|
Unreal seems to come with a lot more in the box, you can click a few times and get some really impressive demos. But once you start digging into making your actual game, they're way behind on documentation and Blueprints are a messy road to go down. Unity's advantages are C# which is a joy to work with and a lot more online references to use. But it feels older (because it is) and seems to have more niggling quirks.
|
#
?
Jun 28, 2016 01:51
|
|
|
|
| # ? Jun 3, 2024 02:31 |
|
|
So now that .Net Core 1.0 is out (and open source), what are the chances we'll see another incarnation of the .Net blueprint integration for the Unreal engine? If I remember correctly, it was licensing issues that killed it off the first time around.
|
|
#
?
Jun 28, 2016 02:47
|
|
|
rarbatrol posted:So now that .Net Core 1.0 is out (and open source), what are the chances we'll see another incarnation of the .Net blueprint integration for the Unreal engine? If I remember correctly, it was licensing issues that killed it off the first time around. Mono's been free for awhile now, surprised it hasn't already been revived(though I don't really follow UE stuff close).
|
|
#
?
Jun 28, 2016 02:59
|
|
|
I have a background and a player sprite, and I've started to make a move script for both Runners, and Lone Star to inherit. The move script is stolen almost entirely from the Unity Tutorial, I'm a bad programmer.  I'm using a 12x12 Open Source font, I'm planning later to inherit color and transparency from whatever the thing is. (Lone Star Will be L's, Infenal Hounds will be H, Shaman's S, etc. etc.) It's not a game but it's start for a couple of hours of work.
|
|
#
?
Jun 28, 2016 03:21
|
|
|
Can't remember if this was mentioned here or not but this is an awesome series of videos on Unity3d that I stumbled upon yesterday: http://www.youtube.com/channel/UCEklP9iLcpExB8vp_fWQseg xgalaxy fucked around with this message at 03:43 on Jun 28, 2016 |
|
#
?
Jun 28, 2016 03:41
|
|
|
Obsurveyor posted:Mono's been free for awhile now, surprised it hasn't already been revived(though I don't really follow UE stuff close).
|
|
#
?
Jun 28, 2016 05:02
|
|
|
Shalinor posted:IIRC, it was less the licensing, and more that they didn't want to get stuck supporting two independent scripting languages. They'd have to officially support both Blueprint AND C#, and make sure they were both kept in parity, and - etc. They chose to double down on Blueprint instead. There's nothing preventing a 3rd party from adding it, the programmer behind the C# experiment even offered some basic explanations of what needed to be done (and possibly source? can't remember), it's just that said things were "hack UE4 so bad that you can kiss staying in sync with updates goodbye." You can watch this thread: https://forums.unrealengine.com/showthread.php?106036-Mono-relicensed-under-MIT-C-for-UE4-can-come-back/page3 It doesn't look like anyone from the Mono/Xamarin/MS side feels like stepping up to complete this. It wouldn't be that bad as it would just call the blueprint endpoints but someone has to do the leg work/continue support. Back when Xamarin was still charging it looked like it might make sense since they were going to charge, now that it's free it seems like no one is willing to do the work. Unity has come a long way since UE4 was released (I'd suggest 5.4B21 not B22) and now is getting close to equal footing in some areas.
|
|
#
?
Jun 28, 2016 11:47
|
|
|
Shalinor posted:Using Unreal. You literally described the polar opposite of the sort of project that it's easy to convince UE4 to produce. You really want an engine better suited to something procedural / freeform / weird. Unity's a fine choice as a default, though it's far from the only one. If there's one dev I want to pick their brains over Unreal it's http://astroneer.space/ they seem to be doing everything people say is very difficult to do in UE4, and they're not a large team. https://www.youtube.com/watch?v=l-ZoIIGrWFc
|
|
#
?
Jun 29, 2016 00:09
|
|
|
BabelFish posted:If there's one dev I want to pick their brains over Unreal it's http://astroneer.space/ they seem to be doing everything people say is very difficult to do in UE4, and they're not a large team. EDIT: ... and just to cover my bases, if it turns out they somehow did that all high level, mostly in Blueprint, my immediate thoughts would be thus:  BabelFish posted:I think they're two engineers and two artists who (at the times of that video) had been using the engine for a year? It's impressive regardless. Hopefully I'll run into one of their guys at a meetup some day, Seattle isn't THAT big. drat, gotta get to the next local dev meetup. So many cool people here. 
