Pretty sure software for waste treatment facilities is much better than another spa garbage full of spyware, that ideally should go straight to those facilities.

Waste treatment is a form of industrial infrastructure. Infrastructure lasts decades. What I've always wondered about is whether there is a hardware/software platform designed to last forever. I don't know much about IBM Mainframes but from what little I have heard they do provide such a platform.

The question is whether we can build such a platform with opensource tooling and make it scale down to microcontrollers (KB of RAM), medium sized MCUs (MB of RAM) and SBCs (100s of MBs to 1 GB of RAM). The support burden is very high so whatever distro/RTOS would have to make it possible to run every binary on any architecture without recompilation.

Lots of people say that the JVM is that platform but the garbage collector makes it insufficient in my opinion. We also need the ability to run existing C code to reduce the porting burden. I think WASM could fit the bill.

If someone doesn't call this cool then I don't know. Of course in practice nobody cares and people just dig out ancient C compilers.

You just don't change anything, and that's how you last forever, like COBOL.

c++ is the standard for most industrial robotics and microcontrollers in my experience, so I'd say that's your answer. Most of the ones that need something more are running that code on some form of linux so you can ssh into them for debugging.

A growing portion are using ROS with their linux

People want to work in the video game industry so they are willing to take less pay in order to work in it.

Too true. The "teenager who likes to play a lot of video games and therefore enrolls in a compsci undergrad program" is an extremely common pipeline.

Not too many teens choose a career in programming because they want to build a robust framework that allows construction of large-scale analytical queries on unstructured data.

Yeah, writing boring HR software pays a whole lot better than game programming because one brings you joy and the other makes you a nihilist.

Based on what I've read about working in the games industry, joy is fleeting between ridiculous crunches, which seem to be nearly ubiquitous.

IMHO, neither is the most interesting work. There are some really interesting jobs writing software for machines with life-and-death consequences. Imagine knowing that your software is implanted in a human, or that we rely on it to prevent a huge pipeline from exploding.

> I find video game programming to be the most difficult of all, just because you have to account for so many things

Can you give some examples? How many things does a VG programming have to account for compared to, for examples, a ML programmer, or a programmer in the HFT space?

Sure, video game programming is not a game, but I don't know why it would be "the most difficult" compared to other fields.

A couple good comments here, but I'll add my $0.02.

Video game programmers often need to be very deep generalists. It's a combination of 1) the reality of small teams working on ambitious games, 2) increased expectations for what a videogame can do by default ("what do you mean I can't join a party with my friends?!"), and 3) video games are deeply, deeply bespoke software, and to implement one feature means you need to know the limitations of many subsystems.

A good friend of mine works on the World of Warcraft gameplay development team. They have ~50 engineers, and they're almost all full-stack to some degree. He said that when he was an intern on the team, to build a feature he needed working knowledge of:

1. Blizzard's battle.net network layer

2. The way textures were rendered onto character models, pixel-by-pixel

3. The intricacies of their in-house Lua engine and parser

4. The way that NPC AI was written

None of these requirements, by themselves, are too crazy; but he was using all of these, every day, as a junior in college!

And that was for a three-sprint project! As an intern!

I had a chat on the topic with a game dev. Essentially, the argument was that games tend to have original UI and interaction models. For example, a web dev can get away with a <Button /> and a CSS/JS framework and make spectacular UI but when it comes to games the creatives would demand that button to be designed to match games concept, therefore for a game dev to build a button it may require to deal with all the styling and button states that a web dev would never touch(because it is taken care by the browser or the UI library).

An average game dev would have much deeper knowledge on the stuff they do than the average dev that build corporate software since the game devs would de bespoke work every time. Sure, they also have assets etc. but in the gaming the originality is much more important since the product itself is meant to be enjoyable, not simply a medium to accomplish something else(like ordering pizza or keep track of payments).

So a game dev would need to do something that a front end dev wouldn't have to do. OK, I'm sure the front end dev has to deal with a lot of things the game dev doesn't need to touch.

>OK, I'm sure the front end dev has to deal with a lot of things the game dev doesn't need to touch

Yep, like NPM and various dependency hells. Maybe that's where the compensation difference comes from?

Skim through some Handmade Hero: https://www.youtube.com/c/MollyRocket/videos

Large breadth of content with lots of esoteric and domain specific knowledge sprinkled throughout. The higher level abstractions available are very limited in scope in comparison to say GUI toolkits, most of the success has come from people drilling down and putting in a ton of work.

I do data science for a job and game dev as a hobby. When all is said and done, I don't know if it's really that much harder. But game dev products must run fast, might need to run efficiently over a network, almost inevitably has tons of unique stateful objects whose interactions can cause something to cease being fun, and that sucks.

IME they're roughly equal in terms of raw skill and experience required to publish a finished product. Its a bit apples to oranges however, there are some things in developing video games that are much more difficult than what I've experienced in non-game programming. A lot of it depends on the complexity of what you're making - you can make games that require almost no programming whatsoever, but some games require a great deal of technical skill to program. For instance, running a large-scale multiplayer game isn't so different from managing a large-scale web business.

Thanks for sharing first hand experience. So basically, "it depends".

Depending on the game, you might need to do some machine learning programming.

Forza heavily uses ML. You have Network programming, in there. Well I'm sure some specialties exist which are more difficult, it doesn't change the fact of video game programmer can make half as less as a normal business programer.

ML and network programming can be found outside of video games.

> I'm sure some specialties exist which are more difficult

OK, so we agree. Video game programming can definitely be very tricky, but it's like any other specialties once you reach a certain threshold of sophistication/scale/whatever.

In addition, real-time systems, graphics programming, modelling, and more.

But usually these are 3 different jobs done by 3 different people.

Depends on the shop.

I've seen video game studios which essentially have a single wunderkind who has to dabble in everything because no one else reaches his or her skill level. No other industry has low pay and crunch

Just like some companies in different industries where some dev need to understand design, front end, backend, database, QA, devops, etc.

> No other industry has low pay and crunch

I wasn't talking about the workload or low salaries. I was talking about the statement that "video game programming" as a discipline is the most difficult of all.

If someone's going to offer me $150,000 to do a job, I'm going to find it much easier than if I'm only offered $75,000.

And plus I started my statement with "I find" , if you happen to program quantum systems and that's 10 times as difficult compared to writing a game engine in CPP, that's cool