Hello Ubuntu MATE community! I was pretty interested in this topic of replacing GNU with Rust so I would really like to hear your thoughts on the whole thing - do you think it will change anything with compatibility with our DE - I am not knowledgeable enough to understand what this change will effect so I would like if somebody could share their thoughts and information on the whole thing.
I have seen some hate towards the change and also some positive stances but I myself really can't tell nor do I really understand the whole mess around it as I am a pretty causal Ubuntu user.
I'd agree that media fuss can be quite misleading. Say, GNU (as a foundation, project, movement, community and software collection) will never be replaced with RUST (as a programming language).
Okay, now that makes more sense! I don't follow Ubuntu news, so first time hearing about it. uutils is the Rust equivalent of GNU coreutils.
Also from the Ubuntu Discourse:
GNU coreutils are pretty much the everyday terminal commands for the OS. These commands that have been around for decades. Rust is "memory safe" - meaning it's harder to trigger security bugs unlike C++.
I'm seeing that the new utilities are MIT licensed versus GPL. That licensing could end up political, but Ubuntu isn't 100% GPL anyway.
Don't really have an opinion. If it is 99%+ compatible then I guess it doesn't really matter. I just hope it doesn't lead to fragmentation within Linux, since some commands already vary ever so slightly between other Unix-like OSes like macOS and BSD.
As long as Ubuntu takes care to ensure it is a 1:1 compatible with GNU's implementation, which they appear to do with the automated test suite and treating any incompatibilities as bugs, then it should be no problem to switch defaults.
I have a mixed opinion. Re-writing things doesnt necessarily make them safer - it depends how big the original project is, how widely used and tested it is, and how the new project integrates the hard-earned lessons from the original into its development.
The ppl who started the project likely have better insight than I have, but I'm not sure they've communicated that in a way I understand.
I remember in the late 80's we were writing our code in "C" and had a bug that we couldn't figure out. Finally we disassembled the machine code in the executable and realized it was the compiler that had the bug.
The lesson: Write everything in assembler. Yes, that's a joke.
I'm only a bit worried about the fact that the Rust implementation will not be GPL licensed so it might open a door to proprietary implementations pushing everything slowly to closed source.
On the other hand, for this particular reason it might also slow down general acceptance by proponents of copyleft software.
As have been mentioned in this thread and elsewhere, the primary concern seems to focus on two points; implications, and licensing.
I share the same opinions with lah7 with regards to the implications of oxidizing Ubuntu. As though I fail to comprehend what exactly sets Rust apart in practice, I sincerely do hope it goes well.
On the licensing end, I am uncertain what to make of the comparisons between GPL and MIT, though I'm aware of what their provisions entail. I have this warm reception towards both of them and I'm welcome whichever one is applied for a software I'm using.
Nonetheless, I'm excited to see how that is bound to change things as they are for us, particularly what implications could the licensing have for this case.
Isn't that what happened with BDS and Apple? I don't know much about licensing so am not sure it is the same, but if it is about the same it was not a disaster.
Yes, it is the same situation and in Apple's case (Next+BSD+XNU) it turned out OK
but I think it's not guaranteed to always be like that.
A malevolent business could take the code, patent any logical extention to it and in that way stifle c.q. prohibit the further developement of that code.
As long as the ridiculous US patent system exists in its current form, this kind of legal monstrosities are still a possibility to happen. GPL prevents that.
Companies like for instance Oracle would surely try to do that.
As far as I know they tried to close down Open Office and MySQL which luckily backfired (almost all developers walked away and forked the projects into respectively LibreOffice and mariaDB) so they bogged down and shoved most of it under the umbrella of the Apache Software Foundation (ASF) but I'm not convinced that we will always be so lucky.
Thanks for the explanation. I am pretty much just a causal user, but it always interests me, and not just Oracle, but Microsoft worries me even more with their: Embrace, extend, and extinguish.
Well the point stands but again I don't think Canonical, as a company registered on Isle of Man, would use same tricks as some US based companies do but maybe I am being too naive in my euro-centric views and I do believe that we shouldn't look at every company model through a US lens since the business culture isn't the same - maybe Canonical deserves a bit more benefit of the doubt than it is given
I recently found a few years old ideas by the founder of Canonical:
At aprox. 2min mark the ideas are really on point about importance of open source
Also this short piece of an interview as well: https://www.youtube.com/watch?v=JKzpdZFB3Zw - this one adds a bit more on the ideas about importance of open source and so when one listens to the ideas of the man who started the whole thing it doesn't seem like jumping to Rust is intended to make it closed source - at least one can hope
Thanks for the links Alarik.
To be clear, It's not Canonical that I fear.
The good part of making these core tools under a MIT or BSD style license is that it brings potentially more security for every platform which in the end is better for us all.
I agree. I don't think a certain company started a Linux subsystem because they love Linux, but did it because they coveted the developers that prefer Linux.
Most of this is a bit beyond my Understanding but I love the way the author sums in upt at the end:
Stick it all up in your bygones Microsoft. You’re a liar and a thief and the only reason most of your higher execs aren’t in prison is that U.S. law and your good ol’ boy network protected you. You aren’t fooling anyone. You don’t love Linux any more than I love liver and onions. You have merely realized that the only way you are going to survive into the next decade is to integrate Linux into your strategies…and integrate it deeply. Let’s face it. You need us. More than we need you. - by Jim Lynch
As far as I'm concerned, as I also expressed in my comment on the Ubuntu forum, the real problem is the MIT license, which is permissive and would allow corporations to take over the code created by the community and develop improved versions of the programs without sharing them back with everyone else. The MIT license is suitable for certain projects funded by industry, but not for such fundamental components in Unix-like operating systems. The success of Linux is largely based on the fact that anyone who modifies the code must then share it back with the community (that's a requirement of the GPL license). Thanks to this mechanism, the industry is obliged to give back, and that's why we've seen constant improvements to the kernel over all these years. If you look at similar projects with different licenses (see BSD or Minix), the results are quite different.
Moreover, even though I don't have any particular objections to modernizing the code of a program using more advanced and trendy programming languages, I don't see the point of completely rewriting well-refined and extensively tested basic programs, especially when they form the foundation of the operating system and are used by virtually all other programs. Rewriting unnecessarily introduces the risk of incompatibility. For these reasons, I find the initiative rather pointless, risky, and harmful in the long term for the GNU/Linux ecosystem. That said, I realize that no one cares about the opinion of someone who doesn't work at Canonical, so I have no illusions that the Ubuntu team will truly listen to our concerns. They will go their own way, as they have done with all other Canonical projects so far.
I think it's unlikely that Ubuntu would completely replace all GNU utilities in the near future. A more realistic scenario would be a gradual migration, starting with the utilities that are most prone to security vulnerabilities or have the most potential for performance improvements. There's already work in this direction with projects like ripgrep (a Rust replacement for grep) and exa (a Rust replacement for ls).
