That said totally not a waste of time. It shows you have the wits to bring something to market and the ability to ship. You coded the whole damn thing which is insanely involved. This is no small feat. However the market is generally the hardest critic and it doesn't matter how many hours you spent or how many lines and bugs you solved.
That was my first thought too. The author talks about the indie explosion, but I don't think they realize they are part of the problem. Clearly the author is a great programmer, as coding something like this definitely isn't trivial, but it lacks in other places, such as design, art style and sound effects.
I understand programmers often like working on their own projects, and sometimes we end up with complete packages like Stardew Valley, but maybe it's better to work in a team where everyone has their own strengths. I see so many games with fantastic code, going to waste because of really underwhelming story, art style or sound.
I consider myself as a big fan of Metroidvanias, I'm even the kind of person who actively looks for new games in this genre, and those were my thoughts as well.
I would even say that this game manages to hit the 3 points in which I avoid on a Metroidvania:
1. Rogue-like. I really dislike rogue-like, no special reason, just a personal preference
2. Lack of a plot. I appreciate the feeling of exploring a world the feels alive, even if it's a very simple one. Going through levels for the sake of going through them, it's not much of a fun experience to me.
3. Huge resemblance to the original Metroid. If I wanted to play Metroid... I would be playing Metroid.
Also... no Linux version? Really? That excludes me entirely from this game.
#1 and #2 are strongly related. Procedural generation makes it much harder to create a sense of a coherent world, rather than a series of disconnected mini-games.
And yet Dead Cells does the Metroidvania Roguelite thing and has a plot.
In fact being a Roguelite is woven into the plot, characters aren't surprised to see you, after all you were just here last session, the mounds of festering corpses prompt you to remark that they're all the same... They're all you.
You're criticizing an indie game developer who feels he just wasted 3 years of his life making a failure, for not spending the extra year or whatever it'd take for it to be cross-platform?
With engines like Unity and Unreal, supporting multiple platforms (including modern consoles like PS4, Xbox One and Switch) has a level of complexity of writing a portable Electron app.
As an indie developer, you need to maximize that market coverage (and develop with portability in mind).
Having said that, it's probably not the primary reason why his game failed.
- Linux desktop market share is tiny. My own app supports Linux, but I'm 100% aware that I'm doing it at a loss in every possible way, it's a passion project. Bang for buck is terrible, so you need to justify as a labour of love.
- The amount of pain one needs to endure to get a Linux desktop to work as it should is huge, and there are several competing packaging providers with no clear winner, and all have very much hidden gotchas that they do a poor job of explaining ahead of time.
- Making a cross-platform Electron app that behaves well and up to to snuff on Windows, Mac and Linux is not even close to easy. The fact that JS theoretically works on all three platforms buys you way less than most people think.
Source: I'm making a cross platform Electron app that supports Linux.
The market may be tiny but it is a lot easier to sell copies of a (decent) game because Linux users will love you for supporting their platform.
You also don't need to support every distro on the planet. Just focus on Ubuntu. With proper planning and choosing your game engine wisely it's not the biggest deal to build for Linux.
My experience (though I'm not a game developer) is the opposite. Linux users are very demanding, and instead of being thankful that your app works well on Linux, they'll take the fact that it works and works well on Linux for granted, and will resent you for charging money for it.
If you support Ubuntu+Debian (1.15% user base globally on average), the next feature request you'll get will be Xubuntu, Arch, and then some smaller distros which has their own undocumented quirks, they'll ask for 32 bit versions (0.0015%) to run on ancient machines that aren't really powerful enough to run the app anyway, and there goes the rabbit hole.
(In the meanwhile, Windows users are 85%, Mac OS is 13%. We're talking about fractions of fractions a percent here when you move out of Ubuntu x64)
These features will be framed as "You're supporting Ubuntu, getting it to work on this {{similar_distro}} is so close, you should do it and you'll have a lot of users". It's not that they're wrong or malicious — it's just that their concept of a lot of users is a whopping multitude of three people.
I'm also purposefully ignoring the more acrid side of the Linux community where they'll call you names, find your personal email, and make sure it's the first thing you read in the morning for not pulling heroics to make it work for their distro of choice (0.0000075% user base).
All in all, not worth it, really. Not financially, not logically. Not from a human point of view, either.
Here are a few things I've found helpful if you're making a desktop app for Linux:
- Consider charging Linux users for support. This is justifiable because for every Windows support request, there are likely 10 people that experienced the problem and haven't written to you about it, for Mac, 2-3, but for Linux, very likely you're only helping that single guy only. This is the best way to do this, but since my app is free, I don't really want to set up a payment infrastructure.
- Make your app free, and ask Linux users to either make their own builds from unpacked releases, or pay for support for their distro on a rolling basis. You don't really expect anyone to take the latter, but it does wonders to cut down on requests in which people demand you support their favourite obscure distro of choice with no help or support from them.
