18 ms·

You made the mistake of thinking that anyone would care how many hours you spent optimizing some C++ function that does something already solved a hundred times

You made the mistake of thinking that anyone would care how many hours you spent optimizing some C++ function that does something already solved a hundred times in a hundred different game engines. It's a natural tendency for all programmers. But making a game in 2018 is far more of a creative endeavor than anything to do with programming really. You need a massive amount of top notch artwork, music, 3D modelling, shader effects, SFX, etc. to have a polished nice looking game. That takes either superhuman talent or a large team of specialized people beyond yourself.

giancarlostoro 8y ago
Seeing the forest for the trees as they say. This is why indie games are focused on the simpler style of video games in some cases like Stardew Valley. It still took the developer 8 years (give or take) and you can tell he worked his tail on it.
- scrollaway 8y ago
  Stardew Valley is insane. The developer did everything. Art, music, sound effects, engine, gameplay, design, writing... everything.
  - derefr 8y ago
    Stardew Valley is an interesting case in that the developer didn't really need to invent anything. They took a formula (the Harvest Moon series), produced a clone, and then polished the hell out of it in response to community demands until it was hardly recognizable. None of that really required an act of genius creative talent, though. Learning enough about audio production to create the set of SFX and music you already know you need from the top-down design of the clone-game you already know you're aiming for is effectively "just" grunt work. Sort of like how an actor learning the exact set of martial-arts moves they need to know to look like they know martial arts in one particular shot, which is intensely choreographed, is "just" grunt work. Stardew Valley's creator didn't need to develop the skills to be an amazing artist or musician, they just needed—like the actor—to develop the skills to create a particular pastiche of existing pieces of art and music. The creator of Stardew Valley is, however, a natural talent at the particular meta-skill of "finding out what people want and tweaking things in response to those wants, never settling for doing 'merely' what they know how to do with the skills they already have, rather pushing themselves to the skill-level required to make people even more happy."
    - TeMPOraL 8y ago
      > Learning enough about audio production to create the set of SFX and music you already know you need from the top-down design of the clone-game you already know you're aiming for is effectively "just" grunt work. (...) Stardew Valley's creator didn't need to develop the skills to be an amazing artist or musician, they just needed—like the actor—to develop the skills to create a particular pastiche of existing pieces of art and music. Would almost agree but you had to bring up this :). I actually talked with SDV developer about his music skills[0]. Turns out, music was his hobby prior to becoming a developer, so he has a lot of experience. So in this case I wouldn't say it's "just', and it definitely wasn't a skill developed for this game. -- [0] -- https://www.reddit.com/r/StardewValley/comments/4acds9/how_do_you_begin_creating_something_like_stardew/d109oz2/ https://www.reddit.com/r/StardewValley/comments/4acds9/how_d...
    - aldanor 8y ago
      > in response to community demands I don't think so. The author mentioned in one of his interviews that he was the only person to actually play the game until the very late pre-release stages. Even those who signed him haven't really played it.
Nokinside 8y ago
If a lone programmer wants to write tight code for games, they should build game engines or libraries that games can use and then sell those to game developers. There exists large value added chain for specialized tools, arts and libraries.
- gmueckl 8y ago
  Except that it turns out that artists and game programmers are too cheap to pay for tools, even if you can present them with a clear benefit in terms of productivity gain vs. sales price. To make matters worse, big companies that could afford those tools will generally only buy exceptionally well established tools and rebuild the rest from scratch in-house. The only way to make money with games is to sell games.
  - azernik 8y ago
    They're absolutely not "too cheap" for that - it's just a saturated market. Don't start a new company, look for a job working at Autodesk or Havok or Unity or Crytek or one of the dozen other companies that makes bank off of selling professional tools to professional game developers
    - gmueckl 8y ago
      I can tell you right here and right now that I know that some of the companies you name do not make a bank off of their products. And it's probably not the companies you would expect and not for the reasons you would expect. I have sources for this, but unfortunately I cannot name them.
      - 1123581321 8y ago
        Can you tell us which companies and the reasons? Otherwise, it’s hard to learn from your comment.
    - slavik81 8y ago
