norswap

11 maanden geledendiscuss

Not really, he explicitly makes the point that it's easier to write useful simple programs in Javascript without the extra stuff, whereas in Rust you need to know about a whole lot of things to make the equivalent simple program.

norswap

vorig jaardiscuss

I read it and was disappointed, it's okay but probably my expectations were too high for this classic.

I would recommend Snow Crash instead, another Sci-Fi classic from the same era which I found to be a much more enjoyable read & more commensurate to the hype.

norswap

vorig jaardiscuss

Wasn't there a better technical way to block the scraper? Like blocking IPs/domains at ingress instead of serving requests? Probably you do still pay for request traffic though...

Isn't there a market for anti-DDOS third-party services for API endpoints (Cloudflare etc) — through probably for "Free Speech Extremist" that wouldn't be suitable solution, and there are charges too (though presumably when facing a situation like this you actually save money).

norswap

vorig jaardiscuss

Am I the only one wishing I could see the original afterwards, for comparison?

norswap

vorig jaardiscuss

I think it did much better on safety in some tests (no I don't have sources, going from memory :/), but is less capable by design on some more trickier scenarios. Basically pretty much what you would something like this to be.

norswap

vorig jaardiscuss

I used to work on compilers & JITs, and 100% this — polymorphic calls is the killer of JIT performance, which is why something native is preferable to something that JIT compiles.

Also for command-line tools, the JIT warmup time can be pretty significant, adding a lot to overall command-to-result latency (and in some cases even wiping out the JIT performance entirely!)

norswap

vorig jaardiscuss

His "attempt" at self-driving is better than most of the industry, at a fraction of the cost.

The man does and tries things. Man in the arena.

norswap

vorig jaardiscuss

While I agree exceptions should not be used for control flow except in the rarest of circumstances (every rule has exceptions — pun fully intended), I think checked exceptions are very useful.

I wrote a whole article about this, way back when: https://norswap.com/checked-exceptions/

But for the tl;dr: checked exceptions represent somewhat expected issues, things like IO errors. There are two okay ways to handle those: checked exceptions and result types, and one bad way: unchecked exceptions.

Unchecked exceptions end up being undocumented in the API, and people forget to handle them.

Result types add a lot of ceremony, as you have to unwrap them at every level of the call stack. I've seen code dominated by error checking logic many times in C and Go (and the frequent alternative is to ignore error, which is easy in C).

Checked exceptions are much less verbose, you just let the exception pass through and annotate the function signature with the checked exception. It's type-safe, self-documenting and easy.

Unchecked exceptions "leak", checked exceptions propagate at the type-level, which is exactly what you want — making sure every caller is aware of what exceptions might occur and letting them make an explicit decision to handle or pass the buck upwards.

By the way, checked exceptions are coming back in favour with "effects" which are basically a generalization of checked exceptions beyond the realm of errors, trying to encode things into the type system.

norswap

2 jaar geledendiscuss

Why use base 36 and not base 26?

norswap

2 jaar geledendiscuss

Wow that's interesting — trying to cross my eyes produces hellish jitter.

I suspect it's because my left eye is slighty lazy.

But I was able to superimpose the right cat picture onto the left one (it's a lot harder for the more complex sky resort picture). It's pretty eerie, the right picture just slides right up the left one (I did need to figure out the right distance for it).

It doesn't help me pick out the differences though, I mostly only see the right picture, and if try to focus my left eye, the right picture slides out. Still, intersting.

norswap

2 jaar geledendiscuss

Zig — here's the choice matrix:

- You need strong performance, your choices are:

   * C: nope, unless it's super simple library-type code, must integrate as easily as possible with everything C compatible

   * C++: unless you have ecosystem lock-in, avoid this mess, stuck in an evolutionary rut where they can't undo bad decisions and keep adding more complexity (mostly good stuff, but total complexity too high)

   * Zig: better than C & fully compatible with it

   * Rust: ecosystem tailwinds, bit better safety than Zig, steeper learning curve and uglier code

   * Go: ok but GC probably means not the best perf

   * Java: JIT means it can handily beat non-painstakingly hand-optimized C code in some particular cases (requires a lot of wisdom to choose this for perf)

   * something that will incur a lot of pain in some other way like Fortran, Forth or Ada

- You need to export to C code (all of the above have options to import from C code with C/C++/Zig being the easy choices)

    * C: ez

    * Zig/C++: harder but still ez, need to write some wrappers when using features nont in C

All in all, for new system or high-perf work with no ecosystem tie in, it's probably a battle between Rust and Zig. I think I'd use Zig on aesthetic grounds — find it less painful to use than Rust, and I'm not too worried about missing on the extra safety. The converse decisions seems reasonable too.

norswap

2 jaar geledendiscuss

The biggest improvement from this (besides maybe performance) is that it should enable much better manually programmed syntax error messages. Those generated by yacc were pretty shit.

This is generally my #1 reason for using a manual parser — nobody has yet made a pretty good syntax error handling / reporting for parser generators or parser combinators.

