It seems that very few indies build there own engine.

[Radical Left Lunatic]

There are very few indie titles coming out that are using there own engine built from scratch.This could be to that fact that designing you own engine requires a lot more low level CS knowledge but in reality is not really that hard.There are quite a few good tutorials on it such as the tutorials at rastertek.com and yet people still choose to use commercial engines such as unity,cry engine ,unreal engine and source.You would think that they would want there own royalty free engine, by designing you own engine you game is royalty free.A common theme i see among indie developers is that they can that and this feature becuase they are using the free version of some software, visual studio is free, DX SDK is free,C++ is free.If you wat a feature that does not exist in your engine simply add it, if your performance is low you have full control over it and can optimize it at a low direct hardware level.And the best part is you dont need to learn how to use your own engine and it accepts what ever model formats you want it to, its really great.Not to mention its fun, i have tried engines such as unity and never really got around to learning how to do it.Using an engine like unity was boring and tedious, if something was not working i had no idea how to fix it becuase i did not make it.If theres a bug in your own engine, it is 100% your fault and you will always be able to fix it.

Anyone elses thoughts?

[Tux]

引用自 zoomdude111：

引用自 Tux：

in all fairness I doubt your customized engine is as good as even Unity.

I am a programmer myself and I know programmers who THINK they can outcode a team of 10 over the course of 10 years and do it part time within a year but in reality they just haven't thought it all thru yet.

also...there is a quote floating around the HTML5 universe, its 'there are more HTML 5 engines then there are developers making actual content on HTML5'

good developers use the tools that already exist

No offence taken, i dont expect to make something as good as unity.My engine is going to be made to run specific games, it will probaly have a simple editor but it wont be like unitys interface with tons of complicated options and features.The interface will be more like a level editor that saves to a custom format that the engine loads.I plan on saving the engine(once it is finsihed) to a library that i can use to make what ever applications i want(not just games).I feel i have acomplished a good amount in the time since i started, here is a link to my github in case you are intrested https://github.com/JustinWeq/Test-Engine .

I guess the question at this point I would have to ask is.

why?

I mean why create a custom engine?

[Radical Left Lunatic]

引用自 Tux：

引用自 zoomdude111：

No offence taken, i dont expect to make something as good as unity.My engine is going to be made to run specific games, it will probaly have a simple editor but it wont be like unitys interface with tons of complicated options and features.The interface will be more like a level editor that saves to a custom format that the engine loads.I plan on saving the engine(once it is finsihed) to a library that i can use to make what ever applications i want(not just games).I feel i have acomplished a good amount in the time since i started, here is a link to my github in case you are intrested https://github.com/JustinWeq/Test-Engine .

I guess the question at this point I would have to ask is.

why?

I mean why create a custom engine?

Custom engines are 100% royalty free, you can do what ever you want with it.They can also be very light weight and run on very low end machines unlike unity which depends more higher technical specs.

[On The Wall Entertainment]

I would say it is good to create an engine yourself to learn how things work and why things should be done a certain way. It gives the developer the freedom to do as they wish, which may or may not be for the best. Biggest downside I see with a third party engine would be an update that breaks how the developer was using something.

With that being said, I think one would be better off using a third party engine. Multi-platform support and advanced real time graphic manipulations. The time and frustration of these aspects would be almost non-existent.

I made the engine for Wills and Wonders myself and it really helped to understand the hows and the whys behind the scenes. Building rules for networking early to make a transfer to server side multiplayer possible. I cannot see doing it any other way but I believe it could all be done using a third party engine with the same understanding; with less effort.

http://steamcommunity.com/sharedfiles/filedetails/?id=379571772


Edited to make my points more clear.

[Tux]

引用自 Willfully Whimsical Wizard：

I would say it is good to create an engine yourself to learn how things work and why things should be done a certain way. It gives the developer the freedom to do as they wish, which may or may not be for the best. Biggest downside I see with a purchased engine would be an update that breaks how the developer was using something.

With that being said, I think one would be better off using an engine. Multi-platform support and advanced real time graphic manipulations. The time and frustration of these aspects would be almost non-existent.

I made the engine for Wills and Wonders myself and it really helped to understand the hows and the whys behind the scenes. Building rules for networking early to make a transfer to server side multiplayer possible. I cannot see doing it any other way but I believe it could all be done using an engine with the same understanding with less effort.

http://steamcommunity.com/sharedfiles/filedetails/?id=379571772

as a proffesional developer of 9 years myself I would have to completely disagree.

Its no more true that its good for me to learn machine lanauage in order to create a webpage

[Tux]

引用自 zoomdude111：

引用自 Tux：

I guess the question at this point I would have to ask is.

why?

