|
I'm glad I can't tweet about work because it would be 100% bitching about UE4.
|
#
?
Oct 1, 2015 05:02
|
|
|
|
| # ? May 29, 2024 01:52 |
|
|
If I recall correctly a Hollow Knight dev used to be active here on SA; RockPaperShotgun just posted a neat article featuring their concept art. Good stuff!
|
|
#
?
Oct 1, 2015 11:53
|
|
|
Omi no Kami posted:I agree that it is, I just don't like it.  BTW, if anyone IS doing the static asset thing, stop. What you're doing forces EVERY SINGLE ASSET YOU REFERENCE to load the second the exe starts. Creates a hellaciously bad stall you can do literally nothing about. Use the drat Blueprints to create a layer of dis-assocation between your code and assets. You don't have to do any real scripting in them, just use them to specify meshes and textures and the like. It'll make your work more usable to any artists/designers on your team, and make it easier to tweak stuff in testing. (also, you should be exposing your stuff to Blueprint anyways, if you want to support modding - learn the pattern that lets you do everything in C++, but have those functions overridden in BP by modders down the road) Shalinor fucked around with this message at 16:06 on Oct 1, 2015 |
|
#
?
Oct 1, 2015 16:03
|
|
I don't know why, but I have a really irrational dislike of Blueprint- if I'm in my office, and I'm in a bad mood, there's a better-than-average chance I've been doing something with Blueprint. I don't think it's a rational reaction, but it's there anyway.
|
I'm on the very opposite end of the spectrum-on the plus side, I'm hyperaggressive about keeping assets out of memory until I need them, but on the extremely negative side I've spent hours improving functions that don't run more than 2-3 times/second by tenths of a millisecond, just because I look at the math and go "...you know, I'll bet I could cut that down from 6 operations to 4".  Edit: thought up the joke too late, but the correct answer is "I don't like to pay for my assets with cache". Omi no Kami fucked around with this message at 18:01 on Oct 1, 2015 |
|
#
?
Oct 1, 2015 17:07
|
|
|
I haven't bothered to check out how to do UMG in C++, but that's mostly because I wonder how crazy would you be to want to do that. Do any complicated calculations in the C++ sure, but expose that poo poo and move those widgets with blueprints.Shalinor posted:BTW, if anyone IS doing the static asset thing, stop. What you're doing forces EVERY SINGLE ASSET YOU REFERENCE to load the second the exe starts. Creates a hellaciously bad stall you can do literally nothing about. Use the drat Blueprints to create a layer of dis-assocation between your code and assets. You don't have to do any real scripting in them, just use them to specify meshes and textures and the like. It'll make your work more usable to any artists/designers on your team, and make it easier to tweak stuff in testing. I haven't yet needed to look into asset loading yet, but what's the pattern here? Use the BP actor spawn event to actually load and create the assets?
|
|
#
?
Oct 1, 2015 20:42
|
|
|
dupersaurus posted:I haven't bothered to check out how to do UMG in C++, but that's mostly because I wonder how crazy would you be to want to do that. Do any complicated calculations in the C++ sure, but expose that poo poo and move those widgets with blueprints. Expose the asset/component reference in an EditAnywhere or EditDefaultsOnly UPROPERTY. Then depending on if your actor is placed in a level or spawned at runtime, set the value of that property in the level or in the defaults of a blueprint child class of your C++ class.
|
|
#
?
Oct 1, 2015 21:01
|
|
|
Is this asset loading thing basically one of those things where every tutorial you find online is doing it wrong?
|
|
#
?
Oct 1, 2015 22:26
|
|
|
dupersaurus posted:I haven't bothered to check out how to do UMG in C++, but that's mostly because I wonder how crazy would you be to want to do that. Do any complicated calculations in the C++ sure, but expose that poo poo and move those widgets with blueprints. I built the UI for the very first published UE4 game, way back before we had UMG. UMG is built on top of Slate which uses its own special syntax to define layout (look here for some examples.) I had a controller layout page that was, I kid you not, >2000 lines of brackets (though part of that was we didn't have absolute positioning, offsets, or spacers back then.) Had to write a custom parser to keep myself from getting lost in it. Animations were based off FCurveSequences and this was over a year before they got hot reload in the engine, so any change required a build. Trying to iterate was a complete nightmare. There are several things I'd really like to have in UMG (masks being one), but overall it's an amazing improvement from the old system.
|
|
#
?