      Autodesk has laid off about a quarter of its workforce in the past two years. Crytek has had employees regularly complaining about paychecks bouncing; they're clearly inches from bankruptcy. Havok is owned by Microsoft; I'm not sure how they're doing. Unity is probably doing well enough, as they're incredibly popular, but I suspect popularity among hobbyists translates mostly into a lot of people on the free tier. The tooling market is hard.
chime 8y ago
> how many hours you spent optimizing some C++ function > You need a massive amount of top notch artwork, music, 3D modelling, shader effects, SFX, etc. to have a polished nice looking game. Top notch in any field requires optimizing hidden, underlying dependencies to deliver the best output. Optimizing code is not sufficient but it is usually necessary, especially if others have already solved it, setting the standard of performance that players expect. The way to avoid optimizing core library code is to use someone else's library and build on top of that. But you can only do that so much. This is just an insanely hard industry to get into.
- eXpl0it3r 8y ago
  It's not that optimizations don't have their place, but when your goal is to create a game and make money off of it, then micro optimizing C++ code simply has no place. Also, a 2D platformer simply doesn't require any highly optimized system. One should spend the time building a game, polishing it, marketing it, presenting it at shows, etc. and not trying to build the most efficient engine, just because engine building is quite fun - as long as the goal is to make money from a game.
mettamage 8y ago
> It's a natural tendency for all programmers. Then I'm not a programmer :) Less optimizing means less work! I like spending my free time in all kinds of ways.
aswanson 8y ago
Minecraft.
- ronyeh 8y ago
  True. But most people aren't as skilled as Notch, or the Stardew guy. It's still definitely NOT a waste of three years, but the author should not expect to make much money off of an indie game. More graphic design and marketing needs to be put into this endeavor, assuming the gameplay itself is any good.
- simias 8y ago
  It's easy to list the games that made it, harder to list the thousand of similar games that didn't and were forgotten. Minecraft didn't make it because it was super well coded (it was decent but I'm sure well-optimized C or C++ could have easily outperformed the JVM), it was mainly at the right place at the right time. I think the reason Minecraft was successful was first and foremost because of its concept. Looking at the author's game is seems pretty obviously like a Super Metroid clone (at least judging from a video, the level design looks like reskinned Super Metroid levels and many of the powers and enemies are extremely reminiscent of Metroid). I really like Metroidvanias but as far as I'm concerned it's really the art style that kills it, I'm really not fond of this pseudo-realistic tile work.
- scrollaway 8y ago
  GP is completely wrong on their last assertion. You do not need AAA quality to make it as a game dev. But citing Minecraft reinforces their first assertion. Minecraft was an underrepresented genre, released early, with an (accidentally) excellent marketing technique. Above all, the code is an atrocious, unoptimized Java pile of crap (Yeah, I used to be into Minecraft dev in the early days). It's the best proof you can give that "making a game [...] is far more of a creative endeavor than anything to do with programming".
  - TeMPOraL 8y ago
    SDV is another example; the author himself admitted on numerous occasions that the code is, well, not stellar, and in the community we also had some fun reading through decompilations of the parts of the game. Yet another proof would be surprising number of very games made in tools like GameMaker - think Nuclear Throne, or Cook Serve Delicious. I actually took a peek at the sources of the latter (they were distributed with some Humble Bundle once), and it's... reinforcing this point.
EthanHeilman 8y ago
> You need a massive amount of top notch artwork, music, 3D modelling, shader effects, SFX, etc. to have a polished nice looking game. That takes either superhuman talent or a large team of specialized people beyond yourself. I think gameplay and marketing win here. Consider two gamesL one has great graphics but bad gameplay, the other has bad graphics but great gameplay. The great gameplay game is going to win that match up. The problem here seems to be that supply is geometricly growing whereas demand is largely fixed. To compete and be noticed you need to make a serious investment in marketing or get extremely lucky.
- rabidrat 8y ago
  > one has great graphics but bad gameplay, the other has bad graphics but great gameplay. The great gameplay game is going to win that match up. Yes, in terms of player satisfaction, but no, not in terms of sales. Terrible movies with great explosions out-earn great movies with low budgets every single summer.
  - baud147258 8y ago
    > Terrible movies with great explosions They usually have bigger marketing than low budget movies.
Kagerjay 8y ago
