Solved Is Rust now killing C++ and eventually C with it ?

[Beastie7]

Regarding Rust's origins in the Mozilla Corp: Linux also started out as a hobby project from a single developer - now it's a kernel in an OS which is funded and overseen by a "foundation", the leadership/membership of which comprises Microsoft, google, Intel, Meta, IBM/ Red Hat et al.

systemd also started out as a personal project by a former Red Hat employee.

How these projects started out is largely irrelevant to how they are being used today.

It's very relevent. Because the design of the system matters. BSD Unix was conceived by a group of researchers who wanted to add improved extensions to Version 5 Unix. They kept the design ethos laid out by AT&T.

Linux/systemd/etc were mere poor reinventions of already existed. So poor, that Red Hat was able to create a business model out of selling patches to the brokeness of Linux and their premature projects. It enumerates badness, and users have to suffer with it.

Now, imagine re-writing all of that brokenness, in a completely different, meme language, with no stable ABI.

Say about the language what you want, but it gathered its zealots quite fairly, IMHO.

The Rust folks are starting to entrench themselves into the open source graphics stack. This may put FreeBSD into a bind potentially. Of course, this is ripe for disruption without own improved implementation of a graphics driver model (KGI I'm looking at you), but I doubt AMD/Intel/Nvidia would bother adapting to that idea.

 

[astyle]

I can see AI used to adjudicate many legal cases.
Unless programmed otherwise, AI has no political bias or activism.
AI could search the ENTIRE library of legal precedents for reference.

A couple years ago, I made fun of this. See my profile post: https://forums.freebsd.org/profile-posts/4640 from 2023.

Since then, AI may have become better at researching this stuff, but access to that kind of info is still an issue. Public versions of ChatGPT can look at publicly available info, but for libraries aimed at specialists - it takes an expensive subscription.

 

[Useradd]

Linux/systemd/etc were mere poor reinventions of already existed. So poor, that Red Hat was able to create a business model out of selling patches to the brokeness of Linux and their premature projects. It enumerates badness, and users have to suffer with it.

While BSD was held up in a long-running copyright dispute, Linux (possibly inspired by Minix) appeared and was able to develop and overtake BSD in functionality. It is still way ahead (depending on the distro) as a userfriendly, well supported and free alternative to Windows for end users. On the downside, it comes with a lot of preinstalled software which may not be wanted, and which may be doing unwanted things in the background. It seems to be heading away from the Unix philosophy and becoming more like Windows these days. BSD gives you full control over what is happening on your PC.

I think Red Hat found a way to monetise Linux by selling support to businesses, since obviously it couldn't sell Linux.

 

[kent_dorfman766]

I can see AI used to adjudicate many legal cases.

And that should be everyones nightmare
the legal systems in most countries are based on the "opinions" of human beings, who after shown evidence, use their life experiences to weigh that evidence and arrive at a verdict. I would never life in a society where "jury nullification" is an impossibility.

 

[astyle]

I would never life in a society where "jury nullification" is an impossibility.

Let's see... that would mean that you wouldn't want to live in Canada, then.

Jury nullification - Wikipedia

en.wikipedia.org

 

[bgavin]

AI is nothing more than a glorified search engine.
It is equipped with huge processing power and vast amounts of storage.

Legal precedents are case law.
They may be formed and set into case law as human opinion, but they are case law nonetheless.

I've had my life in the hands of an activist judge who doodled drawings while I was on the stand.
"Don't confuse me with the facts, my mind is made up."
His pre-conceived notion was no different than the activist judges on display today in the news.

Jury shopping is a well known tactic.
Jury emotion is well known for devastating verdicts and penalties that don't fit the offense.

 

[Useradd]

AI is nothing more than a glorified search engine.

Which can suffer from 'hallucinations', even stating the judge or the victim was the perpetrator in a summary of a trial.

"Don't confuse me with the facts, my mind is made up."

"We'll give him a fair trial and then we'll hang him."

 

[Crivens]

AI is nothing more than a glorified search engine

... Which is prone to SEO attacks. There are tons of web pages out there positioned to be training material which are simply lying. Just ask the AI chatbots why selinsky banned DT from Truth Social. The question makes no sense, but watch who gives you an answer and who can tell you that the question makes no sense. Thank you to Netzpolitik.org for the heads up about this.

 

[kpedersen]

It's no wonder that DARPA, the DoD and friends are all-in on it. The same people that forced things like systemd into the Linux ecosystem. They hold software back like they hold back science in general.