Oct 1, 2015 23:16
|
|
|
xgalaxy posted:Is this asset loading thing basically one of those things where every tutorial you find online is doing it wrong? Not even joking. I don't think I've seen a single tutorial that does it the right way, which is... Bizarro Buddha posted:Expose the asset/component reference in an EditAnywhere or EditDefaultsOnly UPROPERTY. Then depending on if your actor is placed in a level or spawned at runtime, set the value of that property in the level or in the defaults of a blueprint child class of your C++ class. Whatever you want to spawn, make a BP class out of it. Inside that BP is where you do stuff like assign your meshes and other assets (instead of loading them in static-style). Assign that BP class into a UPROPERTY on your spawner thingy. Spawn an instance of the BP runtime. Done. In the event you're hand-placing objects, then you just hand-place that BP class you set up with the mesh/etc. This pattern also lets you actually tweak your objects outside of code, and build your objects such that their functions can be overridden by Blueprint functions. Though that's a whole other discussion.
|
|
#
?
Oct 1, 2015 23:42
|
|
|
BabelFish posted:I built the UI for the very first published UE4 game, way back before we had UMG. Just from my lowly hobby chair it seems like Slate / UMG have turned out to be pretty successful considering Slate itself was mostly a research project born out of frustration with WxWidgets and similar "cross platform" UI tool kits. The decision to go the Slate route even after having already ported (or mostly ported) from WxWidgets to .NET / C# must have been a tough call to make. But seems to have been the correct one considering they can now claim to have Windows, OS X, and Linux editor support. The fact that Slate also turned out to be useful for Game UI either directly or indirectly through UMG seems to be the cherry on top.
|
|
#
?
Oct 2, 2015 00:05
|
|
|
Shalinor posted:Whatever you want to spawn, make a BP class out of it. Inside that BP is where you do stuff like assign your meshes and other assets (instead of loading them in static-style). Assign that BP class into a UPROPERTY on your spawner thingy. Spawn an instance of the BP runtime. Done. In the event you're hand-placing objects, then you just hand-place that BP class you set up with the mesh/etc.  If you find an official reference I could point them to on this, I'd love to see it.xgalaxy posted:Just from my lowly hobby chair it seems like Slate / UMG have turned out to be pretty successful considering Slate itself was mostly a research project born out of frustration with WxWidgets and similar "cross platform" UI tool kits. The decision to go the Slate route even after having already ported (or mostly ported) from WxWidgets to .NET / C# must have been a tough call to make. But seems to have been the correct one considering they can now claim to have Windows, OS X, and Linux editor support. The fact that Slate also turned out to be useful for Game UI either directly or indirectly through UMG seems to be the cherry on top. I might have gone a little too negative there. As a platform for their editor interface, Slate is excellent (also fast as heck.) It's just wasn't originally designed for in-game UIs, and in early iterations it really showed. It's much better now, but UMG is still the best way to go. BabelFish fucked around with this message at 00:26 on Oct 2, 2015 |
|
#
?
Oct 2, 2015 00:20
|
|
|
I would kill for, like, a 12-week course on Blueprints. Or a book. Or anything other than "MAKE THE LIGHT SWITCH TURN THE SUN ON AND OFF LOL" YouTube videos. I guess I haven't looked fairly lately, maybe something like that has materialized.
|
|
#
?
Oct 2, 2015 00:29
|
|
|
mutata posted:I would kill for, like, a 12-week course on Blueprints. Or a book. Or anything other than "MAKE THE LIGHT SWITCH TURN THE SUN ON AND OFF LOL" YouTube videos. I guess I haven't looked fairly lately, maybe something like that has materialized. I think that's likely the root of my own dislike as well- every time i use blueprint I see my productivity grind to a halt, and end up spending a huge chunk of time trying to accomplish comparatively trivial stuff that I know exactly how to do in code, purely because I simply don't know blueprint that well.
|
|
#
?
Oct 2, 2015 01:18
|
|
|
I went to art school so I'm probably more visually inclined than the average programmer, but I just threw myself in and swam around pretty easy. If you don't know exactly what you're looking for you might have to trawl the menu some, but most things map one-to-one to what they're called in c++. Working with struct members gets annoying tho.
|
|
#
?
Oct 2, 2015 02:08
|
|
|
I can punch things
|
|
#
?
Oct 2, 2015 07:09
|
|
|
Do you guys think I should be at all concerned about save scumming? Since crimes are committed intelligently, it would be perfectly feasible for a player to save, doggedly iterate through an entire investigation over a week of in-game time, figure out whodunnit, then revert to last week's save and immediately detain the perp and push him as hard as you want, content in the knowledge that he'll eventually crack and justify the arrest. I'm thinking it isn't worth addressing- it doesn't break the game as much as it does bypass part of the time management economy, and if someone is enjoying the game so much that they're willing to go through that grind, I think it might be better to say "screw it" and let them.
|
|
#
?
