next »

[Faust06]

Hi guys. This is a pretty loaded topic - I wasn't sure where to start it, but since I have questions from a design standpoint I guess I'll throw them here.

First off, I'm beginning work on a roguelike with some graphic flair, and I'm trying to decide how to approach it from a programming standpoint (this is my first full-fledged game and development will continue indefinitely). Though it's somewhat minimalist for better visibility and contrast, it wont be ASCII - sort of a mix of Quarry's bytesoup and brogue's animation from a visual standpoint. Normally I'd go with Java (which I have experience with) or Python and draw from the plethora of resources pertaining to roguelike design. My concern is that going the traditional route and simply applying what's normally an optional tileset permanently will be too limiting, and that I should approach this as a purely graphical game. If that's the case, would I benefit from aggregating (taking a short-cut) with Construct 2 or the like, or will that just get in the way? Should I still go with Java? Etc, etc.

The hardest part is getting started, but once that's out of the way, here's what I have in mind: A roguelike strongly influenced by Shin Megami Tensei. That is, it will feature recruitment, and passive/active squad-based commands which will be fleshed out later. I'm excited about it. It's going to happen.

I had a concern about movement. With most roguelikes featuring a huge map, upon movement the map spatially just moves around the character - this is plainly dizzying to me, especially with fast-travel (e.g. Crawl). One of the reasons I like Brogue is that the map doesn't move, only your character does, and this is much easier on the eyes. The issue, as I see it, is that it would be difficult to implement Brogue's approach if a) the map is huge, or b) there is a greater zoom/magnification on the tileset so as to not show the entire map at once, which with graphical roguelikes is more than likely. I'm thinking of taking a half-assed "Zelda" approach and just "shift" the screen in any direction, but I'm not certain. Thoughts on this are welcome.

[keo]

Love me some SMT.  Another route you could take for camera control, you can have the player control the camera, and a shortcut to center on the PC.

[C.A. Sinner]

would it be less dizzying if the game had full walking animations and non-immediate scrolling a la shiren the wanderer? that'd mean making more assets ofc. maybe implement manual zooming as well?

re: development tools: i'd say just use java if that's what you're experienced with. i can't speak for construct 2 but i'm making a roguelike with game maker atm and i wouldn't recommend it unless youre already familiar with the program because youd probably have to circumvent some engine features etc.

[Faust06]

Popular choice for libtcod users in RL dev seems to be Python or C++. I think Brogue uses that. I may give it a shot and see if I can polish up the visuals just the way I like, but I'm still leaning toward graphic.

From what I remember about Construct, it allows circumvention and the like using Python. I haven't tried Construct 2 yet. It's tempting to exploit something like Construct to cut down coding time, especially if I'll be creating pixel art, but it may be more trouble than it's worth. For now I suppose I'll tinker with both approaches. If someone has experience with creating graphic roguelikes, I'd certainly like to hear it from them.

Love me some SMT.  Another route you could take for camera control, you can have the player control the camera, and a shortcut to center on the PC.

I'm playing Nocturne now. As with most jrpgs, the travel and grind is tedious but I enjoy a lot of the elements that make the series unique. I also want to incorporate the exploitation of weaknesses, and status effects. I enjoy status effects much more than elemental, generally.

Interesting take on camera control. One thing that really drives the importance for a better camera is the co-ordination of allies.

[flowerthief]

Just this week I completed development of a freeware graphical roguelike and I'll never do it again. Not because it was difficult to make or anything (I used RPG Maker XP and it worked out well) but because the lack of interest shown towards it has been deafening. Now I understand why roguelikes are so uncommon outside of the indie scene. The audience for them actually is as small as it appears, it would seem. I wish I could take back all the time I spent on it and put it into something else instead that people would appreciate.

I wouldn't recommend that anyone make a roguelike, but you seem to be passionate about your idea so I probably shouldn't be trying to discourage you. Hopefully you'll have better success than I did.  SMT is a pretty fun series and could definitely be improved if there was less grinding, I agree.

