I refuse to let Amazon define Rust • Hacker News Board

Re: I refuse to let Amazon define Rust

#281

Post by vjust » Mon, Sep 13, 2021, 10:17 PM UTC

nickm12 wrote: post #249

mjw1007 wrote: post #37

Earlier quoted context omitted.

The announcement blog post at https://smallcultfollowing.com/babysteps//blog/2021/09/08/ru... says the initial version was "quite closely modeled on AWS tenets".

This is a misunderstanding. Amazon and AWS use "tenets" as a tool for focusing and aligning a team on what it values. The Rust Principles are an application of that tool for the Rust project, not adopting the exact same tenets that teams at Amazon/AWS use. See this 90 second video for a quick rundown of how tenets are used. https://www.youtube.com/watch?v=5raEKA57PMY (Disclosure, I'm an Amazon employee and am familia…

AWS tenets are good for AWS. Rust has its own tenets. Both should be kept separate.

Re: I refuse to let Amazon define Rust

#282

Post by SkyMarshal » Mon, Sep 13, 2021, 10:18 PM UTC

Fwiw, as far as I'm aware, the principles and priorities outlined in the referenced article [1] were core Rust principles from the early days while it was still a project at Mozilla, rather than anything Amazon has defined recently.

They may have become more formalized over time, into a sort "Six laws of Rustbotics" sort of thing. Eg.:

- First Law: “Reliable: If it compiles, it works.”

- Second Law: “Performant: Idiomatic code runs efficiently, except where doing so conflicts with the First Law.”

etc.

The crux of it all is prioritizing reliability and deterministic elimination of undefined behaviors above all other priorities. Only a handful of languages have done that in the past (Ada, Haskell, etc), but none reached mainstream acceptance in the way that Rust has. That is something new.

I'm not aware of that particular value being closely associated with Amazon, moreso than Google or any other US Big Tech company.

If not for Steve's concerns here, I would have guessed the opposite, that Rust is defining Amazon's engineering culture and principles, than the other way around.

But it also sounds like Steve is being the canary in the coalmine here, and calling out something that may not be a huge problem right now, but could become so in the foreseeable future, and implicitly calling for governance reform similar to other communities [2] to prevent it. An ounce of prevention now is worth a pound of cure later.

[1]:https://www.infoworld.com/article/3633002/the-future-of-rust...

[2]:https://news.ycombinator.com/item?id=28514002

Re: I refuse to let Amazon define Rust

#283

Post by kbenson » Mon, Sep 13, 2021, 10:22 PM UTC

stale2002 wrote: post #264

kbenson wrote: post #261

Earlier quoted context omitted.

> Lets take the claim of "marginalized the core team"? You mean, "let's take an acillary claim, not one of the core three" that are stated to be "undefinable(sp). they're just facts."? My point, which I thought was clear, but apparently not, is that if you have a problem with the statement you brought up, sure, mention that's problematic. But is that a reason to ignore the things mentioned immediately prior, that Ama…

> But is that a reason to ignore the things mentioned immediately prior Its not ignoring! Its asking people to say what the actual problem is, beyond just that Amazon has people on a couple committees. Have these committees done anything bad? Is amazon pushing for features that people don't like? Will some future bad thing happen because of this? What is the value statement here! > Those are specific claims that can…

> Its not ignoring!

Look a the comment I originally responded to. They complained that two out of three items had no substance, therefore we can't have a serious conversation. That is, specifically, what I was addressing.

> I still don't know why it should get attention though. So they don't have an executive director? Why should anyone care?

> You keep trying to say things, without saying why anyone should care about this stuff, or why it is bad.

It's specifically stated in the tweet. Not having an executive director leaves the chair with more power. Amazon is the chair. Amazon has chosen to let the position go unfilled which results in their own position having more power.

He goes into more detail here: https://news.ycombinator.com/item?id=28513656

Actually asking questions about that, like you are here, is the outcome I was calling for, as opposed to ignoring it because of other statements, as the original comment I replied to was.

> It is on you, to both say what is happening, as well as for you to say why it is bad, and what the concern is.

No, you're placing me as someone on the one side of the argument, when the side is irrelevant. My point was that ignoring everything said because of portions that don't add up is not a valid way to assess the information. That doesn't require me to take a side, and in fact taking a side just makes it easier to people to dismiss my point and assume my goal is something else, as I suspect you did.

Re: I refuse to let Amazon define Rust

#284

Post by pjmlp » Mon, Sep 13, 2021, 10:23 PM UTC

eigenspace wrote: post #278

sk2020 wrote: post #242

Earlier quoted context omitted.

So the worst case is Amazon makes the compiler or some tool or integration a paid service? Is paying for a well-maintained compiler a problem? Ada, Java, FORTRAN, and (I think) Julia all do this, right?

The only existing implementation of Julia in case that I know of is fully open source and MIT licensed. It is well maintained and high performance :)

There are the products and tools only available via Julia Computing.

Re: I refuse to let Amazon define Rust

#285

Post by twoodfin » Mon, Sep 13, 2021, 10:23 PM UTC

throwaway17_17 wrote: post #189

rgvr wrote: post #143

Deja vu? This seems like the story of the C++ Committee all over again, isn't it? Maybe this is the price you pay for adoption and popularity..