That said, I still provide Snaps, as it's the closest I can get to a universal Linux runtime. This exposes me to requests to provide AppImage, Flatpak, and some other stuff even then, but it's way better than trying to support distros directly. [0]
[0] I tried to support AppImage, I gave up after a full day of trying. Flatpak had similar issues. One of the core developers of AppImage reached out trying to debug, and I helped him as much as I could — but the point is, while the intentions are pure, and I'm glad for the effort, this is deeper and deeper into the red in terms of price / performance.
I agree that Linux users can also be a pain to deal with :)
But yea from my experience the game market is a little different because there is a growing group of people that rather would not boot Windows for gaming and instead stay on their platform. This group of people is very thankful for ported games.
What I heard third hand from game developers has been that it’s fairly hard to get games to perform well on Linux, and that most of the complaints come because of performance reasons outside the developers’ control.
I disagree with your assertion Linux desktops are hard to set up.
You plug in your Ubuntu drive and install it, it detects your video card and install the drivers.
Everything just works, when people rag on Linux desktops they are talking about Linux from 6 years ago.
A ton of effort has been put into making the experience smooth and there are multiple projects to make it even more user friendly like elementary os and popos
This is true, but also the game in question was coded entirely from scratch, which is part of the problem: the majority of those 3 years were probably spent duplicating the work done already by Unity and Unreal. If you want to make a living as a carpenter, you don't start by planting walnuts.
There's not a ton of information available on the breakdown of sales by platform, but the data that is out there shows Linux sales usually compromise ~2-4% for cross-platform games [1]. If you get linux support "for free" by utilizing an engine, then absolutely go for it, but even major companies with massively popular games are still skipping out on first-class Linux support [2], due to the low market share.
It's also exponentially harder to make a Linux port for a AAA game than for an Indie game, for many reasons. The most obvious ones being that most indie games are made with an off the shelf cross platform engine (Unity, unreal), and that AAA titles require a lot more hardware resources, and optimization of this usage is often platform specific (what helps OpenGL might not help DirectX, etc).
When was the last time you paid for an indie humble bundle though? I think it proved that for a short moment a very small percentage of users were supporting games at a higher rate than the majority in order to support a cause, but once the novelty wore off they went back to not caring. I just don't think there are enough Linux users out there that are also regular gamers to make Linux support worth it for most titles. Sure there are exceptions especially if you're building a game that disproportionately appeals to that audience (like KSP or Zactronics), but in most cases its not really worth it.
>When was the last time you paid for an indie humble bundle though?
Some years ago, but that is explained by a sharp decline in the quality of games included in their offers (which was about the time they dropped ensuring that their bundles were compatible with the three platforms). A better question would be when I last spend money on (game) software for linux. The answer is 28 hours ago.
I concur that the market is small. But when your sales are small too, it is not smart to throw away access to a market with a higher proportion of dedicated users and consumers. From the perspective of tech, less time may have been "wasted" on infinitroid if the developer used more off the shelve solutions that would help with multiplatform targeting.
In the end we are just basing our arguments on our gut feelings. I doubt very much that Valve would push as hard for gaming on linux as they are doing if they didn't see the potentials of sales for the platform in their data. But time will tell.
I think in OPs case a Linux port could have a non-trivial affect on sales because he's on the front page of hackernews and people that see this post and feel sorry for him will buy his game and not play it because he supports Linux.
I agree, C++ projects are quite easy to get running on Linux. I've done cross platform C++ game stuff and almost always had more trouble with Windows than I ever did with Linux. MacOS was pretty annoying in some cases too, actually...
Yeah, the entirety your comment pretty much mirrors what I'd have to say in reply to his post.
The only other thing I'd add for Luke is - do you enjoy playing this game? As in, have you sunk a whooole bunch of hours into playing it, simply because it's the only game available (that you made precisely because there's nothing else) that scratches this very particular itch?
I'm afraid this was my first thought having watched the video — does it look fun?
I'd like to think that the foundations are in place to make it such, but from what I saw, I saw a lot of [highly capable] box ticking, but not a lot to make me want to take it on.
Super Metroid is literally my favorite game, and I don't really have a strong urge to play this.
Metroid is an exploration-based game. The game rewards you for finding secrets and for knowing how to get places. It teases you to find a way to break sequence, and much of the replay-ability of that game is based on the possibility to do that and to bask in what you've already learned about the world. Procedural generation takes a big dump on any sense of familiarity, which is a big part of the reward for exploration.
In fact, the _only_ games I've played where procedural generation were good for the game are story-building games, such as Rogue-likes and Dwarf Fortress. They reward you for building a story, not for traversing obstacles. In every other game, they are just a weak thematic obscuration of the underlying mechanical goals. The Dryad's name in Terraria doesn't matter. Dig deep enough and you'll find diamonds in Minecraft. Kill a boss enough times in Borderlands, and you'll get a good gun. There's no story about achieving these goals. Procedural generation doesn't participate in making the goals more interesting to achieve, it's a forgettable and incidental fact about something almost wholly unrelated.
Game developers need to stop trying to lean on it as a substitute for content. A game is what people can expect from it each time they play it, and if all it is is a bundle of mechanics and throw-backs, then there's not going to be much appeal.
I'm curious about where the story happens? I've played a rogue-like (Pixel Dungeon), while I found it fun, I didn't think it had much of a story. Are others better?