Oct 2, 2015 10:03
|
|
|
Old Man Mozz posted:
Oh yeah, I saw this on Twitter! Best of luck with it. I'm using Haxeflixel and this genre is right up what I'm doing as well. Feel free to hit me up @pizzamakegames, I can help you figure stuff out if you want.
|
|
#
?
Oct 2, 2015 10:24
|
|
|
If it's a single player game, save scumming should be a non-issue. Let people play the game how they want. If there's online leaderboards or multiplayer of any sort, then yeah, it can become a concern.
|
|
#
?
Oct 2, 2015 14:54
|
|
|
The most I'd do to deter savescumming would be to autosave at every opportunity and only have one saveslot. The player can still savescum if they're willing to go to the OS and make a backup of the savefile, but people who might otherwise end up sort of sliding into a cheating mindset will be more likely to play the game "as it was intended". I'd only do this if I felt that savescumming was taking something important away from the game, though. Alpha Protocol did it, because the devs went to a lot of work to make certain that "suboptimal" play resulted in equally interesting gameplay as doing everything perfectly did. There were times where I found myself thinking "Man, I wish I'd gone back and done that conversation better", but since the game saves after every conversation that wasn't easily doable...and it spared me from frankly tedious min-maxing.
|
|
#
?
Oct 2, 2015 15:22
|
|
Not anywhere close to as impressive as some of the stuff posted recently (procedural crime generation!  ) but I've got a nice little UI going for my rogue-style adventure/RPG: Using a variety of tricks, I'm able to work with 16 colours on-screen (out of 4096) and make it look just like a PC-9801 game. I just got done programming a dirt-simple UI that gets the job done and I'm having a lot of fun working on this game. One big way I keep the colours organized to prevent breaking the limit is having a fixed 8-colour Character/UI palette, and another 8-colour palette for each environment. I can also have enemies or characters that only appear in X environment share a palette with said environment, to give the illusion of the gamut being wider than it actually is.
|
|
|
#
?
Oct 2, 2015 15:31
|
|
|
Don't sell yourself short, that looks rad as hell. What's combat going to be like?
|
|
#
?
Oct 2, 2015 15:48
|
|
|
BabelFish posted:We had some bad experiences with blueprints way back, and now none of the other engineers want to touch the things  Blueprints are part of UE4 workflow. You can't use UE4 without touching them, unless you want to incur some pretty serious hits to your initial load time and memory profile. It isn't like you have to do extensive BP scripting, you just need to use a BP class to assign in your assets. You can usually do it without adding a single Blueprint node / touching the visual scripter, even, and even if you DO have to use the visual scripting, we're talking a handful of nodes per object at most. Noyemi K posted:Not anywhere close to as impressive as some of the stuff posted recently (procedural crime generation! Shalinor fucked around with this message at 15:54 on Oct 2, 2015 |
|
#
?
Oct 2, 2015 15:51
|
|
DeathBySpoon posted:Don't sell yourself short, that looks rad as hell. What's combat going to be like? A mix of Rogue and Xanadu (PC-88/PC-98), where movement and attacks are directly controlled through one character and turn based, and each action outside of pausing advances the game's turn counter by one tick. Like Xanadu, enemies drop chests which can be destroyed on accident or serve as obstacles to other monsters. So, killing small minion creatures might be useful if you find yourself facing a tough opponent that you're sure you can't handle up close, and you'd like to run away and avoid taking hits. I might change that chests bit but I thought it added some strategic depth to Xanadu's simple combat and I really enjoyed it!
|
|
|
#
?
Oct 2, 2015 15:58
|
|
|
Omi no Kami posted:Do you guys think I should be at all concerned about save scumming? Since crimes are committed intelligently, it would be perfectly feasible for a player to save, doggedly iterate through an entire investigation over a week of in-game time, figure out whodunnit, then revert to last week's save and immediately detain the perp and push him as hard as you want, content in the knowledge that he'll eventually crack and justify the arrest. I'm thinking it isn't worth addressing- it doesn't break the game as much as it does bypass part of the time management economy, and if someone is enjoying the game so much that they're willing to go through that grind, I think it might be better to say "screw it" and let them. Sounds to me like the problem lies in interrogation being guaranteed to work if you got the right guy. Throw some false positives and negatives in there, or introduce some cost to it that doesn't go away if it yields a confession. Or both.
|
|
#
?