[Paul Jeffries]

I would say just go with Java, unless you have any particular desire to learn something new.  I recently completed a graphical roguelike in Java (devlog here, if you're interested) and found it almost absurdly easy to do everything I needed to do just using Java's built-in graphics functionality.  Not being tied to a specific third-party library makes it easy to extend it how you like as well.

I am unfortunately required to concur with flowerthief regarding the lack of interest, however, since that was my experience as well.  There seem to be substantially more people interested in developing roguelikes than there are people interested in playing them, sadly.

[Faust06]

I appreciate the feedback. I've thought about it for some time, and I think for the sake of 'roguelike-ness' and because I tend to prefer ASCII over graphical anyway I'll go the traditional route and tack on an optional tileset later in development. It's funny, the main reason I wanted to go graphical was for better accessibility, and impressive visuals (which, I think, can still be accomplished with libtcod) but you're correct that they don't draw an audience anyway. There is little reconciliation between hardcore roguelike players and casual gamers with fleeting interest; in fact, I think the graphical middle-ground probably garners the least interest in that genre (DOD and HSL will fall into obscurity in time while Crawl will always be popular)

I also wanted graphical since I originally had a tactical rpg in mind. But the flow, depth, challenge and atmosphere of a roguelike is more appealing. Not to mention the fact that creating art really frustrates me.

[Sharkoss]

Whatever you do, if you have animation on the player sprite, don't make me wait for the movement loop to end before I can move again.

Just this week I completed development of a freeware graphical roguelike and I'll never do it again. Not because it was difficult to make or anything (I used RPG Maker XP and it worked out well) but because the lack of interest shown towards it has been deafening. Now I understand why roguelikes are so uncommon outside of the indie scene. The audience for them actually is as small as it appears, it would seem. I wish I could take back all the time I spent on it and put it into something else instead that people would appreciate.

I wouldn't recommend that anyone make a roguelike, but you seem to be passionate about your idea so I probably shouldn't be trying to discourage you. Hopefully you'll have better success than I did.  SMT is a pretty fun series and could definitely be improved if there was less grinding, I agree.

There are too many recent popular roguelikes to take your unfortunate experience as anything more than that.  I'm sure there are things that can be learned from it, tho.

[ASnogarD]

The lack of feedback may be more due to a lack of a fanbase for your games, hence why it is always a better idea to make a few other games before climbing into your masterpiece... that way you will at least have some players that know you exist to try your game and pass the word.

You could literally make a Diablo that puts Blizzard to shame, and beats Torchlight with a rotten Zombie leg and get no response if you dont first build up some form of a player base.

I have a graphic roguelike in mind to make but I intend to do at least 2 more games and a experimental procedural content generator driven game before I do that roguelike, hopefully by then I'll have more than myself looking at my game 

You could always try doing the roguelike on the side while training your game design skills on other games and broadening your playerbase, but that does tend to dilute the workflow a bit and is a exercise in self control trying to not spend too much time on the roguelike.


In short : First get some games out there so players will know you makes games, then build the big game you want.

PS: If you are already a established maker of games then ignore my waffling 

[Paul Jeffries]

@Sharkoss: There are certainly some very popular roguelikes around, but take the sheer number of roguelikes being developed into account and I think you'll find that those are the edge cases.  Lack of players/feedback is a pretty common complaint amongst the roguelike community - while there are a lot of very dedicated players out there they tend to stick to one specific major game (Nethack, Angband, Crawl, TOME etc.) and are not so interested in trying out other newer games (to be fair, this is probably because there is a constant flood of them and most of them are not very good).  The most successful recent games (Dredmor, Cardinal Quest) have been those which have succeded in drawing in a more 'casual' audience from outside the typical playerbase - which is why I would not particularly advise the ASCII route.

But I hope that my wingeing won't put you off - I'm sure it's possible to build up a player base over time if you're willing to put the effort in.  If you're after quality feedback and a little more exposure you might want to take a look at the new Roguelike Incubator initiative that's just starting up, which aims to help out with some of these issues.

[Sharkoss]

Well, sure.  But if you feel like you have something distinct that will stand out, go for it.  I'm one of those guys who pretty much only plays Nethack, but if the right game were to come along I'd make the switch.