I will try to add every single thing suggested here

I wouldn’t, if you want the game to do well.

And I’d rather not constrain you to do so, because I don’t want to negatively-impact your game.

But I think that the discussion is an interesting one to have.

Let’s see. What are some of the things that I’ve missed or wanted?

Stale Genres

There are some genres that were once bigger, haven’t had a lot going on for a while. I kind of miss these.

• Space-based fighter combat. Freelancer-type games. Based on Star Citizen’s funding, a lot of people want more entrants here. I’ve theorized and read guesses before about why not a lot has happened here. I think that some of it is because the technology advanced; at one point in time, hardware was limited. Rendering outer space was relatively-easy, so it was a good match for back around 2000. Today, doing atmospheric fighters is easier. There were some popular “space fighter” movies, and spaceflight was relatively new; stuff like Babylon 5, Star Trek, and Star Wars all inspired people to want to be in those worlds. These tended to have a world where space combat was a lot like carrier combat in World War II; it used non-Newtonian physics (space fighters flew like terrestrial fighters through a liquid when turning), combat happened at close ranges, spacecraft were piloted, etc. I didn’t really like Elite Dangerous’s “faux online” model, though I suspect that it looked pretty with VR hardware. Egosoft’s X series is more about managing an empire than dogfighting.

Game types with few examples

• Early Close Combat-style games. I liked the early Close Combat games. These are top-view, real-time-tactics games. They weren’t very well-balanced, but they let one build up a force to run around fighting stuff with. They did not have a strategic map – I never really enjoyed the strategic side of refighting the things, kind of missed not having to think about the strategic layer, but still liked the ability to build up a force over time, kind of Homeworld-style. There are some vaguely-similar games, but nothing all that close.

• Air combat flight sims with a dynamic campaign engine where one can affect the flow of the game. Creating good game AI is hard. Creating even tactical AI has been difficult. Creating an AI capable of doing an operational-level dynamic campaign has been extremely rare. Most games that do it have had it modded in and are in very elderly games; IL-2 Sturmovik: 1946 had a dynamic campaign, though not one where a pilot could realistically do much to affect the outcome. I understand that Falcon 4.0 BMS has had it modded in. These are pretty elderly games.

• Terraria/Starfield. I liked these. I’d like to see more similar games. I’d like to see more automation support, ways to create bases that could be doing things (or pretending to do things, have some effect over time) when you’re away. There’s the Factorio factory-building genre, but that’s not really the same thing; it’s entirely-focused on automation.

• Kenshi. I can’t find any similar games. You control a squad, starting with one character. It can grow. You can build outposts and put some characters to work on automated tasks at them. The world dynamically evolves; factions can take control of different towns and such and the world changes as a result of this. There’s a sequel in the works, but it’s pretty unique. Mount & Blade: Warband has some similarities – band of characters, dynamically-evolving world, but isn’t really the same type of game at all. Has “battlefields” separate from the strategic world, whereas Kenshi has one continuous map, has whole armies, doesn’t permit creation of new outposts, doesn’t really model what’s going on at an outpost, only permits a static set of improvements to an outpost, etc.

• A Command: Modern Operations-like game that works on Linux. This is more-or-less a one-of-a-kind game, more of an operational-level hard realism military sim that can handle most modern military hardware. Unfortunately, it’s also unusual in that it’s one of the few games on Steam that doesn’t work on Linux. No real alternatives. I’d like it even more if it supported a dynamic campaign rather than fixed, scripted campaigns, though that might be getting greedy.

• Wargame: Red Dragon-like game with good single-player support. I don’t know quite how to describe this. It’s a bit like a MOBA, but plays more like a real-time-tactical game, with more-accurate modeling of real-world units realism. I am totally uninterested in playing it multiplayer, which is unfortunate, because Eugen saw it as an almost-purely-multiplayer-oriented game. Eugen makes Steel Division 2, which has…well, better single-player support, including a strategic-map campaign and considerably more-sophisticated single-player tactical AI, but that covers World War II. Not that there’s anything wrong with WWII, but there are a lot of games that cover it. Wargame: Red Dragon covered the late Cold War. I’d like to have a more-contemporary game; there isn’t a lot of coverage of newer military hardware.

