Hacker News Reader

Hacker News

Comment by ChrisSD | Hacker News Reader

Parent Full thread

ChrisSD

Announcements also from Mozilla, Amazon and Microsoft:

https://blog.mozilla.org/blog/2021/02/08/mozilla-welcomes-th...

https://aws.amazon.com/blogs/opensource/congratulations-rust...

https://cloudblogs.microsoft.com/opensource/2021/02/08/micro...

View on HN

I‘d love to see Apple joining. They already have the hand on LLVM and support things indirectly, but...

There is no reason that serious things written in Rust can‘t coexist with app-y stuff in swift.

Seriously, I think we all would benefit enormously from some consolidation, especially on such a great foundation like Rust. Think investments like: compiler-optimizations like V8, portability like C, enterprise grade tooling like Java, productive libraries like Ruby, datascience libs like python, ... it would be great to have all of this in one coherent package.

Why? Instead of glueing together brittle and fragile things where the moving parts break within weeks, reinventing the wheel, we could finally focus on the actual things to build. Bonuspoints for Rust: every other techstack can leverage/profit directly, like when binding C stuff.

Everyone still can innovate like the past decades, languages and libs will be re-explored and reinvented and so on, but, really, a solid baseline would be a dream

Pretty much everything you listed is stuff Apple is already doing with Swift, and although Rust is awesome for lower level stuff, and stretches up quite well, Swift really does cover more of the stack of abstraction.

That said, I’d say Apple having Swift is part of the reason everyone else is getting involved with Rust.

Rust is credible competition for Swift where nothing else is.

I have not used Swift much but the use cases overlap more with C# than Rust IMO.

Having used Swift pretty extensively, I’d say it overlaps both.

How does it overlap with Rust where it does not with C#?

Hi, as someone interested in typography I am interested in how and why you wrote a ‘ (U+2018 LEFT SINGLE QUOTATION MARK) instead of a ' (U+0027 APOSTROPHE) or a ` (U+0060 GRAVE ACCENT)? Is it used by a language on your system? is it in your keyboard? Or maybe a conscious typographical statement of proper style? I wonder :)

Personally I have had my dose of fighting with using multiple languages with incompatible keyboards layouts, I would like to hear your story (even if it has nothing to do with keyboard layouts)

I find a compose key (see Wikipedia) quite helpful for diacritics, em dashes, &c.—once you get the hang of it, it’s really quite fast and intuitive. One can even add new combinations.

I recently left the linux world and with it the compose key... the most similar (native) thing windows offers are dead keys which I find deeply abhorring.

I settled on switching keyboard layouts with Windows+space.

I was more curious about the why than the how in a sense.

I use Windows, and haven't used Linux for years. It's probably just a matter of time before I go back, and when I do I know I'll love the compose key.

However, the workaround I use on Windows is Alt key combinations: I memorized the dozen or so patterns [1] I use the most. For example, `Alt + 0160` for non-breaking space, `Alt + 0150` for en dash, `Alt + 26` (and a few more) for arrows and `Alt + 0133` for ellipsis. It must seem crazy when you're used to the compose key.

I haven't heard of dead keys as a workaround. How does it compare? (Wow, this is getting very far from the original topic…)

[1] An important key to memorizing them is to see them as patterns on the numpad, rather than a sequence of digits.

dead keys are an hack that implement a feature of old typewriters. On a computer it simply disables the keys for accents and combine the accent with the next letter, the catch is that if you want to write the accent character (',`,^," etc.) you have to press the key twice, as the first press does nothing.

We don't need monocultures.

If everyone uses the same language, they're all going to write the same bugs! /s

Well, each language has a different tradeoff in terms of usability, performance and model.

Like Brainfuck could program anything.

Then Swift should be replaced by something more mainstream. Adding Rust would be less monoculture than what we have now.

Apple is using Rust somehow internally, they were advertising for network software developers with Rust experience a while back.

And Google:

https://opensource.googleblog.com/2021/02/google-joins-rust-...

Why make the Rust Foundation be its own corporate entity, instead of a project within another major non-profit like the Software Freedom Conservancy?

(For those not familiar, SFC is the home of Git, Godot Engine, Homebrew, Inkscape, QEMU, and Wine.)

My worry would be that this is happening because Amazon, Google, and Microsoft want control over the foundation. (They are writing the checks, so no surprise they'd want to control how that money gets invested…)

I actually feel almost the opposite. Umbrellas are a good thing, but if you have the resources to go your own way, that makes more sense. So if your vision is that in five years nobody cares about Rust, these corporations have lost interest, the few dozen people still working on it need somebody to do paperwork and finances, then yeah, an umbrella is great. If your vision is that in five years Rust is ubiquitous and we are arguing about someone having too much control over the foundation an umbrella is exactly not what you want.