Shalinor fucked around with this message at 03:23 on Jun 29, 2016 |
|
#
?
Jun 29, 2016 01:45
|
|
|
Shalinor posted:I would lay odds their team is almost all low-level programmer types, who basically just wanted a multiplatform graphics framework (with full source) and were happy to throw any of the rest out that got in their way. Of course I have absolutely no idea if that's true, but given the tech they chose and the sort of game they were aiming for... seems like a good bet. It's certainly not a game type you'd start if you were worried about the time cost of low-level programming / all-custom tech. I think they're two engineers and two artists who (at the times of that video) had been using the engine for a year? It's impressive regardless. Hopefully I'll run into one of their guys at a meetup some day, Seattle isn't THAT big.
|
|
#
?
Jun 29, 2016 02:15
|
|
|
Shalinor posted:IIRC, it was less the licensing, and more that they didn't want to get stuck supporting two independent scripting languages.
|
|
#
?
Jun 29, 2016 07:00
|
|
|
OneEightHundred posted:From their responses, it was definitely a licensing issue. They didn't want something like a C# integration becoming de-facto standard unless they could assume the development of it, otherwise UE4 would become dependent on Xamarin continuing to offer acceptable licensing terms. That used to be the case, but then .net went open source and I think that's why people wanted it again.
|
|
#
?
Jun 29, 2016 07:42
|
|
|
 A few years back I developed a Unity3D-ish engine clone for a small company to try and snatch up source code sales Unity weren't supplying, at the last moment the CEO decided we would actually be going toe-to-toe with Unity instead, my team quit (both of us) and so the project folded. Anyway, due to financial destitution I've been thinking of revisiting the idea with a different approach, I've worked on a few Unity/Unreal projects since and I find the same problems still exist. Needing to use an engine I downloaded UE4 and instantly found myself immediately overloaded with interface, with no idea how to do what I want quickly. This is despite being an Unreal veteran since version 2, and been lead dev on a UE3 game where I heavily modified the source. Back when first starting to use UE3 it took a few days to for me to even start working out how to make a game with it, I ended up constructing an UnrealScript framework to do what I wanted. Faced with doing the same thing in UE4, I immediately uninstalled it and re-opened Unity3D. I've recently left a Unity3D horror show so issues there are still very fresh in my mind, I've been thinking what I would do to fix them. * Bugs. Using Unity 5 I found the lighting was next to useless. Enlighten produced awful results and brought my machine to a halt for 3 hours while doing it. A lot of forum Unity forum posts suggested the lighting was pretty much broken currently and there are no plans to fix it. So coming up with a workable PBR pipeline in comparison is probably doable. * The shell development approach is quite at odds with how a lot of game development still happens. Git is designed for the command line workflow, look at the log, make changes, commit regularly, rebase regularly, undo with the reflog etc. However the CTO of the horror show had decided to go with Git without having any experience of it, so the workflow was to do stuff in Unity and then wrangle Git though a GUI that lacked basic commands, then teach the artists and designers to do the same. Unsurprisingly they told me Git had lost a lot of their work when I joined, hmm. Small mobile games companies these days don't want the cost and overheads of Perforce so the Linux dev world and game dev are colliding in an unpleasant way. It's been a long time since game dev has been the domain of Windows, with most Unity3D users preferring OSX so better shell development integration is possible. I'd for like a new engine I create to have the development centered around the Git workflow, its better to get that right early on than have a hash up later. Sure, the barrier to entry is higher but Unity still exists. * As projects become larger use for the Unity GUI tools become a lot less and projects become code driven, I'd like to start with a fully code driven interface like Futile and specifically cater for more dynamic-ness. The horror show company was producing a Boom Beach knockoff so there was no need to place things in the editor, and as Shalinor mentioned somewhere above UE4 finds itself very limited with it's FPS, static level orientation. As far as scripting goes, probably a JavaScript VM with a TypeScript framework. That makes the scripting close to C# and easy to build HTML5 versions. I'll probably go with mixins because I think that's a better way to blend static and dynamic typing over the component model, although TypeScript currently needs some hacking to support mixins better. Anyway dear SA forum users, let me know if you would use such an engine and I'll arrogantly ignore everybody's suggestions and go ahead, creating my dream white elephant project nobody will else will ever use Derek Smart style. Just joking, what could I add or remove from that short list? darkpool fucked around with this message at 08:22 on Jun 29, 2016 |
|
#
?