I mean why create a custom engine?

Custom engines are 100% royalty free, you can do what ever you want with it.They can also be very light weight and run on very low end machines unlike unity which depends more higher technical specs.

as someone who has been working as a developer profffesionally for about 9 years I would have to say the exact opposite is what will make you a great developer.

I assure you 100000%

[kdodds]

The most optimal engine for a game is one written specifically for that game (given like talent, of course), with no more or less code than necessary. With that said, engine coding takes a huge portion of dev man hours, and getting the game to market may require utilizing existing tools (including engines) versus not developing the game at all due to limited resources. Great games can be developed either way, of course. But if it comes down to either getting your idea out there by utilizing an existing engine, or not getting it out there at all, I'd say use the existing engine.

And yes, I have to agree that there are a LOT of young coders out there who think they can do a 10 man, 10 year job in 10 weeks, and I also agree that they're always mistaken as they've failed to consider all of the variables due to lack of experience, not the least of which is the fact that the engine requires more than just display (the usual main focus).

[Radical Left Lunatic]

引用自 kdodds：

The most optimal engine for a game is one written specifically for that game (given like talent, of course), with no more or less code than necessary. With that said, engine coding takes a huge portion of dev man hours, and getting the game to market may require utilizing existing tools (including engines) versus not developing the game at all due to limited resources. Great games can be developed either way, of course. But if it comes down to either getting your idea out there by utilizing an existing engine, or not getting it out there at all, I'd say use the existing engine.

And yes, I have to agree that there are a LOT of young coders out there who think they can do a 10 man, 10 year job in 10 weeks, and I also agree that they're always mistaken as they've failed to consider all of the variables due to lack of experience, not the least of which is the fact that the engine requires more than just display (the usual main focus).

I have alreay considered the things other then graphics and have done my research on what i need and dont need for this project.I am using octtrees for collisions and faster frustrum culling, most likly also going to be using SAT for actual collision feedback.Also i have alreay used path finding algorithms such as A* so i have that down, furthur more i also know how picking works(raycasting) for selecting objects the cursor is over.On top of all this i have a strong knowledge of how the game structure works(init,(update,draw),cleanup,exit) with both a dynamic and liited frame rate.

[Radical Left Lunatic]

引用自 Willfully Whimsical Wizard：

I would say it is good to create an engine yourself to learn how things work and why things should be done a certain way. It gives the developer the freedom to do as they wish, which may or may not be for the best. Biggest downside I see with a third party engine would be an update that breaks how the developer was using something.

With that being said, I think one would be better off using a third party engine. Multi-platform support and advanced real time graphic manipulations. The time and frustration of these aspects would be almost non-existent.

I made the engine for Wills and Wonders myself and it really helped to understand the hows and the whys behind the scenes. Building rules for networking early to make a transfer to server side multiplayer possible. I cannot see doing it any other way but I believe it could all be done using a third party engine with the same understanding; with less effort.

http://steamcommunity.com/sharedfiles/filedetails/?id=379571772


Edited to make my points more clear.

That looks like an amazing game, i voted.I have always ben a fan of turn based stratgy

[kdodds]

引用自 zoomdude111：

引用自 kdodds：

The most optimal engine for a game is one written specifically for that game (given like talent, of course), with no more or less code than necessary. With that said, engine coding takes a huge portion of dev man hours, and getting the game to market may require utilizing existing tools (including engines) versus not developing the game at all due to limited resources. Great games can be developed either way, of course. But if it comes down to either getting your idea out there by utilizing an existing engine, or not getting it out there at all, I'd say use the existing engine.

And yes, I have to agree that there are a LOT of young coders out there who think they can do a 10 man, 10 year job in 10 weeks, and I also agree that they're always mistaken as they've failed to consider all of the variables due to lack of experience, not the least of which is the fact that the engine requires more than just display (the usual main focus).

I have alreay considered the things other then graphics and have done my research on what i need and dont need for this project.I am using octtrees for collisions and faster frustrum culling, most likly also going to be using SAT for actual collision feedback.Also i have alreay used path finding algorithms such as A* so i have that down, furthur more i also know how picking works(raycasting) for selecting objects the cursor is over.On top of all this i have a strong knowledge of how the game structure works(init,(update,draw),cleanup,exit) with both a dynamic and liited frame rate.

And you're a one man team, right? Well, good luck with your project. Let us know when it's released.

