Hacker News Reader

Comment by Decabytes | Hacker News Reader

I don’t understand how they can rewrite the core útiles in Rust with a different license, especially if the intent is a like for like end product?

One of the things “open” game engines run into (IE OpenMW) is that if source code for a game engine they are reimplementing leaks, they avoid it at all costs, since it could “contaminate” the project and ruin their protection as a fair use project.

I feel like this sets a bad precedent for the GPL because if anyone can do that, then it weakens the power of copy left licenses as a whole, but maybe I’m missing something because I’m not a lawyer

View on HN

Think of it like a blackbox reimplementation, not porting existing code to another language:

https://github.com/uutils/coreutils/blob/244693f50e224abf726...

Can we truly believe that not a single one of these authors didn't peak over at the GNU coreutils source to get a handle on how something there worked?

You can believe it or not, but please stop spreading FUD disguised as a question. If you have proofs, you can post them, and then interested people can try to dissect them.

That warning was added 2 months ago.

Maybe they got tired of people making unsubstantiated remarks that were never made for bsdutils, busybox or toybox ? Just look at the level of the current HN discussion.

Is this discussion about bdsutils? Why do you dismiss the whole legal discussion as 'unsubstantiated remarks'? It is certainly a valid issue for a non-GPL rewrite of a GPL software.

I dismiss it because I haven't seen substantiated remarks; i.e proof that anyone actually looked at the source of coreutils to do the the reimplementation.

Nobody has to look at the code to violate the GPL though. An acccidental identical implementation or one spewed out by an LLM counts as a simple translation would largely place the new code under GPL at a technical level. I don’t see the point really for not simply adding rust to enhance a fork of gnu binutils, except for the license change, and the license change is super hard to defend in the long term future when it will be easy to identify translations of the original code.

> Nobody has to look at the code to violate the GPL though

Accidental identical implementation of something trivial aren't covered by copyright, see the Oracle vs Google lawsuit.

LLM completions are something else entirely.

Agreed on the accidental reimplementation of something trivial. The distinction of what is trivial will be harder to clarify in the future, however the language models in the future can also help identify the provenance of substantial parts of the code and eventually it will be easier for GNU, if they care, to have a strong argument. I suspect GNU will not care until somebody plays foul, so GNU can wait for a long time until that happens and until the tools are substantially better. I just find it a little sad that the effort is bifurcated into competing camps. Maybe eventually gnu binutils will also start a rust clone for parts of their binutils and make this effort obsolete at the technical level.

GPL is based on copyright law.

Independent creation is a defense against copyright infringement under US copyright law. (How could it not be?)

They have hundreds of contributors and they don't audit them. This is not a blackbox reimplementation.

> They have hundreds of contributors and they don't audit them

Thank you for copying the SCO FUD playbook, I hadn't heard that one in a while:

https://en.wikipedia.org/wiki/Fear,_uncertainty,_and_doubt#S...

See, the difference here is that SCO was scum, and in this case the people who re-implement GPL tools on permissive licenses are scum.

This was brought up in one of the previous discussion on HN [1], and people found out that indeed this project seem to have copied the original coreutils. There were some name of variables/constants taken from the original code [2]. Also, I am not implying that they are violating copyright (as someone else said, not doing a clean-room implementation does not necessarily imply violating the original license). However, I find it very sad that they replaced the license and are effectively damaging the GNU project. (It is also a bit sad to see your comment, which expresses a perfectly valid concern, down-voted).

I wonder what is the official position of the GNU project about this though.

[1]: https://news.ycombinator.com/item?id=26398251

[2]: https://news.ycombinator.com/item?id=26398538

Reply on news.ycombinator.com