• There are a couple of NSFW games I’d like to see more of, but you ruled those out, and I’m not sure that this is the best forum for those, anyway.

• Moddable, open-world games. This is probably way out of scope for an indie developer, unless they create a whole new ecosystem. Bethesda is probably the premiere example of this, but there’s a lot that Bethesda hasn’t done that I’d like to see. They haven’t really developed a very successful way to make more money off people playing their game in modded, which means that they don’t have a lot of incentive to improve that, which I think is at the root of a lot of the other problems. I don’t mind giving them more money, but I want them to improve the modding situation. I’d like to have solid diagnostic and script profiling tools, and preferably a standard language, like Javascript or Lua or something. I’d like to have support for parallelized scripting of some form, even if those need to run in isolation under constraints. I’d like to have a reasonably-sandboxed scripting environment, so that I don’t need to be paranoid about downloading and using random mods – something that is a chronic problem with many games and their mods. I’d like to have a system that can perform polygon generation and reduction dynamically – like, a map specifies a Bezier curve, and throwing more hardware at it just lets the engine generate more polygons; that’d be friendly to games that are expected to be played for many years. Throw a model at the thing, and it – maybe with some offline analysis – can come up with some kind of a reasonably-simplified model to display at distance, automatically billboard the thing at range (I think that existing tools can do the latter for Bethesda’s stuff). Keep loading out of the render threads so that loading doesn’t cause the frame rate to hiccup (Starfield’s engine is much better at this; I dunno how current Unreal is in open-world stuff). Avoid the limitations of static precombines in Fallout 4’s engine (Starfield may do this; I’m not conversant enough with the internals); this prevented some objects from being removed at runtime, since the engine wasn’t capable of recomputing data at runtime used for optimized display. I’d like to have support for more-sophisticated pathing, like for creatures that can jump and climb. I’d like to see in-game building support (not just for players, but for map development). I’d like to see profiling and suggestions (e.g. “this area of the map has the worst performance issues”). I’d like to see computer-assisted end-user diagnosing mods conflicts in the way that Conflict Catcher used to do for extensions on the classic Mac OS – do a git blame-style binary search across sets, enabling-and-disabling extensions and asking if a problem goes away, to try to find a single-extension problem in with a number of tests logarithmic in the number of mods, and able to find conflicts involving multiple extensions as well. I’d like to have support for streaming data from a remote system and having an engine that can run without data being fully-downloaded so that one can start playing before something is fully downloaded.

Functionality I’d like to see improved upon

• Game AI. For a long time, I’ve heard many promises about what a new game will do in terms of game AI, and this is probably one of the largest areas where games tend to disappoint. I think that that’s partly because this is dealt with later during development, partly because it’s just a hard problem, and partly because there doesn’t seem to be a standard cross-game “game-AI engine”. I’d like it to be possible to record playthroughs and, with some annotation from developers as to actions, learn from player data to optimize strategies. I’d like to be able to play some games optimally in the sense that von Neumann “solved” poker in terms of mathematically optimal play absent tells. I’d like to be able to have generic ways of imposing “human-like” limitations on a game AI so that it can play suboptimally in ways that feel more like a human opponent, like constraining the number of evaluations performed. Basically, I feel like the “game AI” problem should be more-solved in a cross-game way than it is today.

Features that I’d like to see used more

• High-res DLC for pixel art games. Low-res pixel art can reduce art costs. That’s great. But then, sometimes a game does really well. I’d like to be able to buy high-res art for it. Very few games have done this (Cave Story being one of the few).

• Music available for purchase. I can at least understand not doing this for the few games that permit one to use arbitrary music playlists (like Stellaris).

[continued in child]