Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Verilog2factorio

github.com

Post reply on HN
Print view View original

61–70 of 76 posts

Re: Show HN: Verilog2factorio

#61

Post by devit »

jakear wrote: post #42
userbinator wrote: post #36

Earlier quoted context omitted.

This feels a bit like that classic estimation interview question... Verilog originated in 1984 and educational institutions have been teaching it roughly as long, multiply by how many students graduate each year in an EE-related field and I'm pretty sure the number is far more than that in India alone. Of course, the number of people who know Verilog (and Factorio) enough to implement something like this is going to…

Sure, but there’s a difference between “had a couple courses using it in college” and “have an understanding of the language sufficient to develop a novel backend”.

You don't need much understanding in advance to develop a novel backend, since you are forced and guided to learn the whole language as you try to write the backend and start pattern matching on the data structures.

Re: Show HN: Verilog2factorio

#62

Post by emteycz »

interrealmedium wrote: post #40

That's it, I am done. No more serious IT work for me. I am going to become a gamers consultant. If there are people who care enough to make things like this, there must be people who would pay decent money for "professionals" to come in and optimize their factorio game, or their minecraft builds, or to make them feel good about their solution to a Zachtronics puzzle.

The key is in streaming your work.

Re: Show HN: Verilog2factorio

#65

Post by MisterBiggs »

codetrotter wrote: post #52
MisterBiggs wrote: post #5

Always thought programmatically generating Factorio blueprints was a really cool idea and Verilog seems like a great language to model blueprints, but this repo definitely needs more work before being shared with the world. One of my favorite tools that shows what's possible: https://autotorio.com/example/outpost

> this repo definitely needs more work before being shared with the world No, don't say that :( I see way too many people that don't want to share their code because they don't think it's tidy enough or complete enough and that they just need to clean it up or adding some more features. But the thing is that they never get around to cleaning it up or adding those features and therefore they never share it. And often…

The repo totally stands on its own now but when I made the comment it was only a few sample files and a readme with a few lines on how to install the project without any mention of what the project actually was besides its name. Now with a proper readme and some pictures I think it's much better.

Re: Show HN: Verilog2factorio

#66

Post by themulticaster »

jakear wrote: post #42
userbinator wrote: post #36

Earlier quoted context omitted.

This feels a bit like that classic estimation interview question... Verilog originated in 1984 and educational institutions have been teaching it roughly as long, multiply by how many students graduate each year in an EE-related field and I'm pretty sure the number is far more than that in India alone. Of course, the number of people who know Verilog (and Factorio) enough to implement something like this is going to…

Sure, but there’s a difference between “had a couple courses using it in college” and “have an understanding of the language sufficient to develop a novel backend”.

This project actually uses Yosys. From a quick glance at the code, this tool essentially uses Yosys for the heavy-duty logic stuff [1], resulting in a JSON-based logic definition using a couple of standard cells (DFF, MUX, arithmetic, etc.) that verilog2factorio parses [2] and translates into predefined Factorio blueprints. The only part that I couldn't find would be laying out the cells into the overall blueprint (and performing routing).

[1] https://github.com/Redcrafter/verilog2factorio/blob/9691b702...

[2] https://github.com/Redcrafter/verilog2factorio/blob/9691b702...

Re: Show HN: Verilog2factorio

#67

Post by linux_is_nice »

themulticaster wrote: post #66
jakear wrote: post #42

Earlier quoted context omitted.

Sure, but there’s a difference between “had a couple courses using it in college” and “have an understanding of the language sufficient to develop a novel backend”.

This project actually uses Yosys. From a quick glance at the code, this tool essentially uses Yosys for the heavy-duty logic stuff [1], resulting in a JSON-based logic definition using a couple of standard cells (DFF, MUX, arithmetic, etc.) that verilog2factorio parses [2] and translates into predefined Factorio blueprints. The only part that I couldn't find would be laying out the cells into the overall blueprint (a…

It does simulated annealing in src/sim.ts, I think.

Re: Show HN: Verilog2factorio

#68

Post by themulticaster »

harry8 wrote: post #54
userbinator wrote: post #36

Earlier quoted context omitted.

This feels a bit like that classic estimation interview question... Verilog originated in 1984 and educational institutions have been teaching it roughly as long, multiply by how many students graduate each year in an EE-related field and I'm pretty sure the number is far more than that in India alone. Of course, the number of people who know Verilog (and Factorio) enough to implement something like this is going to…

Some saw/earnt vhdl someone after it emerged. ~half? Verilog had a popularity ramp up time of a few years too. I'm going with a much smaller number than you there. Could be way wrong!

