[qayxc]

There's two types of game developers: those who finish and publish games and those who write their own game engine.

Every minute spent on re-inventing the wheel for the umpteenth time is time lost on actually writing and polishing the actual game. Tools like Unity were a godsent for small teams and solo devs who just wanted to write and publish games.

While libraries like SDL can lighten the burden somewhat, they are still no replacement for solid asset pipelines, visual editors, performance analysing tools, and a mature pipeline for multi-platform packaging and publishing.

[badsectoracula]

> There's two types of game developers: those who finish and publish games and those who write their own game engine.

This is wrong, there are tons of finished games that use custom engines. For example one of the most recent indie games i bought was Zortch which was made by a single person and used a custom engine.

EDIT: actually here is a page that collects some custom engines, ranging from big studios down to one man efforts, all with released games:

https://gist.github.com/raysan5/909dc6cf33ed40223eb0dfe625c0...

[amatecha]

Huh, I'm not sure the full intent/scope of the list but Bungie shipped a LOT more games with custom engines than just Destiny. Marathon, Myth, Oni, and Halo, to name a few.

[badsectoracula]

It is certainly non-exhaustive, i think it just mentions some examples. REDengine for example was used not only for Witcher 3 but also Witcher 2 and Cyberpunk 2077.

[aaaaaalop]

This is meaningless without percentages. How many devs didn't release their games for each game in this list?

[badsectoracula]

This is not meaningless as the message i replied to claimed that developers who write their own game engine do not finish and publish games. That was factually incorrect.

What you are asking for is something completely different and unless humanity had some sort of hivemind to have universal knowledge, we can't really know as people start and abandon projects all the time.

The best we can do is some heavily biased guesses - and mine would be that the abandoned games in engines like Unity, GameMaker, etc where it is trivial to start a new project would be much more than even other engines (let alone custom engines).

[Hikikomori]

Only programmers deal in absolutes.

[fxtentacle]

That's an amazing list. And it hammers your point home. There are plenty of custom engines, it seems.

[outworlder]

> There's two types of game developers: those who finish and publish games and those who write their own game engine.

Probably a true statement in a way that's difficult to verify. Nobody is keeping track of never released games, specially those that didn't really get started because so much work went on the 'engine'.

That said, there are some examples. Like Minecraft and Factorio. An existing engine might not have benefited those titles as much as they would a standard FPS game. They could be probably made using an existing engine, but the benefits might not be that apparent if you have to fight the engine and twist it to do what you want.

Going to larger studios, Elite: Dangerous has its own engine too - could they have used, say, Unreal? Probably, but then they would not be able to use many features, like the Unreal's terrain system, for their planetary landings and such(not without ugly hacks - or making it work like Starfield) and would likely have to code their own. If you are doing that, might as well skip the whole thing.

I'll still agree that in 95% of the cases there's an existing engine that will handle most of a game's use-cases. If one has access to the source code, this figure probably expands significantly.

[dgb23]

You can appreciate the utility that unity provides without making false claims in the first paragraph.

There are ofc many successful games that were not built on top of a third party engine.

Including one of HN’s darlings Factorio for example.

[msla]

> There's two types of game developers: those who finish and publish games and those who write their own game engine.

Which is why that silly "Quake" thing never went anywhere.

What was John Carmack thinking? How did he ever imagine "Doom' and "Quake" could ever get released?

[bluescrn]

That was a very different time. That era of pushing hardware to it's limits is mostly history when it comes to games.

These days, hardware is almost limitless* , the real scarce resource is person-hours of development time.

(*Ok, not limitless, there's still some people doing serious optimisation work on games. But for an average indie/mobile developer, it's more about avoiding fairly well known performance pitfalls than worrying about low-level optimisation or inventing really clever algorithms to gain back some speed)

[msla]

Maybe a single person can't make a Triple-A game engine. That's not a problem, because a single person can't make anything else needed for a Triple-A game, either.

Photorealistic art assets? Voice acting from name actors? Really crappy hackneyed plots which fall to pieces under the slightest scrutiny? I think Bethesda actually has a patent on that. Lootbox mechanics which are a couple iffy election cycles away from a Senate subcommittee hearing? None of that is coming out of a bedroom developer. Writing a simple engine for a simple indie game with the kinds of art and mechanics a single person or small unfunded team can accomplish, however, is a lot more feasible.

[vunderba]

This isn't really a very good example for two reasons

1. The concept of licensable game engines during the development of doom in the early 90s was hardly a common thing, and certainly nothing along the lines of a low barrier, low cost entry like unity used to be

2. Id software is very well known for basically being a tech company more than a game company, and later went on to make a great deal of money licensing out the doom and quake engines to other companies who could then focus on creating games without having to build a bespoke engine