> 62,176 lines of code then I watched the youtube video demo in 2017. A 2D gaming scroller that requires 62,126 lines of code written from scratch? There's tons of libraries out there, game engines you could adopt, etc. I would argue you could be an awful programmer and use prebuilt tools but have great design aesthetics, vision, music composition, and storytelling, and deliver something far better
- gmueckl 8y ago
  How can you criticize the amount of code without knowing how much of that is generic engine code and how much of it is gameplay code (game logic, AI, event triggers...) that would have to exist regardless of the engine being used?
  - hannasanarion 8y ago
    Because in the former case, it's a waste of time, duplicating work that others have done already available online for free, so of course there are no customers And in the latter case, it's got to be a poorly written mess, no 2d platformer could need so much game logic. That's twice the size of the entire Super Metroid cartrige, engine and assets included.
    - gmueckl 8y ago
      Wait... did you just compare lines of source code to the size of a compiled binary code of something that was written over 25 years ago and clearly optimized for the size of the deliverable? In that case you might just as well have compared an apple to the Apollo program LAM module. It is just not a valid comparison. I am sure that if Super Metroid were to be developed today in a high level language with the exact same gameplay logic, the resulting code would be in at least the same ballpark.
      - hannasanarion 8y ago
        It is a valid comparison, because what he made is essentially the same game. And I said that in response to the guess that maybe the 62,000 lines of code was all game logic, which is absurd. Game logic, like, "when the player presses the fire button, spawn a missile" and "this enemy goes up 5 units, then goes down 5 units, then repeats". That's not the kind of thing that will change much from one metroid clone to another, regardless of what technology is used to make them. It is practically always the smallest part of the final product. And to have the game logic's source be bigger than the entirety of an almost identical game, with handcrafted levels, art, and engine included? It's absolutely absurd. It was in service of the point above: that code is probably all the engine, which means that, yes, it probably was a waste of time, since engines made by professionals are available for free. At best, it's a line on a resume.
LrnByTeach 8y ago
Great analysis.. this applies to so many areas of software development. Building games, websites, any thing software .. > you spent optimizing some C++ function that does something already solved a hundred times in a hundred different game engines. It's a natural tendency for all programmers. > Making g a game in 2018 is far more of a creative endeavor than anything to do with programming really. > You need a massive amount of top notch artwork, music, 3D modelling, shader effects, SFX, etc. to have a polished nice looking game. > That takes either superhuman talent or a large team of specialized people beyond yourself.
- derefr 8y ago
  You know, there's an interesting relationship between the graphic-novel industry and the TV/film industry at this point. Many TV/film screenplays are adapted from graphic novels, because a graphic novel is, in a sense, a "sketch" of a film—it's something that communicates most of the top-down spirit of the eventual work, while also being something that is able to be produced by a single creator, and so something that is able to let a single creator's vision and talent shine through. I'm left wondering why the video-game industry doesn't have at least some sub-sector with an analogous pipeline, adapting (or in this case, "covering") low-budget indie titles that don't have asset-polish into AAA games, by giving them that asset polish.
  - jvanderbot 8y ago
    I'd buy three beers to hear more about this idea. Current acquisition model by AAA is buy-kill, rather than buy-subsidize.
  - hannasanarion 8y ago
    Because asset polish doesn't matter nearly as much as gameplay: the high-res remake will never make as much as the original. The most successful game of the year, Far Cry 5, has sold 1/20th as many copies as Minecraft has, a game that is ugly as sin and always has been, but it's fun enough that it doesn't matter. There's no point in adapting a property if the adaptation won't grow the audience. The Walking Dead was a comic book superstar, selling upwards of 500,000 units per issue when it was picked up by AMC for the TV adaptation. The TV adaptation is now reaching 8 million viewers per episode, and that's considered low.
mooseburger 8y ago
It doesn't take a large team. Look at latest indie hotness Hollow Knight (https://youtu.be/UAO2urG23S4 https://youtu.be/UAO2urG23S4), which was made by 3 dudes, who were artists and animators (http://hollowknight.com/our-team/ http://hollowknight.com/our-team/). I agree that with game engines, making a successful indie game is primarily about artistic prowess, but you probably don't need that many artists. Just 1 or 2 really good ones.
chrischen 8y ago
PUBG is evidence of the contrary, though the the game is just pure fun.