I think the apparent increase in Verilog popularity is due to pretty much all FOSS toolchains (i.e. mostly Yosys) only supporting Verilog [1], so most of the content publicly visible on blogs etc. uses Verilog.

And I guess the reason that most tooling focuses on Verilog is that the VHDL standard is a little bit of a mess from a parsing point of view. Essentially, in order to correctly parse some VHDL file A you first need to parse every other file that A might depend on. [2] Source: Tried writing a VHDL parser a few years ago.

[1] AFAIK you can buy a license for a VHDL frontend from one of the companies contributing to Yosys, but that's a different story.

[2] Example: What is f(1)? It could be a) calling a function called f with a parameter_association_list consisting of 1 or b) an indexed name, accessing an array called f at index 1, or c) some other things if we're not currently in an expression context. The only way to parse this correctly is to know what f is, but if f is not defined in the file you're currently parsing, there's no way to know other than parsing all other files...

Re: Show HN: Verilog2factorio

#69

Post by meristohm »

interrealmedium wrote: post #40

That's it, I am done. No more serious IT work for me. I am going to become a gamers consultant. If there are people who care enough to make things like this, there must be people who would pay decent money for "professionals" to come in and optimize their factorio game, or their minecraft builds, or to make them feel good about their solution to a Zachtronics puzzle.

I’m guessing you’re mostly not serious, though I acknowledge there may be a market for what you describe.

As a lifelong gamer since the Commodore 64 (since parenthood scaling it way back to mostly crewing a tank in Hell Let Loose with two friends, or a ttrpg with same), looking back it feels like I took refuge in a medium (single-player games) ultimately less valuable than other-format stories, like books, film, and radioplays. What I’m driving at is a feeling that I’ve enjoyed elevated layers of luxury that may be an indicator of too much idle time for too many people in US culture. Is this a common mindshift with age?

Escape has a place, but a whole industry built around spending so much time in abstraction layers feels like a problem if it’s not helping us relate to each other better.

Re: Show HN: Verilog2factorio

#70

Post by scrollaway »

meristohm wrote: post #69
interrealmedium wrote: post #40

That's it, I am done. No more serious IT work for me. I am going to become a gamers consultant. If there are people who care enough to make things like this, there must be people who would pay decent money for "professionals" to come in and optimize their factorio game, or their minecraft builds, or to make them feel good about their solution to a Zachtronics puzzle.

I’m guessing you’re mostly not serious, though I acknowledge there may be a market for what you describe. As a lifelong gamer since the Commodore 64 (since parenthood scaling it way back to mostly crewing a tank in Hell Let Loose with two friends, or a ttrpg with same), looking back it feels like I took refuge in a medium (single-player games) ultimately less valuable than other-format stories, like books, film, and…

Video games are just a younger medium than books and movies. Even TV needed more time to be seen at a similar level to movies in terms of art.

We're getting there though. The past 15 years have given us many masterpieces of gaming that can absolutely be considered art, and they're becoming more and more common. As this happens more and more, the medium is taken more and more seriously.

Many games still suffer from feature-creep by way of trying to please users demanding "content". This incidentally make them feel like the primary goal is to waste time rather than to experience a high quality story. In other words, there are still too many "Big Bang Theory" and not enough "Breaking Bad" so that is why it still feels less valuable.

Because I told you I just binged The Queen's Gambit, you wouldn't look at me like I wasted my time, you'd ask me about the story. If I told you I just binged TBBT, you'd just think I'm nuts. (No offense if you like TBBT, it was a decent sitcom for a couple seasons, but … hey, you do you).

For example, I just finished Control. A wonderful, high quality game. Great storytelling, great mechanics, and … a crapton of completely useless busywork that is just there because it's expected to be there. A talent tree, repeat random side-missions, a crafting system, etc. None of this is useful to the core game. It's a shame because I believe if it had done less in terms of busywork mechanics and maybe focused a tiny bit more on puzzles, that game could have been a masterpiece of the same caliber as Portal and Portal 2. (Can you imagine if Portal 2 had a talent tree and a crafting system?)

Instead, it's a Season 7 of Game of Thrones. Great production value, but so hit and miss in terms of final quality because of avoidable mistakes. Could have been so good, you know?

Post reply on HN