helix278

I write all production code by hand, it's the only way I'm sure I know what the code does. Moreover, when the problem space or code base is more complex, it is necessary to keep in touch with reality and test your understanding by writing the code.

I see colleagues using LLMs more and more, and the number of bugs found by QA is increasing dramatically, because the LLM was not thorough enough and the developer (now reviewer I guess) missed the edge cases too because the code, tests and comments looked convincing.

That is not a benchmark result. That is a different way of thinking about how AI models get built.

tiresome

I'm always surprised at the high frequency of major typos in titles from HN posts

Letting the bytes go by

Ligatures exist

You make it sound like reducing the big O complexity is a dumb thing to do in research, but this is really the only way to make lasting progress in computer science. Computer architectures become obsolete as hardware changes, but any theoretical advances in the problem space will remain true forever.

I have had treesitter crashes in the past editing markdown, causing helix to segfault, but the particular bug that caused my crashes has been fixed since years.

Although I'm more of a functional programmer and ADT addict, I still find this language appealing. Reading the manual, it feels like a lot of correct design decision have been made, contrary to languages like C++. Most notably for me: 1) Memory allocation and reference is restricted, leading to less problems with memory management. 2) Syntax is more flexible, allowing for higher levels of abstraction.

I'm not a compiler dev, but I know that many functional programming languages struggle with this in the same manner if the target platform does not support TCE itself, and therefore require trampolining.

But entirely definable in user code, so an effect is essentially a set of possibly impure operations you can perform (like I/O or exception throwing), and a function that exhibits that effect has access to those operations. Of course the call sites then also exhibit that effect, unless they provide implementations of the effect operations.

TCO (tail call optimization) is often confused with TCE (tail call elimination), the latter is a runtime guarantee whereas the former is a compiler's best effort attempt to statically optimize tail calls.

So cutting edge you wouldn't even need a kitchen knife

I like that there is plenty of room for comments, and the multiline extension is also cool. The backslash almost looks like what I would write on paper if I wanted to sneak something into the previous line :)

Quite nice. Though I like texts with exercises better

You're right, but history has shown that increasing the productivity also leads to increased demand, which is good news as long as humans remain involved for some percentage of the work

What is the main difference with just?

Would it be feasible to build a larger application using a language like Factor? I have no experience with concatenative programming, and it seems interesting, but I also get the feeling that larger systems would become rather complex and hard to read.

There is a .NET library that also wraps DOT and provides programmability: https://github.com/Rubjerg/Graphviz.NetWrapper

Very nice. I would like this for Helix, and allow bringing your own config!

Note that this won't solve all cases. For example, you still have to watch out when merging branches with different submodule commit hashes that you run submodule update while merging.

You might find this git option helpful: https://git-scm.com/docs/git-config#Documentation/git-config...

Software Unscripted, by Richard Feldman

This sounds like a weaker form of algebraic effects, but it is still cool to see such a language and see what you can do with it.

Wrong thread?

What does that mean in normal english? I'm not a native speaker, and my question about the other thread was genuine.

What happened to the other thread about this news item?

https://web.archive.org/web/20240309045142/https://www.newyo...

They're attaching metadata to the videos which can be easily removed. Aren't there techniques to hash metadata into the content itself? I.e. such that removing the data would alter the image.

There are warnings given by analyzers in some cases, but it's not exhaustive

What is hypertapping and tapping?