Hacker News Reader

Comment by higerordermap | Hacker News Reader

A serious number of CS-defying claims, most of the rest being intellectually dishonest.

To be fair I didn't see a ton of "CS-defying" claims as in they were impossible, however he made a ton of claims that would put his language on par with languages with huge teams of people, several with major companies backing them, with far more resources.

It was like, really? You're going to have all these features that teams of engineers have been working on for decades (in many cases)? And do them flawlessly? You're also going to beat Zig, Nim, and Go - in far less time?

I've got no problem with projects like this for fun, but it's being marketed as a full fledge project.

There was at least the CS defying claim of "no GC, no leaks, simpler than rust" (I'm paraphrasing, but it was along these lines). Sounds easy when you just write it as a checkbox on a list of exciting features.

There is no GC, and the language is indeed simpler than rust.

The "only" thing missing for now is "no leaks", but we are working on it.

You may be interested in https://aardappel.github.io/lobster/memory_management.html .

> regarding lobster memory management:

> the scope of anonymous functions cannot outlive variables' scope they reference, however V plans full closure support

> lobster compiles whole codebase at once (so it can properly track data flow across module boundaries), however V plans incremental compilation support

> both points are direct consequence of lobster's choice for memory management, ie solving them would mean moving away from lobster's strategy as far as I undersand

https://github.com/vlang/v/issues/1247

Lobster's approach at least has some theoretical rigorousness to it. What V is doing is quite different both in terms of the implementation and the requirements from the language itself and they have never been shown how their version is sound.

Reply on news.ycombinator.com