Jun 29, 2016 08:04
|
|
|
darkpool posted:
 !I remember this, actually! I was down that you had to scrap it or couldn't release how you originally wanted or etc, forgot the exact reason, but I'm glad you're working on something similar again. I could maybe use an engine like this, I'm always not fond of too much reliance on an editor and prefer working with a slightly more-raw-but-not-100%-raw framework most of the time (in 2D I usually just go with "It can draw textures, use shaders + render targets, with a lot of helper libraries such as vectors & matrices". First thing I usually do is build some kind of "Transform" class similar to Unity, though). Also every major Unity project I've seen has always boiled down to people saying "we cut out the editor and the physics and the MonoBehaviour and only used it as a rendering engine" in the end. If it was C# or maybe C++ I'd probably use it for some stuff? Jewel fucked around with this message at 14:19 on Jun 29, 2016 |
|
#
?
Jun 29, 2016 12:52
|
|
|
darkpool posted:
I would not use such a one, or recommend its use to others. The last thing I want in my games projects is to deal with a bunch of JavaScript BS. It also wouldn't be a very popular engine, so finding experienced people would be impossible. Do you have a marketing strategy to pull people away from Unity or Unreal (or Lumberyard, or extant game engines people aren't using)? Do you have the budget to implement that strategy?
|
|
#
?
Jun 29, 2016 13:15
|
|
|
darkpool posted:
Lighting is horrible in Unity 5 so many people are opting for forward rendering and no baked lighting. The stable released for Unity since 5.2 have been pretty solid. Game dev is absolutely the domain of Windows and generally always has been, I'm really unsure what you are talking about. AAA is almost all Windows from what I hear and most non iPhone devs work on Windows machines because we want a good GPU. Unity has transitioned from being a OSX first product since 2, to also just kind supported on OSX on 5. Every game engine is nearly free now, most with source code access. I think you'd be nuts to want to enter the market. If you want to open source your project that would be cool we can always use more engines. Amazon isn't even trying to make money off of it's engine except to encourage people to use Twitch and AWS.
|
|
#
?
Jun 29, 2016 15:48
|
|
|
darkpool posted:
Are you implying that your engine will support 3D games online with no additional plugins to be installed? That's pretty cool, and I like the mindset of all code first. Is this engine going to be close to Unity's (built in rendering/components/monobehaviour-esq thing)? Good luck, I would use it just to see if I could make a web game. Convincing Unity/Unreal peeps to try a JS engine will be hard, but if TypeScript cuts out the anonymous type part of JavaScript it might be an easier sell.
|
|
#
?
Jun 29, 2016 15:58
|
|
|
I frankly, think you'd be better off focusing on plugins/extensions for unity/unreal.
|
|
#
?
Jun 29, 2016 16:06
|
|
|
Good reasons I can think of for not using unreal/unity would be: * You want to directly compete with them, a noble endeavor but requires a lot of people an investment. Neither of them are perfect after all. * You don't want to pay them for whatever reason * A substantial number of features like their cross platform compatibility are irrelevant to you Remaking unity/unreal for a single or few projects needs a lot of justification. All of their problems are likely to still exist in your own editor along with a ton of problems they already solved.
|
|
#
?
Jun 29, 2016 23:10
|
|
|
Stick100 posted:Game dev is absolutely the domain of Windows and generally always has been, I'm really unsure what you are talking about. AAA is almost all Windows from what I hear and most non iPhone devs work on Windows machines because we want a good GPU. Unity has transitioned from being a OSX first product since 2, to also just kind supported on OSX on 5. I'm an iPhone dev and I still work on Windows. No way I'm using anything but Visual Studio to write code, giving up Perforce + Windows Explorer integration, or making a dozen other sacrifices to work on a Mac. I keep a Mac Mini under my desk to make builds for iOS devices and do nothing else.
|
|
#
?
Jun 29, 2016 23:31
|
|
|
Could someone point me towards an introduction to procedural generation?
|
|
#
?
Jun 30, 2016 00:33
|
|
|
darkpool posted:Just joking, what could I add or remove from that short list? * Cross platform (including web) easy to use 2D graphics. * Cross platform input devices, with, where possible, all input captured with precise timestamps, not just the common "things that happened this frame" or the slightly better "things that happened this frame, in the order in which they happened". (Falling back to the worse options with made up approximate timestamps where the better options aren't available.) * Minimal builds (no 50MB dlls or js imports or whatever). * Not slow. * Optional (I don't mind writing or integrating my own physics really): choice of fast / precise / deterministic physics engine integration. These are all areas of weakness for Unity and Unreal.
|
|
#
?