An example of a project I think would be less healthy under an umbrella would be LibreOffice. An example of a project I think would be more healthy under an umbrella would be the Haiku Operating System.

You largely cannot fight those big cheques, if you say "We don't want any corporate influence, go away", you lose, Amazon, Google and Microsoft do go away and build I dunno, "Titanium", a language eerily similar to Rust and with lots of manpower and money behind it.

You can put some structures in place to reduce the worst influences of money, you obviously don't want a situation where one day everybody senior at the Rust Foundation wouldn't know an IEEE float from a UTF-16 code unit and the language is allowed to decay and fail while executives fill their pockets from the dwindling coffers. But there are practical limits, at some point you must trust in the basic goodness of your fellow humans.

Our industry hasn't really been around long enough for the worst rot to set in, but there's a Vox story (I think?) years back about the problem of zombie foundations, with a set mission that ceased to make sense in the modern world. There's plenty of money but the intended charitable purpose no longer makes sense. A free primary school for the "disadvantaged" children in a long-since gentrified inner city region where the average child today is a millionaire is the sort of problem that arises.

Titania, to follow the oxide metaphor. Which is honestly a pretty darn decent name.

They have an unhelpfully generic answer to that in their FAQ: "After spending a significant amount of time researching potential umbrella organizations, we decided that our best option was to incorporate an independent entity. Rust is a technology and community that is value driven and we simply didn’t find an organization that we felt was aligned with our community goals. This does mean more work for us, especially upfront, but we think the tradeoff is worth it."

https://github.com/rust-lang/foundation-faq-2020/blob/main/F...

As they are registering as a 501(c)6, I would assume that tax reasons would have impacting their decision not to join the SFC.

> My worry would be that this is happening because Amazon, Google, and Microsoft want control over the foundation. (They are writing the checks, so no surprise they'd want to control how that money gets invested…)

I find this attitude surprising. They are paying for development, why they shouldn't have authority to dictate the priorities? If it was SFC, some other interests would be dictating the priorities anyway.

> They are paying for development, why they shouldn't have authority to dictate the priorities?

If you are sponsoring a foundation and can dictate the direction you aren't a sponsor but an employer and owner. Then the foundation is just a way to make good looking PR or get tax-cuts.

Companies shouldn't be able to control other people's projects just because they pay developers to work on it. The project's vision can't be corrupted to serve any one company's interests. I'd even say this should also hold true for every company's own products.

Realistically, it all comes down to leverage and power. Who needs the other the most?

As I mentioned below, the way to get good answers here is to contact them directly; nobody from the foundation is in this thread.

I find it puzzling that aws uses rust so much yet don't have an official rust aws sdk

That metric doesn't really matter for that purpose. If AWS gets enough customer requests to warrant building a rust SDK, they'll do it. It doesn't have much to do with internal rust adoption at AWS.

I don't think any of these companies use Rust much. You may find teams here and there using it but the vast majority of the Rust type stuff is written in C++ or Java.

It depends on what you mean by "much." If you mean "as a volume of LOC of total software development," absolutely. But if you mean "not being used for anything big" or "not being used for anything important," that is very much wrong.

In what concerns Microsoft, I am yet to see anything besides Win32 and WinRT bindings generation.

UWP/WinUI foundations are in C++ and .NET.

Azure Sphere, despite all the security selling marketing uses C based SDK.

Windows kernel seemed to be in the right direction adopting more C++ and less C, specially with the userspace drivers COM SDK, but they just reverted to more C support on the latest versions.

So in what concerns Microsoft, I look forward to their Rust's adoption, although currently its adoption is yet to be reflected on the Windows development ecosystem.

Firecracker is written in rust and is from Amazon.

Firecracker being a modified fork of Google’s crosvm project.

Not really. They share some dependencies, and there's quite a bit of desire to work on those dependencies together (via the rust-vmm project).

It literally forked from crosvm.

While not official, there is at least one or two Amazonians working on https://github.com/rusoto/rusoto which is a pretty great Rust AWS SDK.

Just wanted to point out in case you didn't see it, that rusuto currently needs a fulltime maintainer! They're in maintenance mode currently, no one is spending much time on it since the primary maintainer stepped back.

An AWS tech lead mentioned that they were working on an official Rust SDK a while back

Reply on news.ycombinator.com