Talk about anything

KanataEXE Cool! I am just starting making a Udemy course! Would love to know more about how you're going about it

edit: in other news, crossing my fingers to see if unity crashes and burns. I'm happy just learning godot and unreal for now.

I mean, I dont want Unity to fail. I think it's decent technology (the engine itself), but John Riccitiello doesn't know what he is doing and took the company in the wrong direction. They might still survive, but it will require new leadership and a new direction. And it would be nice to get some refugees coming to Godot, but at the same time Godot isn't quite a 1:1 replacement for Unity (it's better in some areas, worse in others) so it may be too soon to declare victory.

Megalomaniak That's a big if tho...

Possibly the biggest if ever. But there are still games with smaller success on Steam and also a huge potential. It is not the case that you need to develop on consoles. Look at these stats from 2020.

• Over 50% of indie games never make more than $4,000. These games never generate enough revenue to cover the development time and other costs. Learn more
• An average indie game makes $3,947 (this is actually median, average is not a good measure as it’s significantly distorted by top earners).
• Over 2/3 of the games don’t cross $10,000 in life time earnings.
• However, the top 1/3 of the indie games are doing pretty well. Almost every 1 in 5 indie games makes more than $50,000. Still, this might not mean the game is profitable or a good investment. If the development team was any bigger than one person and took over a year to make, chances are these people would have been better off as full time developers for a big studio.
• Top 9% cross the $200k line. Most of these games are probably profitable and a proof that indie games can survive.
• The top 1% of indie games make over $7,000,000!

https://vginsights.com/insights/article/infographic-indie-game-revenues-on-steam

You can also see here that indie games make up 95% of Steam.

https://vginsights.com/insights/article/indie-games-make-up-40-of-all-units-sold-on-steam

So if you do the math, you have a 50% chance of losing all your money. And about a 67% chance of not making over $10K, which probably means it does not cover development costs. However, if you get in the upper 33%, you will be doing okay. The top 20% are going to make over $50K. If your game takes under a year to make, and it's a solo project, this is enough to live off of, even in America (and may be much more lucrative overseas since you're making USD). Granted, you could get a normal programming job and make more, but it's still a livable wage to work for yourself and make art. If you are in the top 9%, then you're making over $200K, so a game that took one person 2 years (or 2 people 1 year) you'd be making in the 6-figures. And if you get lucky to get to the top 1% you will make over $7 million. Honestly, this seems like a good gamble, provided you have a good quality game you have faith in.

C++ is really a pile of steaming bovine waste and should never be used to write safety critical systems. And the more I learn about it the more I understand why the Linux kernel guys do not want it in the kernel.

Example, who would remember if they not made that error before that indexing with an operator[] into a map for instance actually changes the map if the key is not found, thus provoking ub in a concurrent environment just with supposed concurrent reader access (that actually is no reader access), if not taken special care of ? And who knows what else is hidden as an error waiting to be made ?

What did Linus say "... and if it is only to keep out the mediocre C++ programmers" or so. Right he is. I would certainly have made that error, wasn't it mentioned as an example in an exercise.

grumblegrumble

:-)

[was piffle from my part]

There's no way around C++. Rust is gaining territory but has the paradigm "You don't know what you're doing so I won't let you". That's a bit the opposite. No, I fear we must sit down on our behinds and learn ...

C++ is great, I just don't think most people need that much power or performance. If you do, then you have no choice. But even some of the more advanced ML stuff is moving to GPUs. I think at some point CPUs and GPUs will merge, but the architecture will probably be more like GPUs. Basically a huge set of parallel streams that are fully programmable. Because x86 is a dead end at this point. Or at least it has the potential, but the software can't take advantage. While with GPUs you can essentially reach 100% potential today.