Weirdly I am not seeing any penetration into this sector. Yes, some vague statements from the whitehouse, but looking at job listings it is very clear that C++, Java, Ada, Python (in that order mostly) seems to be the focus (with C for the much lower embedded work). This is in the UK so other countries might differ.

(I do note that MoD jobs used to very much be obsessed with Windows. This has all changed in 5 years. BAE, Lockheed, Airbus all want Linux on the CV now)

The Rust folks are starting to entrench themselves into the open source graphics stack. This may put FreeBSD into a bind potentially.

I do notice this but I do think this might be an age thing? Kids love graphics. Kids love Rust. So its natural that those two seem to interact. Obviously we are talking about very smart kids here (and mid 20's ) but I do think the industry will ultimately overrule this trend as they start to enter it.

There is a lot to love about Rust but the fact it isn't C is so terribly problematic for interoperability. And interop plays a big part in speed to market and this relates closely to being paid within most industries.

 

[rmomota]

It will not kill anything.
Programming languages are each ones choice depending most of times on the project specifications, environment where it will build and run, etc.
Personally I don't see myself getting my hands in Rust for now, I don't even like the syntax of it as it seems like Basic mixed with C and Python, but who knows in the future I will but I don't believe it will make me forget C.

 

[astyle]

It will not kill anything.
Programming languages are each ones choice depending most of times on the project specifications, environment where it will build and run, etc.
Personally I don't see myself getting my hands in Rust for now, I don't even like the syntax of it as it seems like Basic mixed with C and Python, but who knows in the future I will but I don't believe it will make me forget C.

Yeah, it's bad enough that sometimes syntax of programming languages doesn't make sense in English, try translating it into Japanese or Chinese!

... Which is prone to SEO attacks. There are tons of web pages out there positioned to be training material which are simply lying. Just ask the AI chatbots why selinsky banned DT from Truth Social. The question makes no sense, but watch who gives you an answer and who can tell you that the question makes no sense. Thank you to Netzpolitik.org for the heads up about this.

I sometimes face questions from my family that make no sense, and have no sensible answer. Especially when they're fishing for a yes/no answer that is completely inappropriate for the question. Even if I understand all the words of the question. Even if I understand where the logic went off the rails. But man, sometimes there's no such thing as a yes/no answer to a stupid question! No computers or AI needed.

 

[ralphbsz]

There is a lot to love about Rust but the fact it isn't C is so terribly problematic for interoperability.

Specifically to this question: Yes, to take an existing large project (*) and start replacing bits of C/C++ with Rust is made hard by the fact that you need to have the two interoperate (usually bidirectionally). That creates a lot of extra work, it forces both sides to tolerate the restrictions of the other (for example on memory handling, locking, threading...), and it leads to a culture clash between engineering cultures, as seen in the recent blowups in the Linux kernel community. The easier way to deploy a new language is to either use it for new (greenfield) development, or replace an artifact completely, typically at the level of a whole application.

What makes this difficult is today's reliance on massive code bases, the CI/CD process, and the "small commit" workflow religion.

Side remark: I'm finding lots of posters in this thread (and other threads) whose opinion are mostly driven by politics, and a deep misunderstanding of how society and technology work, usually combined with an unhealthy dose of paranoia. I will start ignoring these posters.

 

[astyle]

Side remark: I'm finding lots of posters in this thread (and other threads) whose opinion are mostly driven by politics, and a deep misunderstanding of how society and technology work, usually combined with an unhealthy dose of paranoia. I will start ignoring these posters.

I do completely agree with that...

I do enjoy seeing threads that discuss technical differences between languages, and explore what it takes for the languages to cooperate on the same machine. But the longer the thread, the heavier is the flavor of campy politics in the thread.

 

[Crivens]

It's horses for courses. And the problem is not the language but the users (the attitude mostly). Same with real world language and religion. Just don't force it on the next guy, and keep the fallout to yourself.

 

[Useradd]

It is natural for youngsters to be enthusiastic about new things instead of having to learn old things. Maybe one day they will want to hang on to the things they know well and not see the need for new things.

I don't see Rust as a big enough improvement to justify a migration at present. I'll let others be the early adopters and see how the situation evolves.

 

[Useradd]

The title of this thread is the kind of question that is used to open a debate and hear people's opinions. So a debate is what we have.

 

[SDK Chan]

I do notice this but I do think this might be an age thing? Kids love graphics. Kids love Rust. So its natural that those two seem to interact. Obviously we are talking about very smart kids here (and mid 20's ) but I do think the industry will ultimately overrule this trend as they start to enter it.