[Radical Left Lunatic]

引用自 kdodds：

引用自 zoomdude111：

I have alreay considered the things other then graphics and have done my research on what i need and dont need for this project.I am using octtrees for collisions and faster frustrum culling, most likly also going to be using SAT for actual collision feedback.Also i have alreay used path finding algorithms such as A* so i have that down, furthur more i also know how picking works(raycasting) for selecting objects the cursor is over.On top of all this i have a strong knowledge of how the game structure works(init,(update,draw),cleanup,exit) with both a dynamic and liited frame rate.

And you're a one man team, right? Well, good luck with your project. Let us know when it's released.

As far as making my engine goes that is correct, i am not going to be making any content for it though.Probaly build a couple of simple games that tests its performance and graphics.

[Radical Left Lunatic]

引用自 Gremmerz police officer™：

Unless your game is very basic = requires a very simple basic engine you will conserve a lot of effort, time and even money by using an existing engine that is constantly updated, have hundreds or even thousands of users who report bugs which are corrected by the dev team and has tons of source material (guides, support articles, etc.) also including assets store (mods, addons, effects, scripts, even 3D models, GUIs and GUI scripts, and so on and so forth) out of the box.

Good luck achieving the same... unless if your goal is some minimalistic game and engine that is too simple to have real bugs and require minimal time, effort and budget to complete OR if you are making 2D games and engines which are quite easy to make even with added 3D visual effects and such. Every other case you are better off with an existing engine.

Don't believe me? Sure. Check around at the major gamedev portals like gamedev.net and alike and do your research. Also sniff around the game engine portals / related forums as well. You will see much more than enough evidence and reasoning there.

Also remember the good ole' saying: "If not broken no need to repair it"... or in this case: "If not broken no need to reinvent (the wheel)".

You sound completely like the "first game to be MMO" kinda people who pop up each day on every gamedev related site and message boards claiming they will make the new WoW and whatnot, claim they have done plenty of research, pretend they know what the undertaking will be and how much effort, time and budget it needs but in the end they all will face the harshness of reality like how a person faces the same when hit by a train, and when that happens facts will result in these overambitious and completely ignorant devs to either run away from game development for the rest of their lives or be hurt permanently and have that wound never to heal... only a few I saw who take it well and learn from it, people who are ignorant (out of pride, out of thinking they are better and/or smarter than the rest and unable to face that is not the case) will never recover their shattered ego.

No amount of useful advice, facts referencing or explanation will convince them, they 99% of times just ignore it and pretend as if they know better just to later realize how wrong they were. But looking back to their accounts years later, doing some research and checking their project pages always prove that facts and reality can not be argued with and their projects fail all the time... for obvious reasons.

You seem to be fixated on your approach just as much so I don't think anyone or anything will convince you otherwise so I don't even recommend you the most basicest and yet most useful of approaches: use common sense. You will learn this eventually. Smart people learn from other people's mistakes, while others doom themselves to learn from their own.

Yet I recommend you something that may help you out still: do some research (much beyond what you already did), this may open your eyes. I frankly doubt it and you seem much more stubborn already than to change your fixed mindset

Either case good luck.

Oh and be sure to come back once you have a completed project even if just a fully working and fluid 3D game engine you made because for that you need nobody else, no assets (aside of placeholders / test assets which can be acquired for free) and no budget only spare time you already mentioned you have during summer break or something like that which should be enough. Show us how easy and reasonable it is compared to using an existing engine.

I already have a working example of a basic 3D engine, it has basic lighting,2D and 3D rendering,alpha blending,a terrain engine(uses height maps),a sound interface(direct sound),a input interface(directinput) and a working example of a camera you can navigate the scene with(makes debugging easy) i now how to use tools such as intels GPA and VS own built in graphics performance anaylyzer.Even with a scene with milloins of trianlges and no pruning or extra culling what so ever i still get almost 2000 FPS.It also has AA,Full screen and borderless window support.When i work on it i can easily pump out a couple thousand lines per day of working effective code, i have done my research and it has taken me years of dedication and a couple dozen computer science courses to get to the point where i can even attempt such a engine.Trust me i have done me research and have looked at websites such as gamedev.net and code project.The two most valueble resources by far have been rastertek.com and braynzersoft.com (i my have the name wrong there), the are well written and i still refer to them for reference and to see just what can be acomplished.I have built game engines already in XNA before and well low level things like graphics where partially taken care of(still had to write my own shaders) i had to handle the game logic and physics myself.Even though i am inexperienced i do have engough knowledge to know what i can and can not do.I still have a ways to go and plan on implementing things like post processing and a nice collision system.I also know how to write my own scripting languages so having objects loaded into the scene is a peice of cake, i simply write a parser for what ever format its in and load objects into the scene(use hardware instancing to draw multiple of the same model at one time).