Why am I building a programming language in private? | Hacker News Reader

Why am I building a programming language in private?

drewdevault.com

100 pointsals069 comments

> The reason they don’t is because I asked them not to, and we maintain a mutual understanding regarding the need for privacy.

This is the only sane definition.

But otherwise I don't totally get the point of this post. "We're private for reasons XYZ, please don't share it" followed by "reach out if you want to work on it". And all we know about the language is that the goal is "simplicity".

So, ok.

If I squint a bit I could interpret it as asking for help but at the same time he throws up a barrier to keep out the people that don't have the skills or the motivation to put in the work. It sort of makes sense.

Over the past two years there have been multiple public posts on his blog about the language, you can learn a lot about it from there.

Likewise.

I'm afraid years of seeing people talk in great detail, in public, around the edges of something they "aren't ready to talk about" has led only to greater disappointment or greater cynicism.

Conversations like this in private, sure.

A straight recruitment post for contributors to a private project? Yeah.

This sort of thing? Come back when you're ready to talk about it. It's fine, we all have our own thing going on!

[Edit for weird tiredness error]

Remember the Segway?

Sure, that's the thing that was going to revolutionise transport and is now used to to play polo at corporate team-building days, right?

Not sure what you're getting at here. Is it your contention that Kamen talked about it too soon, or too late?

The Segway got a lot of pre-announcement hype. It was referred to as "It" and a lot of well-connected people were allowed to see It and talked It up (at least in quotes splashed around the news) and the hype was way too much for the actual thing. Which got a "That's it?" response from most people and the press. It was supposed to be revolutionary, as big as the PC. It was a scooter.

It was revolutionary in the self-balancing aspect.

I don't think it's the hype cycle that killed the Segway so much as the unwillingness of really any national government or major municipality to encourage people to whizz around on pavements using a powered device.

Nowadays, regrettably, that horse has bolted, and all sorts of sociopaths think foopaths are the right place for their badly-maintained future-electrical-fire on wheels!

Self-balancing may have been revolutionary, but not as revolutionary as the PC, which is how it was playing out in the news at the time (pre-announcement). Though maybe that's a matter of opinion, but I doubt most people think that a self-balancing scooter is going to change the world, which is how it was getting hyped up. The hype really did help kill it because expectations were way too high, and the cost ($5000 or so, used car prices), and then it ended up with the public image as the thing tourists and mall cops used.

> It was revolutionary in the self-balancing aspect.

What exactly did it revolutionize?

Ah yes... "Project Ginger" was promised to fully revolutionize society, bigger than the Internet, etc. People do seem to forget how old the tech-hype circus is, and with a lot of the same promises, always "too early yet" with the promised benefits lying just around the corner.

> But otherwise I don't totally get the point of this post.

Reverse psychology? "Please don't look at my language!"

A quick googling finds this: https://harelang.org/

My first impression is that it's somewhere in the trade space around Go and Zig.

> Reverse psychology

Seems to have been effective here on HN. It worked on me.

There's a social media trick where you make some really obvious mistake (eg pronounce "Mario" as "mah-RYE-oh" in a video) and then rely on people's "well akshully" in the comments causing you to go viral. Regardless if it's intentional or not, it's probably working since this already has 18 comments after only being posted a half hour ago hahah

> Cunningham's Law states "the best way to get the right answer on the internet is not to ask a question; it's to post the wrong answer."

https://meta.wikimedia.org/wiki/Cunningham%27s_Law.

I feel like Drew doesn't need to try a reverse psychology approach. He could just post it.

Anyway, I definitely have plenty of thoughts about this language but I don't think anyone working on it would care, so I'll refrain.

and it uses qbe as compiler backend, which is very interesting.

I read the first design principle and I'm already done. The programmer should the last person to trust.

I can't tell why you're being this petty. The man explicitly asked for people to respect his wish of the project being public but not advertised.

If a quick search leads anyone to find it, why not leave it at that? The ones who are interested will find it. All you did was to be be contrarian for no good reason.

And to answer parent's question with my own impression: the post serves to tell people like me (that tangentially knew the language is in development) that work is being done, and if any of us has more than a cursory interest, they can get in touch to help. It also serves to tell people that are not immediately interested in helping that spreading the word about the project is not encouraged. I can't think of why there needs to be more of a "point" than this.

