cobbal

How long before a plague of brighter-than-white ads, I wonder...

For a really in-depth, online resource on Bézier curves, I constantly consult https://pomax.github.io/bezierinfo/

Goodhart's law kills again.

I never realized wikia was so closely connected to the wikimedia. I'm a bit shocked.

blacker_than_the_blackest_black_times_infinity

It's not like this is a new thing. Those of us at high risk for COVID because of other conditions have been living in that reality for a while now. More so before the vaccine was available.

True. And the case of what upscaling does is technically truly interesting and nuanced... but that doesn't mean that people don't have a sense of what zooming in does to an image, CSI "zoom and enhance" be willfully ignored.

Most states use a form of common law, which is completely based on what is common practice and what is common sense. Or at least was in England hundreds of years ago...

Not to mention the fact that dithering will look different depending on the display’s gamma calibration...

“It’s not real engineering if someone can’t get hurt” seems like a strange definition to me.

What’s flexible for a dynamic scripting language is often cursed from a static perspective. Knowing what imports resolve to statically can be nice.

The problem is that “radical” is in the eye of the beholder. If a language is so poorly specified that developers and the specification don’t agree on the meaning of a program, that leaves optimizers in an unwinnable situation.

That can be fixed by using “EDITOR=true git commit --amend --allow-empty”

Eventually, a long enough ship should be supported by 3 or more waves at the same time, assuming waves have some sort of maximum wavelength. (Disclaimer: no idea what I’m talking about)

The fact that CompCert, seL4, and similar projects exist is proof that a large amount of composition is possible if you have enough time to spend on it. It’s true that you can’t compose arbitrary formalizations, but you can architect your formalizations to be composable. Functional or stateful matters less than the engineering work required.

The ISA alone doesn't mean much, but there's OS compatibility too. MacOS will run iOS applications out of the box.

So what you're saying is that we need more militarization of the police?

True... but I think it’s just called an “investigation” when done by the authorities.

You were downvoted because you stepped into a long-running and bitter ideological war: Bayesians vs frequentists.

As far as I can tell... Bayesian statistics is a model of the real world, fuzzy logic is just an easy to compute, bad approximation of reality that assumes everything is independent. (Someone correct me if I’m wrong; I’ve only studied logic and Bayesian stuff, not fuzzy logic)

Cool! I've recently started writing many of my shell scripts in swift [shameless plug], but nice command line argument processing was something I wasn't looking forward to tackling. Now it looks like I won't have to :)

[shameless plug]: https://github.com/cobbal/swsh

Description:

A shell-scripting library for Swift, inspired by scsh.

swsh makes writing shell scripts more fun by exchanging bash (or similar) for a better thought-out language like Swift. In the process, a small amount of conciseness is traded for better quoting, error handling, and access to libraries.

That’s some good, bad sci-fi waiting to be written. A new strain of the flu spreads overnight across the globe once it is able to jump between people and the internet...

I can tell the difference at 12. There's a flash test at http://jakemandell.com/adaptivepitch/ and the (probably heavily sample-biased) histogram at the end suggests 10-20% of people can't differentiate 1/4 of a semitone and 60% ish can't differentiate 1/12.

It's two different processes. Normal counting is saying, "take a thing and add one" (a successor ordinal). It's how we get from 42 to 43. If a number can be reached from zero using just this kind of counting, we call it finite.

The new type of counting also allows you to take a different step and say, "take a collection of things and add a new thing at the end" (a limit ordinal). It's how you get from the set of finite numbers to ω. There's nothing that's "one before" ω, but all the finite numbers are before it.

Repeat these steps as much as you want (making sure none of your sets are circularly defined), and you're counting with ordinals.

Call stacks are almost always used in implementations of functional languages as well, even those that support first-class continuations. Even implementations that use CPS as you describe don't necessarily lose call-stack discipline (e.g. guile); that's a separate choice that some (e.g. SML/NJ) choose to make.

The call stack provides very cache-friendly and gc-friendly memory. It's still a very useful construct in functional languages.

Looks like a variant on ml-like `let`.

    let <variable> = <expression 0> in <expression 1>

will bind the value of expression 0 to variable in expression 1.

A likely reason is that it makes it possible to iterate through a list like `'((1 2) () (a b c))`. If you have to do in-band signaling, make it something unique and obscure.

I was getting that error until I disabled the "HTTPS everywhere" extension for the page.

Oops, looks like I made several mistakes in the above calculations.

For a 768 bit number the primes would be 384 bits each, gaps between numbers of the form "27a^2 ..." would be 10^58 apart, and weak primes would be at worst one out of 10^56.

Still extremely unlikely, but not by as much as I thought.

Assuming at least 768 bits (an RSA number of this size has been factored), the gaps between such numbers are more than 10^232. Even if every number of this form was prime, the prime number theorem tells us we can expect one out of every 10^229 primes to be weak.

There could or course be a larger class of such easy primes, but it seems like it's extremely unlikely a prime of this particular form would ever be generated by chance.