Want to make a PSO clone, need help with this and that.


PSO is pretty simple as far as RPGs are concerned. It's pretty much a simpler, 3D version of Diablo. There really is not much gameplay to it beyond character positioning, good timing and item collection madness. A homebrew game like POWDER is much more complex I imagine.


There IS one potenitally difficult thing though, I guess: I don't know wether SEGA did this by hand or had some randomizer, but the Dungeon layouts would change on a weekly schedule.


Edit: Having said that, hell yeah I want some of this! Can't offer any help though as I'm still stumbling over pointers :p


Edit2: Well putting more thought into this, a timing critical MP game could be difficult in terms of networking, again just an ineducated guess :D
 
Last edited by a moderator:
Well, I don't even need a complex RPG Story or anything like that. I'm happy with a "Simple Diablo style" gameplay as long as I can smash Monsters, grow the own characters and find awesome Items. :D Thats why I enjoy Diablo 3 at the moment. But as you may know, they needed 12 Years to make such a "simple" game like Diablo 3 and I bet they've spent the minority of the time on making graphics and engine.


The big challenge is the game balancing, the actual gameplay and "how it fits together". This need the most of the time in such games. The entire "sub systems" of gameplay, formulas, calculations of skills, actions, Monsters, behavior, items, also map design... You can spent Years on that I guess.


Compared to that, the engine, Graphics, sound etc. is the smaller part of such a game. ;)
 
The dungeons had a number of different complete designs were individual sections were in different spots, and wich complete design you got to play was chosen randomly by first person to enter server game.


The multiplayer timing dosnt really happen, the server will decide how enemys move and when they attack, and when you attack your client send the damage and a hint for enemy animation/ai. And first to pick up items is measured by clients own time. If multiple same time reports happen the server can randomly select a candidate.


I finnished the editor to use for this project, but I havent thaught about how to do the dungeon random thingy... In my editor you cant select several polygons at once... and theres no copy function.... Lets just make static levels, and static caracter models as well, its easier and quicker :) Variations of a caracter can still be with different color schemes for textures for ex. We could place enemys and stuff around the level with random locations instead, that would still provide some variations that would be nice and easy to do.
 
How about this: You could make a level that has a multitude of different paths and have collapsed doorways/bridges generated randomly so players have to take a different path. Dunno how complicated it would be making the rules for something like that. Edit: And the levels would have to be pretty big to make it worthwhile.
 
Last edited by a moderator:
Yea that should be easy to do, I was thinking 1000 polys for an entire level, since levels are usualy very flat and very planar walls mostly and not to much details overall, but if levels are gonna be big mayby a bit more would be needed. The renderer does 50 000 polys at 20+fps, I dont remember if 100 000 sinks below 20fps, but theres gonna be lots of other processing to, Ill make a test with new engine when its done and then I should be able to make a rough estimate of how many polys can be used. The reason I went with 1000 is because I thaught it would be enough for a level, and if all data packages is same size its convient, but then ofcourse it dosnt matter if level data package has different poly count number, but for ease lets say all stuff is 1000 polys, thats 4000 for caracters, plus 4000 for MAGs, plus 20 000 for up to 20 enemys at same time plus then 10 000 for a level, its still within the limits of my old renderer.
 
OMG! I just tryed to model a caracter, its really the first time I try to do anything in my editor and I realised I dont have a function for deleting a polygon... Well anyway, its insanely difficult and frustrating to try to do caracters, enviroments shouldnt even be half as bad to do, but I think I will have to deal with the caracters cos that nor fun nor easy... Lets just make 6 caracters at all, one male and one femal of all classes, hunter, ranger and tech. We will see how it goes, mayby I get used to it and can quicker and easier create caracters in the future, but lets start with this anyway.
 
The different classes sounds good, just different colour schemes would be good enough to set people apart for me, I never though about PSO being fairly simple (I'm fairly uneducated in game design and tech specs) but it doesn't sound too bad, as a PND it would be easier to update as you go along too, so you could create more elaborate animations further down the line.


The cities wouldn't need that many polygons like you said, it's fairly simple in design, the space stations maybe more, I'm drawing an internal shot at the moment and i've got it with a central circular hub, something like that would take a few more polys, as for the worlds I think we should have mainly static dungeons, so people coming in later get the same experience, just adding more and more in later.


MAGs could be made simpler, maybe even 2D sprites that change depending on perspective, like the Doom sprites or something, and then develop them further down. I'm no expert though maybe it would be easier to do them in 3D, I'm still very excited and optimistic for this.
 
I think random stuff could really save time. Random maps/dungeons are fun and always a new challenge, they also did this in Diablo 3 (but not everywhere) And for performance reasons, the brandnew Diablo 3 also uses some nice tricks, for example, the trees and other level decorations are only 2D, or better "2.5D" sprites but it still looks good. :)


Random quests could also save much time because you don't have to actualy make complex quests per hand. And I personaly don't need a story or anything like this. Random Items are a must but thats nothing new.


Of yourse, you still have to develop the formulas for all this random stuff. But once finished these groundwork, the game could grow pretty big in a compared short time.
 