Jun 30, 2016 01:47
|
|
|
Music Theory posted:Could someone point me towards an introduction to procedural generation?
|
|
#
?
Jun 30, 2016 02:02
|
|
|
Shalinor posted:Wang tiles are a pretty good starter: http://procworld.blogspot.com/2013/01/introduction-to-wang-tiles.html That reminded me that I've already read that article, and in fact have that website bookmarked.  (Thanks though)
|
|
#
?
Jun 30, 2016 03:37
|
|
|
Music Theory posted:That reminded me that I've already read that article, and in fact have that website bookmarked. http://natureofcode.com is not bad http://galaxykate0.tumblr.com/post/139774965871/so-you-want-to-build-a-generator is a nice 'intro' article
|
|
#
?
Jun 30, 2016 03:40
|
|
|
Shalinor posted:Wang tiles are a pretty good starter: http://procworld.blogspot.com/2013/01/introduction-to-wang-tiles.html This isn't what I was hoping for at all.
|
|
#
?
Jun 30, 2016 03:44
|
|
|
Jewel posted:Also every major Unity project I've seen has always boiled down to people saying "we cut out the editor and the physics and the MonoBehaviour and only used it as a rendering engine" in the end. Yep, that's what I'm going for. The engine will be C++ and bindable to whichever scripting VM; JS, Lua, Mono, etc. leper khan posted:I would not use such a one, or recommend its use to others. The last thing I want in my games projects is to deal with a bunch of JavaScript BS. It also wouldn't be a very popular engine, so finding experienced people would be impossible. Do you have a marketing strategy to pull people away from Unity or Unreal (or Lumberyard, or extant game engines people aren't using)? Do you have the budget to implement that strategy? I'm not sure that many people mind if their C++/C#/TypeScript gets compiled to JS and executed in V8 under the hood, if it really bothered a lot potential users I'd consider a different route. There are millions of former Flash developers out there who did develop games in an ECMA script variant, so I don't think the issue would bother them at all. I have a small series A funding, but we're talking open source here because I have customers for specific projects using the engine I'll make. I'm more concerned about dealing with issues in the development industry that I've come across lately than I am business strategy. Stick100 posted:Lighting is horrible in Unity 5 so many people are opting for forward rendering and no baked lighting. The stable released for Unity since 5.2 have been pretty solid. I should have said exclusively the domain of Windows and I'm definitely not talking AAA here, but I did just look up the stats and you're right, there has been a big shift to Windows with newer versions of Unity. I've contracted for or worked at 5 different Unity outfits of 40+ employees since 2010 and they've all been exclusively OSX, so I'll have to dig further in to the stats to work out the Windows/OSX split for pro and indie and see if I just have anecdotal bias. My focus is large professional teams, as that's where I've experienced Unity not scaling well. This engine will be free and open source. Nude posted:Are you implying that your engine will support 3D games online with no additional plugins to be installed? That's pretty cool, and I like the mindset of all code first. Is this engine going to be close to Unity's (built in rendering/components/monobehaviour-esq thing)? Good luck, I would use it just to see if I could make a web game. Convincing Unity/Unreal peeps to try a JS engine will be hard, but if TypeScript cuts out the anonymous type part of JavaScript it might be an easier sell. Yes, exactly. I wouldn't suggest anybody use JavaScript directly for a big project, I'm a big static checking enthusiast and I want more errors to caught at compile time not less. Most people love compositional inheritance and that's been a big part of Unity's success, the problem I expereince though is the mix of dynamic and static typing, GetComponent can return null at runtime and the compiler can't help if you need to change the type of a bunch of GameObjects and you missed some. So my suggestion here is to have a mixin definition to create static GameObject types up front so the compiler can catch more errors upfront. This is doable with JS running under the hood, but not so much with .NET/Mono. Triarii posted:I'm an iPhone dev and I still work on Windows. No way I'm using anything but Visual Studio to write code, giving up Perforce + Windows Explorer integration, or making a dozen other sacrifices to work on a Mac. I keep a Mac Mini under my desk to make builds for iOS devices and do nothing else. There would specifically be no restriction on code editor or build setup. roomforthetuna posted:I'd use this engine if you removed everything from that list, and added: You can assume all of these, except for 2D, there are plenty of good 2D engines out there. Unormal posted:I frankly, think you'd be better off focusing on plugins/extensions for unity/unreal. Not my field of expertise really, most jobs I do involve throwing away almost everything from Unity/Unreal to work with them, so I figure why not make my life easier by throwing the last bit away too? darkpool fucked around with this message at 04:18 on Jun 30, 2016 |
|
#
?
