would never touch it |
23:57 |
verisimilitude |
I've been in awe at the behaviour these companies have inspired. |
23:58 |
verisimilitude |
``Noooooo, I can't talk to a person on the phone to order my food, I have anxiety, noooo!'' |
23:59 |
mats |
there were a lot of these in taipei, food couriered around by kids on mopeds |