There doesn't have to be a point. It just seemed really odd, so I pointed out that it was odd. I also refrained from commenting on the language as I had understood it.

So like... Drew Streisand Effect'd himself here?

Writing a big post about your secret language to a tech crowd seems like a really bad idea. People like puzzles.

Here are the available 'details' on harelang.

> Hare is a systems programming language designed to be simple, stable, and robust. Hare uses a static type system, manual memory management, and a minimal runtime. It is well-suited to writing operating systems, system tools, compilers, networking software, and other low-level, high performance tasks.

A code example:

  use io;
  export fn main() void = {
 const greetings = [
  "Hello, world!",
  "¡Hola Mundo!",
  "Γειά σου Κόσμε!",
  "Привет, мир!",
  "こんにちは世界！",
 ];
 for (let i = 0z; i < len(greetings); i += 1) {
  io::println(greetings[i]);
 };
  };

And a list of Design principles

    Trust the programmer.
    Provide tools the programmer may use when they don't trust themselves.
    Prefer explicit behavior over implicit behavior.
    A good program must be both correct and simple.

Without any elaboration on "Provide tools the programmer may use" assuming that these will relate to language concepts of constructs, there's no differentiating it from many of the well-established 'better C' languages. If there's anything we've learned it's that anything close to C in the area of "Trust the programmer" leads to broken software and we need to aim higher.

it doesn't really trust the programmer. it is even more opinioned than go, where every unused variable or function is an error. eg. it has weird identation rules, like mandatory hard tabs, and alignment demands. it uses the reverse go styles types.

but if you look at the stdlib, it looks pretty clean.

The irony is that the C creators were all aware of this, that is why they also created lint (in 1979!), but many macho devs think they know better and dealing with compiler warnings is too much trouble.

Very petty to put the website that's been up for months behind a login page: https://harelang.org/

ddevault is an extremely talented developer. this looks like a cool project. like many of his projects there's a weird undercurrent of hostility towards the greater developer community (e.g. `--my-next-gpu-will-not-be-nvidia`). i never understood the need for that, and it's a hard stop for me when choosing tooling.

I can understand it. So many tools are developer hostile, we can only take so much.

I could easily mistake that flag for being humorous or tongue-in-cheek instead of hostile.

Not if you read the author’s blog posts (or just the URL) about it

https://drewdevault.com/2017/10/26/Fuck-you-nvidia.html

This makes it pretty clear. Thanks for clarifying

> When people complain to me about the lack of Nvidia support in Sway, I get really pissed off. It is not my fucking problem to support Nvidia, it’s Nvidia’s fucking problem to support me. Even Broadcom, fucking Broadcom, supports the appropriate kernel APIs. And proprietary driver users have the gall to reward Nvidia for their behavior by giving them hundreds of dollars for their GPUs, then come to me and ask me to deal with their bullshit for free. Well, fuck you, too. Nvidia users are shitty consumers and I don’t even want them in my userbase. Choose hardware that supports your software, not the other way around.

this anger seems misplaced and dare I say, bit entitled

> ddevault is an extremely talented developer.

I fully agree with this, and that's why every single post of his about the language fills me with dread. It's a complete waste of talent, boiling the ocean, instead of simply "taking over" an existing minor language with his talent. Or just do his thing with Go, or Rust, or any of the other zillion system languages in the world.

Emotionally I want to be proven wrong, and see that language thrive, Linux/gcc-style against all odds, but rationally, I think there's only one language that can even approach C-killer levels and that's Rust. If Rust with a literal Evangelism Strike Force can't kill C, then nothing can.

I clicked through to the list of projects written in the language & was immediately taken back to 25 years ago, having an internet-facing public_html directory sitting at ~username on a web server. :)

What’s wrong with that? (Serious literal question.)

I don't see the original comment saying anything's wrong with it. It's an older way of doing things, sure, but that's not necessarily a bad thing. I actually quite like the way it's done.

> the specification is CC-BY-ND (the latter is notably less free, albeit for good reasons)

Anyone know what these "good reasons" are? I recall Stallman himself does the same thing, licensing his prose in a non modifiable format. To me it seems in screaming opposition to the idea of copyleft -- "make all the changes you want, but be sure to distribute them".