giancarlostoro 8y ago

Seeing the forest for the trees as they say. This is why indie games are focused on the simpler style of video games in some cases like Stardew Valley. It still took the developer 8 years (give or take) and you can tell he worked his tail on it.

scrollaway 8y ago
Stardew Valley is insane. The developer did everything. Art, music, sound effects, engine, gameplay, design, writing... everything.
- derefr 8y ago
  Stardew Valley is an interesting case in that the developer didn't really need to invent anything. They took a formula (the Harvest Moon series), produced a clone, and then polished the hell out of it in response to community demands until it was hardly recognizable. None of that really required an act of genius creative talent, though. Learning enough about audio production to create the set of SFX and music you already know you need from the top-down design of the clone-game you already know you're aiming for is effectively "just" grunt work. Sort of like how an actor learning the exact set of martial-arts moves they need to know to look like they know martial arts in one particular shot, which is intensely choreographed, is "just" grunt work. Stardew Valley's creator didn't need to develop the skills to be an amazing artist or musician, they just needed—like the actor—to develop the skills to create a particular pastiche of existing pieces of art and music. The creator of Stardew Valley is, however, a natural talent at the particular meta-skill of "finding out what people want and tweaking things in response to those wants, never settling for doing 'merely' what they know how to do with the skills they already have, rather pushing themselves to the skill-level required to make people even more happy."
  - TeMPOraL 8y ago
    > Learning enough about audio production to create the set of SFX and music you already know you need from the top-down design of the clone-game you already know you're aiming for is effectively "just" grunt work. (...) Stardew Valley's creator didn't need to develop the skills to be an amazing artist or musician, they just needed—like the actor—to develop the skills to create a particular pastiche of existing pieces of art and music. Would almost agree but you had to bring up this :). I actually talked with SDV developer about his music skills[0]. Turns out, music was his hobby prior to becoming a developer, so he has a lot of experience. So in this case I wouldn't say it's "just', and it definitely wasn't a skill developed for this game. -- [0] -- https://www.reddit.com/r/StardewValley/comments/4acds9/how_do_you_begin_creating_something_like_stardew/d109oz2/ https://www.reddit.com/r/StardewValley/comments/4acds9/how_d...
  - aldanor 8y ago
    > in response to community demands I don't think so. The author mentioned in one of his interviews that he was the only person to actually play the game until the very late pre-release stages. Even those who signed him haven't really played it.

Nokinside 8y ago

If a lone programmer wants to write tight code for games, they should build game engines or libraries that games can use and then sell those to game developers. There exists large value added chain for specialized tools, arts and libraries.

gmueckl 8y ago
Except that it turns out that artists and game programmers are too cheap to pay for tools, even if you can present them with a clear benefit in terms of productivity gain vs. sales price. To make matters worse, big companies that could afford those tools will generally only buy exceptionally well established tools and rebuild the rest from scratch in-house. The only way to make money with games is to sell games.
- azernik 8y ago
  They're absolutely not "too cheap" for that - it's just a saturated market. Don't start a new company, look for a job working at Autodesk or Havok or Unity or Crytek or one of the dozen other companies that makes bank off of selling professional tools to professional game developers
  - gmueckl 8y ago
    I can tell you right here and right now that I know that some of the companies you name do not make a bank off of their products. And it's probably not the companies you would expect and not for the reasons you would expect. I have sources for this, but unfortunately I cannot name them.
    - 1123581321 8y ago
      Can you tell us which companies and the reasons? Otherwise, it’s hard to learn from your comment.
  - slavik81 8y ago
    Autodesk has laid off about a quarter of its workforce in the past two years. Crytek has had employees regularly complaining about paychecks bouncing; they're clearly inches from bankruptcy. Havok is owned by Microsoft; I'm not sure how they're doing. Unity is probably doing well enough, as they're incredibly popular, but I suspect popularity among hobbyists translates mostly into a lot of people on the free tier. The tooling market is hard.