I get what you mean, but there is a key difference and I think that difference is driving a good portion of then for/against comments in this thread and elsewhere. The entirety of the Rust Principles, etc, seems to stem from the belief, supported by Niko in the Principles blog post and others by wording, that Rust is not just a programming language, but is a ‘community’. I can find no analogue in the C++ Committee, o…

The C++ Committee story was primarily about Google and other large corporate interests advocating a large techno local change in the language to benefit their use case with no input or consideration for other users of the tool

Any links for background on this debate?

Re: I refuse to let Amazon define Rust

#286

Post by c-cube » Mon, Sep 13, 2021, 10:24 PM UTC

ajross wrote: post #148

Had to chime in on the first of the named "principles" of the language, because it makes a good example of what I think is wrong with the Rust community: > “Reliable: If it compiles, it works.” This isn't a principle. This is a subtweet. The principle being elucidated is something more like "fully specified semantics" or "no undefined behavior". And that's fine. But phrasing it like this is (1) obviously a lie as ple…

It's not just about UB, it's about the design encouraging good code and defensive programming. That includes enums, non nullability, immutability by default, etc.

Re: I refuse to let Amazon define Rust

#287

Post by pjmlp » Mon, Sep 13, 2021, 10:32 PM UTC

throwaway17_17 wrote: post #189

rgvr wrote: post #143

Deja vu? This seems like the story of the C++ Committee all over again, isn't it? Maybe this is the price you pay for adoption and popularity..

I get what you mean, but there is a key difference and I think that difference is driving a good portion of then for/against comments in this thread and elsewhere. The entirety of the Rust Principles, etc, seems to stem from the belief, supported by Niko in the Principles blog post and others by wording, that Rust is not just a programming language, but is a ‘community’. I can find no analogue in the C++ Committee, o…

When ISO C++98 came out there was no Google, or better said, it was just been founded.

Re: I refuse to let Amazon define Rust

#288

Post by stale2002 » Mon, Sep 13, 2021, 10:42 PM UTC

kbenson wrote: post #283

stale2002 wrote: post #264

Earlier quoted context omitted.

> But is that a reason to ignore the things mentioned immediately prior Its not ignoring! Its asking people to say what the actual problem is, beyond just that Amazon has people on a couple committees. Have these committees done anything bad? Is amazon pushing for features that people don't like? Will some future bad thing happen because of this? What is the value statement here! > Those are specific claims that can…

> Its not ignoring! Look a the comment I originally responded to. They complained that two out of three items had no substance, therefore we can't have a serious conversation. That is, specifically, what I was addressing. > I still don't know why it should get attention though. So they don't have an executive director? Why should anyone care? > You keep trying to say things, without saying why anyone should care abou…

> in fact taking a side just makes it easier to people to dismiss my point

The problem is that, it is pretty clear, that based on how many people have agree with me, that this guy was unclear.

There was obviously a lot of confusion here. And that is a valid criticism of this person.

> My point was that ignoring everything said

The fact that so many people are confused about what the guy is saying, is on him.

He should have communicated better. And no, it is not everyone else's fault, that basically everyone misunderstood his point.

> He goes into more detail here

Well maybe he should have done a better job of communicating his ideas, from the very beginning. That's on him.

The fact that he was so bad at saying what the problem was, and why we should care, from the very beginning, is a valid problem.

> is not a valid way to assess the information.

No, it is not on the public to have to cross reference hacker news posts, with tweets, with blogs posts, to figure out why we should care.

That is the job of the communicator. And if you communicate it poorly, then that is your failing.

Re: I refuse to let Amazon define Rust

#289

Post by type0 » Mon, Sep 13, 2021, 10:43 PM UTC

dmix wrote: post #98

ipaddr wrote: post #63

Earlier quoted context omitted.

Because he is well known many will listen. The not airing dirty laundry and not detailing the issue means no one knows what is going on and no one can help. If it's a private fight I would understand if they wanted to keep it private. If it's a public issue what is happening.

This is one of my pet peeves on social media. Nothing is worse than vague references to personal issues or something requiring inside knowledge, while the person refuses to elaborate (not saying this is true here, I haven't dug into this issue). If you're going to make it public, decide to actually make it public.

He could have expressed his concerns in a coherent blog post, that maybe could have sprawled a more serious discussion, we all know he's a good writer. Choosing to do this in a series of tweets tells just one thing - "nothing of substance here, move on".

Re: I refuse to let Amazon define Rust

#290

Post by Sindisil » Mon, Sep 13, 2021, 10:44 PM UTC

smolder wrote: post #224

xdavidliu wrote: post #193

Earlier quoted context omitted.

genuinely curious, has there been any evidence in the past of large corporations decreasing the quality of programming languages that they have influence over?

I know some people took issue with Oracle buying Sun and controlling Java, and I think it had something to do with licensing. Perhaps that doesn't qualify, though.

I was ... let's say "apprehensive" ... when I heard Oracle had purchased Sun, and thus control of Java.

However, with the exception of some communication missteps around the licensing changes (which actually ended up in fully open source JDK, as opposed to the mostly open source JDK in Sun's time), Java has actually done great under Oracle.

Amazing what having resources can do for a project, eh?

I mean, I'm not thrilled with the licensing model around GraalVM (since it's another "mostly" open source situation), but business is business, I guess. I'd personally prefer that they treat it the same as Java proper, and fund it via support contracts/subscriptions.