It's genuinely very complex — I read the whole literature on that as of 2019 (there's surprisingly little). You basically have to inject custom logic, though there are a few heuristics that you can prepackage and can be useful in a lot of places. But the custom aspect of it means this doesn't play nice with traditional LL/LR parser generators. It could be done for parser combinators (PEG etc) however. Didn't have enough time in my PhD thesis to play with this, and I moved on to other things, but I'm hoping someone will make this eventually.

norswap

2 jaar geledendiscuss

Is there a way to import existing corpuses?

Thinking in particular of browser bookmark exports and Notion, I'm sure other people would like something else. Being able to parse any kind of text file for links would be great + full-text search in text files / markdown.

norswap

2 jaar geledendiscuss

Oh that looks cool! I was really hoping it would have better emoji support than MessagEase as that is the one think I miss (the ability to search for emoji by name) but alas no.

norswap

2 jaar geledendiscuss

See MessagEase for a similar keyboard (not programmer-focused) with less keys but letting you use the swiping motion to type ordinary — great for fat-fingered people.

norswap

2 jaar geledendiscuss

I used to respect patio11's opinions, and obviously he has a lot of valuable knowledge and there are things to learn from him, but some of his takes on crypto have led me to heavily discount his writings on the topic.

On the face of it, this writeup is reasonable.

I'm a crypto founder, and I struggle with banking. We've had two banking applications rejected in Singapore.

Here's the thing: we're not doing anything remotely like what Patrick describes. I think he strawmans a lot.

We're a crypto-focused IT consultancy, but closely affiliated with an upcoming blockchain. The business is set up to be super boring: we bill the client for work and that's it. The company doesn't handle a token, doesn't get revenue from onchain activities, ... The blockchain in question is not anymore risky than any other and in many ways less. Not that the banks know or ass.

They just say "you're rejected and we can't tell you why".

Now this is Singapore, not the US. It's not a targeted plan but just general policies: "crypto is potentially bothersome for us, let's not bother". In the meantime we don't have a bank account. Though I do believe US attitudes and (lack of) regulation probably impact these policies a fair deal.

norswap

2 jaar geledendiscuss

Frontend frameworks really do make things simpler once you go beyond small applications though. Automatically re-rendering when the data/model changes is the right thing to do.

We can argue over the specifics for sure, but simply imperatively changing your UI whenever your data changes doesn't scale very far, neither in term of codebase nor in terms of team.

norswap

2 jaar geledendiscuss

I think what actually convinces me more than tone is nuance. If you can fairly assess arguments on both sides, recognize when either side makes a good point, and mention when you're confused about your conclusion, or when some point of evidence doesn't mesh with the theory. Things that are all one-sided are usually wrong (though I suppose there are cases where the truth is indeed one-sided, it's just pretty rare, and less likely to be things that people that you otherwise consider serious would argue about).

Even that attitude can be weaponized I suppose, if nuance convinced more people, than more people would learn to fake nuance to push their favorite outcome. Though I'd like to think that the process would change them a little bit for the better too.

norswap

2 jaar geledendiscuss

A heuristic is a heuristic, and we can do some of our own research too. I was just defending the OP.

As for my personal position, I remember looking into things a while back and just coming away with the conclusion that nobody really seemed to know. If there's a truth here, it's either considerably more complex than "SFA good" / "SFA bad" (all these interacting pathways, man), or it just makes little enough difference not to matter at all in a way that we can meaningfully measure above the noise.

It seemed clear SFA are clearly not that deadly. Whether you'd be more healthy or live slightly longer if you chose to consume less of them, I have absolutely no clue.

norswap

2 jaar geledendiscuss

I shouldn't enter into consideration, but it's a shockingly good heuristic to evaluate the credibility of people.

I repeatedly found that people who are thoughtful and not snarky — and often, nuanced — when discussing a topic I have no deep expertise on were much more correct on topics that I did have expertise on and could properly evaluate.

We're all on a time budget at the end of the day. And I do share the sentiment: the scientific literature in nutrition is known not to be very good. You don't have to be an expert, it suffices to notice that there are a lot of people coming to contradictory conclusions, and that the consensus seems to have changed drastically over the past decades, not being particularly driven by any groundbreaking changes in available scientific methods.

Submissions

TypeScript Infers Type Variables

TypeScript Distributes Unions

Heritability Puzzlers

Futarchy's Fundamental Flaw

Have you ever seen a goth downtown?

The Draw (and Danger) of Data

How are optimistic rollups secured?

Freaking blockchains: How do they work?

GameStop: Anatomy of a Short and Gamma Squeeze

The Hindley-Milner Type System

Chrome unlikely to support XPath 3.1

Google vs. Oracle – Case Overview

Google vs. Oracle – Case Overview

Covid: 19% of cases responsible for 80% of transmissions

An Implementation of SHA256 in Ada

Swiss Political System: More Than You Ever Wanted to Know (II.)

Tutorial: Building Your Language with Graal and Truffle

Does Tweeting Improve Citations? Results from one-year randomized trial

Between the Spreadsheets

How to Break YouTube (Copyright Claim your own video)

comments