chime 8y ago

> how many hours you spent optimizing some C++ function > You need a massive amount of top notch artwork, music, 3D modelling, shader effects, SFX, etc. to have a polished nice looking game. Top notch in any field requires optimizing hidden, underlying dependencies to deliver the best output. Optimizing code is not sufficient but it is usually necessary, especially if others have already solved it, setting the standard of performance that players expect. The way to avoid optimizing core library code is to use someone else's library and build on top of that. But you can only do that so much. This is just an insanely hard industry to get into.

eXpl0it3r 8y ago
It's not that optimizations don't have their place, but when your goal is to create a game and make money off of it, then micro optimizing C++ code simply has no place. Also, a 2D platformer simply doesn't require any highly optimized system. One should spend the time building a game, polishing it, marketing it, presenting it at shows, etc. and not trying to build the most efficient engine, just because engine building is quite fun - as long as the goal is to make money from a game.

mettamage 8y ago

> It's a natural tendency for all programmers. Then I'm not a programmer :) Less optimizing means less work! I like spending my free time in all kinds of ways.

aswanson 8y ago

Minecraft.

ronyeh 8y ago
True. But most people aren't as skilled as Notch, or the Stardew guy. It's still definitely NOT a waste of three years, but the author should not expect to make much money off of an indie game. More graphic design and marketing needs to be put into this endeavor, assuming the gameplay itself is any good.
simias 8y ago
It's easy to list the games that made it, harder to list the thousand of similar games that didn't and were forgotten. Minecraft didn't make it because it was super well coded (it was decent but I'm sure well-optimized C or C++ could have easily outperformed the JVM), it was mainly at the right place at the right time. I think the reason Minecraft was successful was first and foremost because of its concept. Looking at the author's game is seems pretty obviously like a Super Metroid clone (at least judging from a video, the level design looks like reskinned Super Metroid levels and many of the powers and enemies are extremely reminiscent of Metroid). I really like Metroidvanias but as far as I'm concerned it's really the art style that kills it, I'm really not fond of this pseudo-realistic tile work.
scrollaway 8y ago
GP is completely wrong on their last assertion. You do not need AAA quality to make it as a game dev. But citing Minecraft reinforces their first assertion. Minecraft was an underrepresented genre, released early, with an (accidentally) excellent marketing technique. Above all, the code is an atrocious, unoptimized Java pile of crap (Yeah, I used to be into Minecraft dev in the early days). It's the best proof you can give that "making a game [...] is far more of a creative endeavor than anything to do with programming".
- TeMPOraL 8y ago
  SDV is another example; the author himself admitted on numerous occasions that the code is, well, not stellar, and in the community we also had some fun reading through decompilations of the parts of the game. Yet another proof would be surprising number of very games made in tools like GameMaker - think Nuclear Throne, or Cook Serve Delicious. I actually took a peek at the sources of the latter (they were distributed with some Humble Bundle once), and it's... reinforcing this point.

EthanHeilman 8y ago

> You need a massive amount of top notch artwork, music, 3D modelling, shader effects, SFX, etc. to have a polished nice looking game. That takes either superhuman talent or a large team of specialized people beyond yourself. I think gameplay and marketing win here. Consider two gamesL one has great graphics but bad gameplay, the other has bad graphics but great gameplay. The great gameplay game is going to win that match up. The problem here seems to be that supply is geometricly growing whereas demand is largely fixed. To compete and be noticed you need to make a serious investment in marketing or get extremely lucky.

rabidrat 8y ago
> one has great graphics but bad gameplay, the other has bad graphics but great gameplay. The great gameplay game is going to win that match up. Yes, in terms of player satisfaction, but no, not in terms of sales. Terrible movies with great explosions out-earn great movies with low budgets every single summer.
- baud147258 8y ago
  > Terrible movies with great explosions They usually have bigger marketing than low budget movies.

