klodolph

I see that a lot of these are markets.

Yes, it’s hard to predict markets. Because anybody who can successfully predict markets, does so, makes money, and changes the market so their predictions lose their edge.

Time series forecasts are a lot easier if you are forecasting, say, disk use in your servers or whatnot. (By “easy” I mean you can do a simple prediction and get useful insights.)

Rust is designed to make this type of issue impossible, but that assumes that you can correctly encode object lifetimes in the kernel in a way that allows the compiler to check them.

So I would say that any easy answer like “this would not compile” would just be a guess, because you would want to know more of the particulars in order to answer this question.

I know that this is kind of a non-answer, but if you want to write a kernel in Rust you have to figure out boundaries for where unsafe {} are. In a kernel, there are probably large chunks of unsafe {} and the Rust compiler prevents certain bugs outside unsafe {} assuming there aren’t bugs inside unsafe {} that would prevent the type checker from doing its job correctly.

Copybara can do that but I think it will be annoying and tedious to use it that way. More annoying than the problem of extracting a library or shoving some files in a separate repo.

Been using this for a while, mostly when I make a tool as part of a larger project and the tool is big enough to deserve its own release.

It’s powerful enough to do a whole bidirectional shipping operation where you export and import code—no thanks, that’s a hassle. I use it mostly for a simple fire and forget export, where I take a folder out of its original repo and preserve the history. Then I just move development to the new repo. The new project layout can be completely different, but Git blame works and I’m happy with that.

There are some really powerful things in Williams’s performance. You’re right about the nature of Hollywood—you work with what you got. But Hollywood also eats its own tail. Filmmakers grow up watching films, and they tend to draw from other films, the same way that LLMs just kinda rearrange pieces from ingested text (which is an oversimplification, bear with me).

Like, there’s something special about The Lord of the Rings that is not there in, say, Wizard’s First Rule. I don’t want to pick on or make fun of things too much, but Wizard’s First Rule seems to more rearrange existing ideas in the genre rather than drawing from something else.

What I’m saying is that there’s a broader problem with stories in film and books where you can tell that the stories are written by somebody who leans too heavily on other stories and books. Movies are a kind of alchemy where writing, direction, and acting intersect so we can’t explain everything away as easily as we can with books, but I want to say that the monologue is weak in the script, and Robin Williams and Gus Van Sant manage to elevate it. The direction is absolutely stellar, the acting is on point, but I hate the actual words in the monologue.

That's the entire point.

Yeah, that’s the entire point, and I think it makes no sense.

People will take off their mask if they feel like it’s safe. They don’t take off their mask because you make a big speech and confront them about it. The perspective that this speech has—it is telling me that a forceful, paternalistic approach can fix people. If somebody needs to talk but won’t, you can break down the walls. I disagree with that.

I remember feeling like the world worked this way. I remember feeling that maybe I could be broken down and fixed by the right person. Back when I was the age Matt Damon was when he wrote this movie, maybe I would have agreed with it.

But what have I seen since then? I’ve seen that these speeches alienate people. That the person giving the speech rarely understands their target well enough to know which buttons to push. That trained psychologists know better than to try and take their patients down a peg.

Yes, that’s what I was getting at.

Matt Damon didn’t have the kind of life experiences to write a good version of that speech, so we got the version in the movie, which sucks.

I completely agree. I watched the movie recently and really hated that monologue. I truly hated it. It seemed just so out of character for somebody who is supposedly a psychologist in their mid-40s—the whole speech is taking Matt Damon’s character down a peg. The fact that he’s downplaying his own experiences (he doesn’t understand what it’s like to be an orphan) doesn’t make the speech any better.

This movie was a disappointment when I finally got around to watching it, recently. The movie was just so naked in how much it presented the perspective of somebody in their 20s. All of Robin Williams’s dialogue sounded straight out of the mouth of a 20-something kid trying to sound mature for his age.

And if I asked you about love you probably quote me a sonnet. But you’ve never looked at a woman and been totally vulnerable.