Jun 30, 2016 03:46
|
|
|
aegof posted:This isn't what I was hoping for at all. "Ah, yes, I see your Wang Manager is terribly optimized. This will never handle the sheer number of Wangs you'll have in a finished game. Right now, though, you're seeing glitches with your Wangs mostly because you need more Wang Tiles. You see wangs are defined by their number of exits - I mean shape, wangs have a basic shape, you know? - and that shape indicates which wangs fit together. Wangs are just hard in general, don't worry, you'll get it working..." Shalinor fucked around with this message at 16:00 on Jun 30, 2016 |
|
#
?
Jun 30, 2016 15:58
|
|
|
Music Theory posted:Could someone point me towards an introduction to procedural generation? What type of procedural generation? There's, like, procedural mesh generation stuff (like what gets used to make the huge areas of land in Elder Scroll games), then there's more simple tile based stuff (like what people use to make roguelikes and roguelites), and then there's just intelligent random placement (like making monsters and items spawn in random places on a premade map and making sure the things chosen are balanced and appropriate). In general, my big advice is to basically treat it like you just have a big deck of hand-made content and then the generator is just a way to semi-randomly deal out that content whenever you need it.
|
|
#
?
Jun 30, 2016 18:17
|
|
|
Music Theory posted:Could someone point me towards an introduction to procedural generation? Try chapter 1 of http://pcgbook.com/
|
|
#
?
Jun 30, 2016 18:38
|
|
|
Shalinor posted:Everyone should use Wang tiles. Not so much because it's a good algorithm (though it is), but because it guarantees a huge swathe of your code can't be taken even remotely seriously. "Also, this is running really slow. What sort of PC is this?" 
|
|
#
?
Jun 30, 2016 23:03
|
|
|
roomforthetuna posted:choice of fast / precise / deterministic physics engine integration. Is there such a thing (for rigid body dynamics)?
|
|
#
?
Jun 30, 2016 23:41
|
|
|
Vlad the Retailer posted:Is there such a thing (for rigid body dynamics)? Infinite precision math with a fixed time step? Maybe I'm missing something though.
|
|
#
?
Jun 30, 2016 23:50
|
|
|
You don't need infinite precision, just the fixed timestep. Bog standard floating point is completely deterministic. Without a fixed timestep it's not deterministic because one of the inputs is randomly varying (the delta time).
|
|
#
?
Jul 1, 2016 02:49
|
|
|
Bog standard floating point is deterministic on a single compiler and OS, but it you're going to have a complete nightmare in practice making the results match up across platforms due to compiler optimisations and reordering of operations unless you use fixed-point math instead.
|
|
#
?
Jul 1, 2016 12:57
|
|
|
Yeah, true.  I'm purely talking in theoretical terms.
|
|
#
?
Jul 1, 2016 16:30
|
|
seiken posted:Bog standard floating point is deterministic on a single compiler and OS, but it you're going to have a complete nightmare in practice making the results match up across platforms due to compiler optimisations and reordering of operations unless you use fixed-point math instead. In how many modern languages is fixed-point even an option? I think Python supports fixed-point operations, but are there many others that wouldn't require you to write your own implementations?
|
|
|
#
?
Jul 1, 2016 20:36
|
|
|
Centripetal Horse posted:In how many modern languages is fixed-point even an option? I think Python supports fixed-point operations, but are there many others that wouldn't require you to write your own implementations? There are fixed point libraries for pretty much everything.
|
|
#
?
Jul 1, 2016 20:41
|
|
|
If I want to develop a 2D game with aspirations of eventually putting it on steam and I hate programming in C, where should I start? Game would include sonic style movement. Looking for suggestions on engines/frameworks.
|
|
#
?
Jul 1, 2016 21:55
|
|
|
|
| # ? Jun 3, 2024 02:31 |
|
|
Benson Cunningham posted:If I want to develop a 2D game with aspirations of eventually putting it on steam and I hate programming in C, where should I start? Game would include sonic style movement. Looking for suggestions on engines/frameworks. Unless by 'C' you also mean C#, Unity.
|
|
#
?
Jul 1, 2016 21:59
|
|