Oct 2, 2015 16:12
|
|
|
Moddington posted:Sounds to me like the problem lies in interrogation being guaranteed to work if you got the right guy. Throw some false positives and negatives in there, or introduce some cost to it that doesn't go away if it yields a confession. Or both. You could also be required to release the perp after a few days if you don't make any progress, which would encourage the player to find some additional evidence if they couldn't quite close the case last time.
|
|
#
?
Oct 2, 2015 16:17
|
|
|
Maybe rather than just getting a confession, you need to have enough evidence to actually convict them. So even if someone brute forces the right suspect by trial and error savescumming, they still need to get witnesses or physical evidence to actually solve the case. Alternatively, maybe just have it so that the perp will never crack if you don't have any evidence at all, even if they actually did it. After all, why would they confess to the crime when you have no hope of proving they did it without their confession?
|
|
#
?
Oct 2, 2015 16:18
|
|
|
Noyemi K posted:A mix of Rogue and Xanadu (PC-88/PC-98), where movement and attacks are directly controlled through one character and turn based, and each action outside of pausing advances the game's turn counter by one tick. Like Xanadu, enemies drop chests which can be destroyed on accident or serve as obstacles to other monsters. So, killing small minion creatures might be useful if you find yourself facing a tough opponent that you're sure you can't handle up close, and you'd like to run away and avoid taking hits. Which version of Xanadu did you play? The only one I have fond memories of is the one where they took one look at it at the original with all it's stolen artwork from the Ultima 3 manual (yes this happened, lol) and were like "uh, no thanks" and redid the entire game from scratch when ported as Faxanadu.
|
|
#
?
Oct 2, 2015 16:24
|
|
Mercury_Storm posted:Which version of Xanadu did you play? The only one I have fond memories of is the one where they took one look at it at the original with all it's stolen artwork from the Ultima 3 manual (yes this happened, lol) and were like "uh, no thanks" and redid the entire game from scratch when ported as Faxanadu. Oh, I remember that. Well, reading about it, because I wasn't born when any of them were released. I played the PC-88 original, and Revival Xanadu on PC-98, but not Faxanadu.
|
|
|
#
?
Oct 2, 2015 16:28
|
|
|
I'm thinking the reward you get for clearing a case is commensurate with its capacity to be prosecuted, which gives a major economic motivation for collaring the right guy (as a mistake/purposely fabricated arrest will be worth less) while refraining from punishing you for being a corrupt dick (and also, a corrupt detective). The more we discuss it, the more I'm convinced that save scumming isn't an issue at all. My original concern was that it could be used to completely eliminate the time management stress from having a big caseload, kind of like how Persona 4 meant for you to run through the dungeon-of-the-month several times before clearing it, which made the pacing feel slightly hollow when you cleared the entire dungeon on the first night and bought yourself an unplanned-for week of free time. However, I don't actually think you lose much of the experience from doing it this way- even if you're solving the case, then reloading and sprinting right for the perp and the key evidence, you're still getting the complete experience of the case. In some ways it even works in a mildly meta sense- the cheaters are the sherlock holmes-type detectives, who take a single look at a crime scene and instantly go "I know what happened. Find the red-haired bobby who was on patrol at 2:35 last night on Knickerbocker Lane and check his left-front pocket. The whistle discovered therein was most assuredly the murder weapon."
|
|
#
?
Oct 2, 2015 16:34
|
|
|
Omi no Kami posted:Do you guys think I should be at all concerned about save scumming? Since crimes are committed intelligently, it would be perfectly feasible for a player to save, doggedly iterate through an entire investigation over a week of in-game time, figure out whodunnit, then revert to last week's save and immediately detain the perp and push him as hard as you want, content in the knowledge that he'll eventually crack and justify the arrest. I'm thinking it isn't worth addressing- it doesn't break the game as much as it does bypass part of the time management economy, and if someone is enjoying the game so much that they're willing to go through that grind, I think it might be better to say "screw it" and let them. I would consider if there's benefit in adding an ironman mode. Even if the advantage in save scumming is small and harms their enjoyment, some players can't help it.
|
|
#
?
Oct 2, 2015 16:38
|
|
|
That's a good idea, and it would be trivial to add. I keep going back and forth on whether I want the standard saving mechanic to be completely free, or constrained to 1 manual save + autosave. I like the commitment of the latter, but as a player I have had so many incredibly frustrating experiences with bad save systems (MGSV, I am looking squarely in your direction) that I think unless there's an excellent reason not to do it, "Save anything, anywhere" should be the default, and anything like 1 save/ironman should be player-defined optionals.
|
|
#
?