“don’t put words in my mouth”, i assume.

Sure, but nothing stops the would-be word-putter from crafting an entirely novel false narrative.

I hope this isn’t incorrect, but I recall Drew once saying he didn’t want people to make variations of the language by forking the compiler and changing a page or two of the spec, since that would result in slightly incompatible code.

So the list of 'system'(GC less) language is now (approximately chronologically): C, Ada,C++, Rust, DasBetterC, Odin, Zig, V, Jai, Hare.

That's a big list, and I'm most certainly forgetting about some language!

ATS is another very promising one

Only JOVIAL, ESPOL, NEWP, PL/I, PL/S, PL/M, BLISS, MESA, Modula-2, SOLO Pascal,...before C was known outside Bell Labs.

Pascal OK, but the other are mainly interesting from an historical perspective.. Why would you start a new project in any of these other old language?

Parent did not mention anything about being a selection for new projects.

So a historical perspective to kill the "C was first" myth is always good.

> Hare fits on a 3½" floppy disc, which are available for purchase.

Computers with a 3½" floppy drive may not be available for purchase :-)

Looks like 3.5" external floppy drives (USB) are still readily available for about $25-30.

5.25 floppy drives are a lot thinner on the ground.

I work on an open source language, none of these (e.g. jai) style of work's rationale ever really rings true to me day to day. Designing languages is hard but most of the feedback I receive on my ideas is from people who are very smart and usually right.

This feels like an "it is what it is"-style viral stunt, even though I actually suspect it's a sincere(?) post about reasons for developing without an audience.

It's gonna end up with more of an audience now, tho... :eye::mouth::eye:

couldn't they do it in the open and just ignore everybody?

edit: it seems they're doing that indeed. kudos

Then you could not write mysterious blog posts about it.

Until it’s open source it’s a proprietary project. If it makes sense for the author I guess it’s a good idea.

> Until it’s open source it’s a proprietary project.

I do not quite understand what you mean by this, are you saying that unless they accept contributions it's propietary?

It is open source.

presumably it will have enumerables instead of for-loop with iteration variables to avoid off-by-one errors… oh, wait

Don't know what to make of this - if you are doing this in private, who is exposing it to the public?

Also, you should be able to create a new language in less than two years. I've certainly done so, though I admit ones not widely adopted :-)

Edit: Genuinely interested on why the downvotes? But nevermind.

You’re almost certainly being downvoted in the first case for having snarked about somebody taking their own time for their own project. How fast you could do it isn’t really relevant, so it’s just a pointless dig.

And then in the second case because the rules ask you to meta-converse about downvotes

Where snark? I asked a question - is that not allowed here?

> Also, you should be able to create a new language in less than two years

This is the snark.

But you should!

Why isn't Zig done yet then? Or Rust?

create not done.

I can go to the websites of those, download the stable release, read tutorials and then use them for actual work, surely?

(Whether I should is another matter)

Given that Zig hasn't reached 1.0 yet, I'd say that no you cannot download a 'stable release' of Zig.

AFAIK the main reason for that is because AK wants to have a fast (debug) compiler before reaching 1.0 and that Zig isn't just a thin layer above C mainly due to its compile time feature, unsigned overflow being UB in Zig is another.

Imagine your friend says “I’ve been working on this painting for a year”.

You can say “That’s a lot of dedication, you must be really into this project” or “you should be able to paint faster than that.”

One of these is the right and answer and the other is wrong. HN doesn’t require that you hype up every poster, but if you’re going to minimize their efforts we instead ask you to just not submit that comment.

Javascript was created in 2 weeks. It then took two decades to become nearly tolerable to use. That it would cost 2 years to make a language is an absolutely arbitrary amount of time. Most great languages took significantly longer than that to become truly great.

> Edit: Genuinely interested on why the downvotes? But nevermind.

> Also, you should be able to create a new language in less than two years. I've certainly done so, though I admit ones not widely adopted :-)

This community isn't particularly fond of minimizing others' work. I'm not particularly interested in learning this language but I can still respect the effort.

> I've certainly done so, though I admit ones not widely adopted :-)

This is minimising _my_ effort.

No one cares if you minimize your own effort

They're obviously referencing the "it shouldn't take that long to create a language" bit.

Reply on news.ycombinator.com