The monologue is just so damn trite! When I say that it sounds straight out of the mouth of a 20-something kid trying to sound mature for their age, that’s because I remember hearing a lot of speeches like that, extolling the virtues of life experience, from kids in college, back when I was in college. Kids in college understand on an intellectual level THAT experience is valuable, but when they try to articulate it, these speeches end up sounding parroted, sounding like they’re putting on an act, sounding like they’ve gotten their life lessons from movies. Kind of the same way that ChatGPT gets its lessons by ingesting massive volumes of text.

I’m going to be honest here—I kinda hate the Good Will Hunting script. I really do. The movie was saved, SAVED from oblivion by some truly stellar acting from a few phenomenal actors. But that script, that script… there is so much wrong with it.

If there’s one thing that the movie really taught me is that “write what you know” is serious business. LLMs don’t know much, and that causes a lot of problems with their output. Matt Damon didn’t have the experience that comes with age, and so when he tried to write a monologue that extolled the virtues of experience that comes with age, it had similar problems. :-( The movie has an interesting thread of a story at its core; I don’t want to give the impression that I have nothing positive to say about it. There are some really good bits. The monologue from Robin Williams is not one of the good bits.

Isn’t species fitness somewhat weak, though? I would want to think about this in terms of gene fitness. Over long periods of time, in different situations, genes that promote one strategy or the other will dominate part of the time and lose part of the time.

Unfortunately they seem to be useful, if for no other reason than parents want to take picture of their children at a moment’s notice. AFAIK this is the killer app for these glasses.

I never claimed it's representative of anything.

When I write “it’s not representative”, I am hoping to communicate an opinion and not hoping to refute a specific claim you made.

I appreciate unorthodox explanations, and I like to collect them—but sometimes the explanation just doesn’t “land” and in this case the explanation landed especially poorly for me, and I also identified some errors, like the claim that “si” is not in the kana, or that hanasimasu is incorrect—I know that you don’t accept my viewpoint that this is incorrect—sometimes it happens that explaining your point of view or reasoning in more detail doesn’t result in agreement.

What is certainly true about linguists is that they do not present explanations in ways that are consistent with each other—and sometimes not even self-consistent, but they are upfront about the tradeoffs (they present theories and acknowledge that the theories contain errors) and they decompose what they present into different topics like phonology and morphology.

I have read some linguistic texts on Japanese (not very well! It’s a difficult subject). What I saw is a lot of variation.

I think it’s fine that your explanation is good enough for you—that’s exactly what you expect when people make individual breakthroughs in understanding when learning a subject. Sometimes those breakthroughs do not translate well to other people or translate well to lessons and that is the main gist of what I am trying to articulate.

I also don't think this is especially deep—that I think this approach to teaching verb conjugation is a bad approach and I don’t think there is this much to say about it.

People off the street are obviously not learning Japanese verb conjugations in isolation. If they are learning it at all, they probably have some broader goals involving spoken or written fluency, and these people are gonna fire up DuoLingo or sign up for a class or something. Japanese verb conjugations are simple and easy to learn but they are usually not taught day 1, and you are not expected to learn the whole table at once, but one or two conjugations at a time along with practice using that conjugation.

So if the pitch is, “this system works for teaching English speakers off the street how to conjugate verbs in Japanese” it seems to me like the goal is a little artificial and maybe not representative.

I think the call to “engineering mindset” may be illuminative, because engineers are likely to have unwarranted confidence in fields outside of their expertise. Engineers in practice often think that they can use engineering skills (broadly speaking) to solve education problems, learn foreign languages, or solve social problems. The phenomenon is sometimes called “engineer’s disease” or “engineer’s syndrome”. What I wonder is whether there is something about engineering mindset that is counterproductive outside of engineering fields—this seems plausible, because it explains why we don’t just teach everyone to use an engineering mindset.

I understand why you wrote the article this way, I think the lesson here is “we have learned why Japanese textbooks do not teach the content in this order” and there are a couple reasons why this order is not good:

1. It relies on people not understanding certain things. In general, you cannot expect people to have exactly the right misunderstanding necessary for a lesson.