Oct 2, 2015 16:49
|
|
|
I want to make an RPG, with pixel art graphics (probably 32 bit) with the potential for it to be a fairly long term project, with decent potential to expand my ideas on the same software. I have very little programming experience (used to know Visual Basic way back, and took a class on Java years ago) and would prefer something I could jump in pretty quickly. From looking around, it looks like Construct, Stencyl and GameMaker are the most popular options for relatively simple 2D game making, but in theory they seem quite similar and I'm not sure how to choose which one to commit to. I've poked around with them a bit, but it's kind of hard to really get a feel for which one is best for this sort of thing. I'd appreciate hearing about any experiences y'all have had with these, which you'd recommend, pros/cons etc. a lot. I'm sure I'll have quite a few more questions as I get into this, but for now I'm kind of stuck at this very early stage.
|
|
#
?
Oct 2, 2015 17:19
|
|
RIP ZEEZ
|
So Microsoft bought Havok. Chiefly known for their suite of physics libraries. Other things they do that are less well known are AI libraries, LUA scripting system, and a full fledged game engine called Vision Engine. http://blogs.microsoft.com/blog/2015/10/02/havok-to-join-microsoft/ I wouldn't be surprised in the slightest if these things all become free for indies.
|
|
#
?
Oct 2, 2015 17:26
|
|
|
foutre posted:I want to make an RPG, with pixel art graphics (probably 32 bit) with the potential for it to be a fairly long term project, with decent potential to expand my ideas on the same software. My recommendation is pick any of the above and do a tiny project. The chances of you picking up an engine for a long term project as your first game project and actually finishing it are astronomically small.
|
|
#
?
Oct 2, 2015 18:12
|
|
|
People who use Blueprint in UE4, is it viable to make a shippable game? I'm not talking about someone's GRAND OPUS DREAM OPEN WORLD GAME but like, simpler fair like a Tetris-like or a match 3 or something. In your opinion could someone make a shippable product like that in Blueprint? Edit: I guess, a similar but more specific question to the above, haha.
|
|
#
?
Oct 2, 2015 18:13
|
|
|
mutata posted:People who use Blueprint in UE4, is it viable to make a shippable game? I'm not talking about someone's GRAND OPUS DREAM OPEN WORLD GAME but like, simpler fair like a Tetris-like or a match 3 or something. In your opinion could someone make a shippable product like that in Blueprint? Yeah, definitely.
|
|
#
?
Oct 2, 2015 18:37
|
|
|
mutata posted:People who use Blueprint in UE4, is it viable to make a shippable game? I'm not talking about someone's GRAND OPUS DREAM OPEN WORLD GAME but like, simpler fair like a Tetris-like or a match 3 or something. In your opinion could someone make a shippable product like that in Blueprint? It's definitely possible, but like every other tool, blueprint is perfect for some use cases and not well-suited to others. Blueprint is great for rapid prototyping, my advice would be to keep data and high-octane computations in C++, then expose your interfaces to blueprint and use it to actually make the moving parts you've set up work together.
|
|
#
?
Oct 2, 2015 18:41
|
|
|
xgalaxy posted:So Microsoft bought Havok. Chiefly known for their suite of physics libraries. Other things they do that are less well known are AI libraries, LUA scripting system, and a full fledged game engine called Vision Engine.  EDIT VVV: mutata posted:The main issue there with me is I am an artist and despite good intentions, I have yet to begin my programming education journey but I'm also a whiny millenial baby and want to make a game RIGHT NOW (next year). More importantly, you're making a game, learning UE4, and learning scripting / code flow in the mean time. Which is awesome. All of that is necessary as a foundation before you dive into C++ anyways (and would be useful even if you went "gently caress this" and hopped over to Unity for C# scripting). Shalinor fucked around with this message at 18:49 on Oct 2, 2015 |
|
#
?
Oct 2, 2015 18:46
|
|
|
Omi no Kami posted:It's definitely possible, but like every other tool, blueprint is perfect for some use cases and not well-suited to others. Blueprint is great for rapid prototyping, my advice would be to keep data and high-octane computations in C++, then expose your interfaces to blueprint and use it to actually make the moving parts you've set up work together. The main issue there with me is I am an artist and despite good intentions, I have yet to begin my programming education journey but I'm also a whiny millenial baby and want to make a game RIGHT NOW (next year).
|
|
#
?
Oct 2, 2015 18:46
|
|
|
|
| # ? May 29, 2024 01:52 |
|
|
There's always RPG maker, and I just saw Steam advertising a new version of it. I don't have much experience with it myself, though.
|
|
#
?
Oct 2, 2015 18:56
|
|