Kagerjay 8y ago

> 62,176 lines of code then I watched the youtube video demo in 2017. A 2D gaming scroller that requires 62,126 lines of code written from scratch? There's tons of libraries out there, game engines you could adopt, etc. I would argue you could be an awful programmer and use prebuilt tools but have great design aesthetics, vision, music composition, and storytelling, and deliver something far better

gmueckl 8y ago
How can you criticize the amount of code without knowing how much of that is generic engine code and how much of it is gameplay code (game logic, AI, event triggers...) that would have to exist regardless of the engine being used?
- hannasanarion 8y ago
  Because in the former case, it's a waste of time, duplicating work that others have done already available online for free, so of course there are no customers And in the latter case, it's got to be a poorly written mess, no 2d platformer could need so much game logic. That's twice the size of the entire Super Metroid cartrige, engine and assets included.
  - gmueckl 8y ago
    Wait... did you just compare lines of source code to the size of a compiled binary code of something that was written over 25 years ago and clearly optimized for the size of the deliverable? In that case you might just as well have compared an apple to the Apollo program LAM module. It is just not a valid comparison. I am sure that if Super Metroid were to be developed today in a high level language with the exact same gameplay logic, the resulting code would be in at least the same ballpark.
    - hannasanarion 8y ago
      It is a valid comparison, because what he made is essentially the same game. And I said that in response to the guess that maybe the 62,000 lines of code was all game logic, which is absurd. Game logic, like, "when the player presses the fire button, spawn a missile" and "this enemy goes up 5 units, then goes down 5 units, then repeats". That's not the kind of thing that will change much from one metroid clone to another, regardless of what technology is used to make them. It is practically always the smallest part of the final product. And to have the game logic's source be bigger than the entirety of an almost identical game, with handcrafted levels, art, and engine included? It's absolutely absurd. It was in service of the point above: that code is probably all the engine, which means that, yes, it probably was a waste of time, since engines made by professionals are available for free. At best, it's a line on a resume.

LrnByTeach 8y ago

Great analysis.. this applies to so many areas of software development. Building games, websites, any thing software .. > you spent optimizing some C++ function that does something already solved a hundred times in a hundred different game engines. It's a natural tendency for all programmers. > Making g a game in 2018 is far more of a creative endeavor than anything to do with programming really. > You need a massive amount of top notch artwork, music, 3D modelling, shader effects, SFX, etc. to have a polished nice looking game. > That takes either superhuman talent or a large team of specialized people beyond yourself.

derefr 8y ago
You know, there's an interesting relationship between the graphic-novel industry and the TV/film industry at this point. Many TV/film screenplays are adapted from graphic novels, because a graphic novel is, in a sense, a "sketch" of a film—it's something that communicates most of the top-down spirit of the eventual work, while also being something that is able to be produced by a single creator, and so something that is able to let a single creator's vision and talent shine through. I'm left wondering why the video-game industry doesn't have at least some sub-sector with an analogous pipeline, adapting (or in this case, "covering") low-budget indie titles that don't have asset-polish into AAA games, by giving them that asset polish.
- jvanderbot 8y ago
  I'd buy three beers to hear more about this idea. Current acquisition model by AAA is buy-kill, rather than buy-subsidize.
- hannasanarion 8y ago
  Because asset polish doesn't matter nearly as much as gameplay: the high-res remake will never make as much as the original. The most successful game of the year, Far Cry 5, has sold 1/20th as many copies as Minecraft has, a game that is ugly as sin and always has been, but it's fun enough that it doesn't matter. There's no point in adapting a property if the adaptation won't grow the audience. The Walking Dead was a comic book superstar, selling upwards of 500,000 units per issue when it was picked up by AMC for the TV adaptation. The TV adaptation is now reaching 8 million viewers per episode, and that's considered low.

mooseburger 8y ago

It doesn't take a large team. Look at latest indie hotness Hollow Knight (https://youtu.be/UAO2urG23S4 https://youtu.be/UAO2urG23S4), which was made by 3 dudes, who were artists and animators (http://hollowknight.com/our-team/ http://hollowknight.com/our-team/). I agree that with game engines, making a successful indie game is primarily about artistic prowess, but you probably don't need that many artists. Just 1 or 2 really good ones.

chrischen 8y ago

PUBG is evidence of the contrary, though the the game is just pure fun.