2. Spending extra time with Hepburn reinforces it, and it shouldn’t be reinforced.

I am in general extremely skeptical of lessons which try to engineer a way for the students to make mistakes. What I have seen in real classrooms and in informal teaching is that the mistakes are habit-forming and the outcomes of this kind of engineering are unpredictable.

Mistakes are appealing to the developers on HN because we understand things more by seeing them fail. But this does not mean that you can engineer somebody to experience the same moment of enlightenment that you did, because it requires constructing the same (incorrect) mental model that you had when you made that mistake that led to useful insight, and it both difficult and counterproductive to try and make that happen to students. Give people the best chances to learn by giving them the best chances to avoid mistakes, and the mistakes and insight will happen organically on their own, in unique ways for each student.

In general, I find your attitude a bit condescending.

Yeah—I can understand why I’d come across as condescending. There’s a balance here—I want to be clear when I say that I have problems with the article, but I don’t want to be hurtful and I don’t want to make criticisms that are not supported by the text.

Rather than defend my comments as “correct” let’s say that I failed in my goals of not coming across as condescending. The reason I want to frame it this way is that similarly, I think the article failed in its goals as coming across (to me) as “look at this neat thing about Japanese”.

It is just kind of the nature of written communication that it takes a lot of editing and polish to make it clear, correct, and concise. I had the good fortune to sign up for Japanese 101 when my professor was in the middle of writing a new Japanese textbook—it was pretty exciting, with the changing lesson plans, the flock of master’s students hanging around, revisions and drafts to teaching materials, and those endless hours of classroom observation. The teachers occasionally gave us a “peek behind the curtain” and explained why they chose to teach things a certain way or another. I’ve rarely gotten that kind of explanation in any class that I’ve taken so I thought it was pretty special.

I don’t expect you to put in the textbook-level of polish into your article but there is a kind of verbosity (the article is long, which makes it kind of hard to respond to because there is just so much to sift through), there are some problems with clarity (the issue of romanization and orthography is mixed in with the conjugation, and maybe it would be better to separate those issues) some problems with correctness (various) and some problems with completeness (the patterns omit some conjugations that I think you don’t know, and I don’t think they follow the pattern).

I have certainly put effort into articles that have gotten brutal negative feedback; I think it was right for me to write the article, and then feel like shit from the feedback, and then maybe retract and revise it. If there is one actual error here, a true error, I think the error is fighting out criticism in the HN comments.

The online spaces can be really discouraging, true—but it can also be really discouraging to be in a classroom or in a foreign country, struggling to use a language you barely know. Meanwhile, there are also a lot of ways to spend effort trying to learn a language without gaining mastery. Truly frustrating.

If you want some decent level of fluency then there is something you have to do, which is to communicate with other people in context and with specific goals in mind (get information, give information, make a request, etc). Whatever you can do to arrange for that to happen is probably more valuable than anything you can do online or with books. I personally like to recommend finding classes at a local college.

If you can’t get that, then I think the next best thing is reading and listening.

Drills are also necessary but you can easily fill your time with drills without advancing your ability to communicate or understand people.

There is plenty of research about what is / is not effective when it comes to learning languages so I encourage people to at least take a look at the results of that research rather than just go with whatever people recommend online (I’m just some random person online, I may be no better than the next). AI tools reportedly have a positive effect but they are not nearly as good as human interaction.

I think the reader would just read the next section where I use your argument to critique my own approach? And then make up their own mind whether it’s defensible to do something in the article, to raise pros/cons for why I did it, and then to keep on with the choice.

I think that’s a long wait; I don’t want to rely too heavily on analogies but it is like teaching somebody arithmetic roman numerals and then explaining in a parenthetical that there are other ways to do arithmetic (but not naming them). Maybe the reader can make up their own mind—but I don’t think the pros and cons are raised in the article, or if the are raised, I couldn’t find it.

I don’t want to pile on here but it sounds like you are, in this conversation, learning about why the different romanizations exist and what the pros and cons are. Or if you already knew, you are getting what they call an object lesson. (Like you noted—in Hepburn, ji and zu correspond to two different kana each.)