Im getting really excited about this now. Random dungeons is hard to do with my crap tools, I really like the idea of blockades forming different paths instead. The MAGs will be easy to do, its the human and round shapes thats difficult to achieve, the more angular and more planar surfaces the easier to do. I have figured out how to release the tools now, the editor will be a pre-compiled program that runs in linux and need SDL and GLEW runtime libs I think theyre called. Textures need to be bmp and RGBA format. You can make that format in GIMP thats free for linux. Ill make 2 editor programs, one for 1000 poly packages, and one for 10000 poly levels. It will save and load static names, so after saving a file for a completed level for that file needs to be renamed otherwise something else will overwrite it later, and loading needs to be handled same way. And the textures, Ill make a list in the editor that reads names and specific number of textures, for ex t1, t2, t3 etc to for ex t100 for max 100 textures, and all textures created will have to be named that, there needs to be 100 textures in place with correct name for editor to read, so make a small 8x8 bmp and copy it and name it to all numbers, then replace the number textures one after another as they get created. I need to fix a few more things in editor then Ill wait until I have the game renderer and physics up and running on Pandora and release all that in one package and people who want to can create content and immidiatly test it on Pandora, its uncomfortable, but atleast its doable. :)
 
If you can create random creature placement, create a bunch of dungeons that have segments, like MGS Peace Walker or Monster Hunter, where the map is split into multiple areas, and have random combinations of areas connected.


No idea if that's possible or practical, but I thought I'd throw the idea out there.
 
Weren't there five or even six classes in the original PSO? My Shinji was a hybrid focussed on melee, but could also cast and wield guns. I remember using those submachine guns a lot because of the insane dmg output at close range.
 
I havent played either MGS Peace Walker or Monster Hunter so I dont know what you men, Ill take a look at it on youtube later.


In the original PSO there were only 3 classes, you could use different weapons with them but for ex only rangers could use rifles and only hunters could use the more powerful sword weapons and so on. Also each class had 3 different species, humans, robots and newmans that looked a little like elfs or something.


I think instead of classes mayby theres shouldnt be any, and instead you spend earned experience points freely on what you want to make your caracter more specialized and more unique and to your liking. And then mayby buy weapons licenses to use a particular weapon type and level.


I realised how I could make multiple floors above and underneath eachother work with same collision detection system, if a position reports a hit with multipe polygons, the polygon height is choosen wich creates the smallest difference in height from previous caracter position, thats not a solid system and could theoretically produce wrong results in some occasions, but I think it will work just fine here for various reasons, and if anyone builds levels just playtest entrances to different floors carefully, if you end up some place you shouldnt it was because of height difference calculations choosing the smallest difference value, and you need to move different paths to the multi floor sections further away from eachother, youll figure it out, I dont think it will ever happen anyway due to slow move speed mainly.
 
^ah right, I had a vague image in mind of a character selection screen that had more than 3 options. Those were the combinations of classes and races, of course! I had a hunter then. Yay for greatswords bigger than your character :D
 
Incredible... I cant get anything to run on my Pandora, Im beggining to think theres something wrong with the librarys I have or something... Ill take another look at crosscompiling again, if I cant get that to work Im throwing in the towel on Pandora development altogether... Its a billion times easier to develope for IOS, isnt that funny? Why isnt there a coding program with a finished context that works for Pandora?
 
My new code wich is basically same as old render code I had there, system wise, ir compiles and when I try to run it it says "killed" I think, didnt saw that error message in the egl code either, have no idea what that is about... My old render code cant display textures, dont know why, no errors and code runs fine, but polygons are blank.....


I managed to install a toolchain now, but have no idea how to use it.... so tired. Im gonna carefully do some gameplay coding still, cos Im learning and doing things I can directly transport to my first commercial game, and then mayby someone else could compile it in the end? I can ask on dev forums if someone can compile it and see if all works then and there, if it dosnt then I probably stop worrying about Pandora at all and focus on other platforms.
 
I dont know how to use the crosscompiler toolchain, and there dosnt seem to be any interest from anyone who would be able to compile it. It all depends on what happens in the future regarding ease of compiling, I would be happy to start working on it at any time, but until I can compile it or someone who can comes along I wont be working on this game specifically but focus on other projects wich can easily be grabbed the system from to have the gameplay up and running and functioning quickly.


Did you do any specific work? Its best to consider this project dead unless I say otherwise in this thread.


I just tested the collision detection system and it worked nice, I started doing the camera, it needs to have a way of not sliding through walls otherwise thats done also, so pretty much the hole game mechanic was working and ready for being put to use, I realised what someone said about my textures and changed so same textures dont get recreated constantly, it saved quite a few fps, thanks man I dont remember who said it.
 
Wait, I just remembered something, I could compile GLES1 code, the textures were garbage but I think it was because of wrong format, I just had the same effect in other code this evening were I forgot to change a parameter, cos usually I just got wrong colors but here it was same effects, wich reminds me of similiar looking effects with wrong sizes and so on. Ill try with correct format texture tomorrow, but then a new problem arises, without shaders and without alpha channel textures hmm, well you could still make a OSD with multiple layers, but there wont be any trasnparent bits, so it will be square windows OSD and quite ugly... Well screw it, hole game is gonna be ugly as heck :/ But if the textures work out tomorrow the project is BACK IN BUISSNESS! Ill just have to hope for being able to compile GLES 2 code for the Pandora 2 in the future so I can make a better sequel then with kick ass graphics! :)


Ah yea, it seems both my nubs are completly dead, I never looked into what the nubs are actually, joysticks? the rest are just SDL keyboard, but the game needs analog controls, I guess I have to make do with ddigital controls, and I look into analog programming later and make an option so people can choose what they want.
 
Use the nubconfigurator or a startupscript like the one of mupen to set the joysticks to joystick or mousemode.


The yactfeau toolchain is quite simple and I can compile some stuff for you, if you like. Just send me the code.


Edit: have a look here for eglport/sdl use and textureformats:


http://pandorawiki.org/Porting_to_GLES_from_GL
 
Last edited by a moderator:
Back
Top