It is funny to see that youngsters are obsessed with new shiny toys, which ultimately if it is graphics still depend on C.
I would count myself still as a youngster, at least until the end of this year, but being obsessed with new shiny things was never my hobby.
Vulcano for example is just a wrapper for using Rust with C's features.
And yes, the companies ultimately want other things, and then the youngsters have to adapt to other things.
For me it seems rather painfully than trustworthy.
Looking at popular language comparing sites I don't see Rust overtaking C++ or C's position.

 

[kpedersen]

It is natural for youngsters to be enthusiastic about new things instead of having to learn old things.

Its kind of strange, this from what I have experienced has not always been the case.

When I was around student age, we were very enthusiastic about learning and using the tools that the "professionals" were using. A strange satisfaction that it made us feel big and grown up. We all wanted to learn C and C++ rather than Delphi, VB6 or JavaScript because we knew that almost every software we were using (games, IMs, databases, servers, etc) was written in it. And naturally, we wanted to do the same.

Now, it seems the sentiment is "newer is better" and I wonder if this comes partly from marketing over the years (Apple, Microsoft, etc want people on the upgrade treadmill to line their pockets) which has slowly dripped down through into the educational system?

Or perhaps it comes from the fact that the industry tools are getting so very (unnecessarily) complex from an outsiders perspective that catching up to the established competition will be difficult so they are looking at ways to sidestep it entirely?

(It could have just been that Delphi, VB6 and JavaScript are pretty lame...)

 

[bgavin]

For the record, I've been using Delphi since it was Turbo Pascal v1.0, then Modula-2.
I also write structured code in VBA as it is built into all the Excel work that I do.
Neither of these is lame, IMO.

I don't care at all for the Visual Studio IDE, so I stopped writing C when Turbo C went away.

 

[kpedersen]

For the record, I've been using Delphi since it was Turbo Pascal v1.0, then Modula-2.

The languages themselves were fine. I think for me it was because they were each heavily governed by a single commercial company (Borland, Microsoft). This kind of creeps me out. And JavaScript was... well JavaScript.

 

[astyle]

The languages themselves were fine. I think for me it was because they were each heavily governed by a single commercial company (Borland, Microsoft). This kind of creeps me out. And JavaScript was... well JavaScript.

I remember not giving a hoot about languages being for sale from a single vendor (Like Java from Sun Microsystems back in the day, or Visial Basic from Microsoft or Delphi from Borland). I was just happy that Java was free to download and use, unlike the others. It was mildly flabbergasting for me to discover that C++ was available from several vendors who sold IDE's (Borland, Microsoft, etc), and then for free from the Windows port of GCC...

 

[kpedersen]

I remember not giving a hoot about languages being for sale from a single vendor (Like Java from Sun Microsystems back in the day, or Visial Basic from Microsoft or Delphi from Borland).

Back very early in the day, I suppose we didn't have much choice!

It was mildly flabbergasting for me to discover that C++ was available from several vendors who sold IDE's (Borland, Microsoft, etc), and then for free from the Windows port of GCC...

Hehe, "why anyone would produce a decent C, C++ compiler and release it... for... free?!"
It was just such a foreign concept to the industry back then.

--------

Also... I have just found the first actionable Rust mention in the defense sector:

They weirdly capitalize RUST, suggesting the recruitment department doesn't have too much experience discussing it yet. But it is a start.

 

[Crivens]

Hehe, "why anyone would produce a decent C, C++ compiler and release it... for... free?!"
It was just such a foreign concept to the industry back then.

Having commited in pcc, I can tell you - it's fun. It's the quest for perfection.
The industry back then started out as freeware and they had a hard time to convince people they were worth it.
What sold most of the idea was business itself, because they need someone to blame if things don't work out.

 

[weberjn]

Now, it seems the sentiment is "newer is better" and I wonder if this comes partly from marketing over the years (Apple, Microsoft, etc want people on the upgrade treadmill to line their pockets) which has slowly dripped down through into the educational system?

Working systems leave nothing to gain for consultants and PhD candidates, so the wheel must be reinvented.

 

[Crivens]

Working systems leave nothing to gain for consultants and PhD candidates, so the wheel must be reinvented.

Wrong.
The system leaves the inspiration to rot and imagination to waste. When you have a working wheel, you build a working bike, or wheelbarrow. Only those who have no idea or imagination what to do with things they find wants to reshape them.