As I noted somewhere else, you could imagine that I’ve chosen IPA notation instead.

This just resurfaces a similar problem with different symbols—if you put your IPA notation in slashes // you get phonemes, which will get you something mostly equivalent to Kunrei-shiki romanization. If you put your IPA in brackets [] then you get something sort of equivalent to Hepburn (in that it’s designed to show pronunciation). Both choices will on some level obscure a regular pattern that could be revealed with kana or romaji. Orthography is funny like that; in both Japanese and English it can show the origin of words even when the pronunciation changes.

I think the other lesson here is that students will mostly learn morphophonology intuitively by absorbing examples with some light explanations of the rules, and if you overexplain the rules you end up with too much “scaffolding” which gets in the way. Like when people use mnemonics or try to memorize kanji by thinking pictorially.

That note isn’t much to go on.

I think the choice is not a good one, whether it is deliberate or by accident, it is not a good choice either way. The main caveat to Hepburn is that it’s unsuitable for explaining how Japanese works and it’s unsuitable for learning Japanese—so before you start working on verb conjugations, you pick up kana or one of the romanizations which is more aligned with Japanese.

The idea that you “shouldn’t think in romaji” is really “you shouldn’t think in Hepburn”. This is an important distinction! Japanese has a relatively small inventory of phonemes, somewhere around 20 or 22 of them, and they map very neatly to the latin alphabet.

But the article doesn’t make this distinction, and seems to rely on confusion induced by the Hepburn romanization in order to make its points.

IMO, this is kind of like seeing an article about how monads are burritos. Thinking that a monad is a burrito does not help me understand monads.

Nomu -> noma-nai / nomi-masu / nome-ru / nomo-u

Miru -> mi-nai / mi-masu / mire-ru / miyo-u

The ichidan and godan verbs are not assigned different categories because existing scholars of Japanese are just bad at explaining how they work, and you can still understand them just fine in romaji. I put the hyphens above to mark a place where you could think that the verb ends and the common conjugation forms end, and you can see that the part on the left has somewhat different rules for ichidan and godan verbs, even when you apply the “tricks”—but some of these forms may be unfamiliar if you are are starting out (are you familiar with miru -> miyou conjugation, or miru -> mirareru?)

The systems are obviously not isomorphic—Japanese kana are not entirely phonetic (they are just mostly so) and the different romanization systems choose differently whether to follow orthography or phonetics more closely.

hanas* + (i)masu = hanasimasu (wrong!)

I cannot wrap my head around how this line in the article could be defensible. Like, if I don’t understand how Japanese is pronounced or written, and I just rely on Hepburn, I guess pasting these fragments of Hepburn together don’t produce the right Hepburn in the end?

YMMV indeed, but I think the lesson here is “this is why you don’t use Hepburn when you’re writing an article about Japanese verb conjugations”.

Hepburn does make sense for somebody with zero knowledge of Japanese but it just gets in the way when you are trying to explain how Japanese works. So lesson zero is “don’t rely on Hepburn” and IMO if you are interested in pronunciation and listening you should be using audio as your primary source.

First, As a basic part of Japanese language education, students are expected to be familiar with different romanization systems. If you ask a student where “si” is in the table, they should be able to find it. If a student says “it’s not in the table” then they’ve failed the lesson or there is something wrong with the teaching material.

Second: am I arguing that the choice of using Hepburn here is somehow bad? Yes, that’s correct. I think Hepburn is a bad choice here. A good choice is Nihon-shiki. JSL romanization is also fine.

That’s just false, “si” is in the hiragana table as し. The romanization “si” is /si/ which is pronounced [ɕi] (or [ɕi̥] or some other possibility). This is basic Japanese phonetics.

If you fix all the errors that are in the article, at best there is an argument buried here that Hepburn romanization should not be used to teach Japanese to English speakers—but I think that point is really my own argument that I’m making with the fragments of the article that make sense.