In say nethack you can pick up random overpowered or weird items that change the game or get into crazy adventures when you meet some weird rare creature.
I kinda like games like nethack and brogue but my only problem with them is since it's extremely easy to die, it emphasizes attention a bit too much imho. I'm not a very attentive person and when I play games I'd much rather it be a bit more relaxed. When I know that even if I play the game for a whole Saturday, one floating eye can just randomly kill me for no reason, it kinda demotivates me...
It works like gambling. You play the game 100 times and you just lose most of them, but on the rare occasion you get very lucky... and you walk away with an unforgettable story about how you defied all the odds and did something wonderful and amazing. That's where procedural generation shines the brightest.
I suppose that's true. But one can imagine a game just like nethack except it's not as easy to die so after, say 20 run, you learn the ropes and can enjoy the game every time you play it. Obviously, eventually you'll get bored (unlike playing 100 times and losing every time) but that few times you beat the game would be enjoyable.
Not nethack, I was talking about a hypothetical clone of nethack where you can supposedly make a whole run without "surprises" that can randomly kill you, like floating eyes (or Jelly monsters in brogue). I think such a game would be very enjoyable first few times you play, but then since not challenging would be boring. Entirely different genre but Universal Paperclip might be a good example, extremely interesting story but the game is almost too easy after you learn to reliably win it. I think nethack is "boredom proof" because it's so challenging. I must be honest, I never finished brogue nor nethack (but I suck at video games because I play them once a year or so) but every game still gives the same rush of enjoyment every time. Still, though I don't keep playing it because the "demotivation" thing I was talking about earlier stops me from starting a new game (I'll lose anyway).
That's right, Tariru has won 61 times in a row. So that makes my 100 losses 'avoidable' in some sense, which spurs me on to do better. It only requires luck if you play inattentively :)
At the risk of sounding like an elitist - Pixel Dungeon isn't a "real" roguelike. Check out something like Nethack, or ADOM, or Dungeon Crawl Stone Soup. Or maybe Tales of Maj'eyal if you want something with lots of graphical polish (this one also has a decent amount of actual scripted plot).
These games have the extra depth and detail that cause the kinds of emergent gameplay and "storybuilding" that people are talking about.
In Sproggiwood, the story takes place around dives into the dungeons. Each dive is "rogue like" but the overall progression outside each dive is persistent.
I think this is the best way to do a random dungeon roguelike with a story.
That said, I wouldn't sell Sproggiwood on its story in particular, just the arrangement of everything. The game is fun because it's a good bite size style of roguelike for a phone with interesting/fun mechanics in play.
I would recommend Rogue Survivor. Pixel Dungeon is an ok game, but it doesn't really encourage you to get attached to your character or environment. There's very little backtracking or foreshadowing in it, so learning the lay of the land or remembering where things are just isn't very important.
If you are aiming for a game that people play for more than 1000 hours, sure. Commercially, this would be a terrible segment of the market to aim for. The people who play the same title for that long are unlikely to buy your game, because they don't have time to play it. Unless you happen to find a niche first, you will end up with a few players, but supporting them for a long, long time.
Shorter games are fine, and sell to people who want novelty over repetition. Don't make the game longer than its content, and price accordingly.
It doesn't make a game better, it just superficially defers the realization that the game is played out. Yatzee's only claim to being a game is that you roll dice... if you didn't, people would immediately see it as the pointless activity it really is. Video games are no different. If you put 1000 hours into a game, you will see it for what it is, and proc gen is not even a factor at providing interesting gameplay at that point. A game ultimately succeeds based on the merits of the structure it actually has.
Now, if the proc gen is sufficiently complex, we're talking about something else, (but that is rare if it exists at all.)
I would say it is a perfect example of a story-building game. The specific details that drive the main gameplay mechanics are based on the circumstantial arrangement of procedurally generated components. Where any two players are in the world matter to the events that subsequently play out, and those relationships are based on proc gen.
I don't know if I'd agree about your dislike of the graphics being because of your own personal preference. The human brain has evolved to find certain harmonies of shapes/colors attractive while others are not. The author undoubtedly has a lot of grit—but, he's not an artist, and his game suffers because of that. Metroid has superior graphics because they had dedicated artists, and surely put a lot of time/thought into the look and feel of the game.
Also, the design scheme between different 'blocks' which make up the platforms in the game is not uniform, which makes the design feel much more disjointed.
Games these days are a lot about marketing and huge budgets. The indie ones that do well need to be targeted as well as refreshing.
Some hard thoughts just watching the video and reading a bit about the game:
1. I don't really like the graphics as much as I liked the original Metroid. This is probably just a personal preference.
2. It mostly screams low quality "Metroid clone" and not something cool I'd tell my friends about.
3. Procedurally generated levels doesn't sell me. I don't really care.
That said totally not a waste of time. It shows you have the wits to bring something to market and the ability to ship. You coded the whole damn thing which is insanely involved. This is no small feat. However the market is generally the hardest critic and it doesn't matter how many hours you spent or how many lines and bugs you solved.