Romanization can be more consistent with Japanese phonetics or it can be more consistent with English phonetics, and the Hepburn romanization is more consistent with English phonetics, which is why it’s a good choice for English speakers that don’t know Japanese, but a bad choice for English speakers who are trying to learn Japanese.

I think it’s probably a mistake to use Hepburn if you’re learning Japanese, it kinda gets in the way. Either learn kana (which takes what, a week?) or use one of the other romanization systems which maps more cleanly to Japanese orthography

Because the author of the article hasn’t internalized that si is pronounced “shi”, is my guess.

They’re sometimes called “semi-irregular” because they are mostly regular with, like, one deviation. The list is not long and it is quick to memorize.

hanasimasu = 話します

:(

Romaji are great, and in some ways more instructive because they reveal patterns which are otherwise a little hidden. You just have to realize that S+I is shi, T+U is tsu, etc. I don’t want to get too deep into it but there is a regularity to the language, and rules, and different choices of writing system reveals different pieces of the puzzle.

Next, the conjugation itself. There are massive categories of conjugations missing! Like, how do you get from taberu / nomu in this system to tabereru / nomereru? It turns out that these ichidan and godan verbs actually do have some differences in conjugation. Who’d have thought? (There is the -i stem, but there are other forms.)

I would encourage people who have the pragmatic bent to read about JSON-LD from the Google documentation for web sites;

https://developers.google.com/search/docs/appearance/structu...

You’ll also notice that a lot of the information is relevant to only a small subset of sites. Rotten Tomatoes can publish the critic rating for movies using JSON-LD, but that’s not relevant for me (even if I write a review for a movie).

JSON-LD is nice because it’s easy and it is actually used by search engines. Yes, it can duplicate information in the web page itself, but I think the dream of perfectly annotating information so it only appears exactly once in your document is, well, a dream of spherical cows and massless ropes. It takes human effort to make a webpage and I am ok with a little duplication in the final product. My <h1> duplicates information in <title> anyway.

I think if you are using Dublin Core, it’s because you’re a library. Maybe I am off the mark, but that is the sense I get from this—not all these standards should be used for all pages on the web.

I think you should just think about what metadata you actually care about, and the main metadata I care about (choose your own list) is authorship, publish date, last update, subject keywords, thumbnail (OpenGraph 1200x630), and summary.

There’s a long list of additional metadata that I could put in my webpages because there are standardized ways to do it, but, why bother?

IMO this is going overboard. Any time you are duplicating data from HTML into JSON-LD, consider just omitting that data from JSON-LD, unless the data isn’t consistently present in HTML (because it is a bitch to be consistent about this stuff).

I tried using RDFa and liked the property that it was theoretically less redundant, but switched to JSON-LD because it JSON-LD is just easier to get working. And this is speaking as somebody who uses a hand-rolled static site generator—the issue here is that whether information is present in the raw HTML is something contextual, and if something isn’t present in the HTML then you need to put it somewhere else or it’s not mechanically parseable from the page. Like, to a human reader, a post on “Alice’s Blog” is assumed to be authored by Alice, so I may omit the “by Alice” text from the document, and then I would want to put that metadata in the page some other way.

Putting the metadata in JSON-LD lets me just be dumb about it. The metadata is always in JSON-LD, and the HTML may or may not contain an explicit representation of that same metadata. Easy.

But the JSON-LD does not need to contain the URL of the page (which is <link rel=canonical>) or the title (which is in <title>), for example.

I have used JSON-LD in my own websites and found that it fills a separate need from semantic HTML. Your semantic HTML will specify things that the browser processes, like the title and headings. The JSON-LD data is metadata, like date created, date updated, tags, authorship. These things can be expressed in the HTML using micro data, but I stopped using micro data because JSON-LD was easier.

The JSON-LD I populate from the same data that I use to generate my site, and I use the JSON-LD metadata to generate things like index pages (list of blog posts from 2024, all posts related to topic X, etc). The main consumers of JSON-LD are search engines.

If you are interested in getting offended, then think about how we are also putting OpenGraph metadata in our web pages. Two different metadata formats for the same page.