I've been playing the excellent Cardinal Quest by Ido Yehieli and I can't recommend it strongly enough. You can download the demo from links at his site, read the IndieGames.com article about it and there's a great interview with Ido at True PC Gaming which outlines some of the challenges in designing an indie game.
While Cardinal Quest doesn't necessarily have the depth of some roguelikes - and isn't intended to, it has one outstanding feature, which I also praised Terraria for: amazing user interface design. Here's the first screen you see once you choose which class you'll play (click to see full size):
The user interface entirely describes what you can do in the game. There are keyboard shortcuts for every function you need to do but you don't need to look up help menus: everything is discoverable by mousing over it and a small pop-up reminds you of the key to use, as well as describing the item in the slot.
Not only that, but whenever you pick up an item, it automatically is placed in the correct slot if it is better than what you're already using, replacing any existing item which is converted to gold.
And perhaps best of all, the spell system has separate time outs on each spell, which are displayed by a clock ticking down effect which slowly highlights a slice of the spell icon as the spell recharges.
The user interface is so good, it has seriously forced me to reconsider my priorities in Unangband. I typically find other (non-Angband variant) roguelikes impenetrable because of the subtle but important differences in keyboard shortcuts which means I can't easily shift between games. But I had no problems at all with Cardinal Quest, I could start playing straight away and enjoy the progression of my little avatar (Down to level 8 first time).
With this and the success of Dungeons of Dredmor, I wonder how much the lack of attention to user interface typical of hobbyist and indie programmers has been holding roguelikes back. It's not just that of course: the verb-object model of using items - which has important properties for emergence - and the large number of items in a typical roguelike are also important, but most of those could be overcome by a smart redesign. I'm thinking about going through this process myself.
Monday, 22 August 2011
Is user interface design holding roguelikes back?
Subscribe to:
Post Comments (Atom)
Posts
10 comments:
I've been saying (and wishing) that roguelikes would put some more effort into UI for years. Unfortunately this usually devolves into an ascii-vs-tiles argument, with most people completely failing to realise that UI is about much more interesting things than what exact symbol is used to represent each monster.
I got so fed up with this that my roguelike, Albion ( http://albion.triangularpixels.com/Game/Albion.jsp ) is largely an experiment in seeing how to make an accessible UI without sacrificing the complexity (particularly the verb-noun combinations).
I think it's reasonably successful in that goal - I only let the user perform 'valid' actions (like not drinking swords) and put the most likely action on left-click (like hitting someone with your equipped weapon, or healing if they're friendly), but can put the other valid actions on a right-click menu (like throwing potions, or hitting your allies).
Unfortunately I got bogged down in the party mechanics, but the principles of a simpler UI are pretty solid I think.
I haven't played cardinal, will probably try it.
To the main point:
Angband's interface is rather meh, but better than DF for example. I don't mean the graphics, those are very good in both games, if you like ascii.
Angband's most evident problem here is the aAuzrmp problem. All of those keys do almost the same thing.
If I had the skill´s to experiment myself I would switch from verb-noun to noun-verb. One button for default action of selected object, one button for a list of possible actions on an object.
That means two required buttons to memorize, plus movement keys, will be all you need to play the game.
Tobias: You shouldn't just try it, you should buy it.
Noun-verb is definitely the way to go but there's a lot more than just this change to getting the UI right.
I think mouse support can instantly transform the UI of a roguelike. Tooltips can explain all the necessary information, and right-click menus can give all the available options. There doesn't need to be any sacrifice of complexity (though admittedly it works better with less complex systems). Of course the interface itself must then be mouse-friendly, with sub-windows and easily accessible buttons.
UI is drastically under-rated for almost all games, particularly any sort of turn-based or strategic games. It was a huge problem for Hinterland and every strategy game I ever worked on went through 2-4 complete UI revisions before and after playtesting. If you really want to see how your UI performs get a girlfriend/wife and her non-gamer friends to try and play it.
Dwarf Fortress would definitely have a much greater following if it had a more useable AI. All of my hardcore gamer friends want to play it, but can't get past the UI.
I'm surprised you praise Terraria. For all its polish I think the UI is really opaque and poorly designed. No one I know has had an easy time getting in to the game or understanding the UI.
I recommend not picking one of noun-verb or verb-noun, rather pick both. Introduce new players by teaching noun-verb since there is less to learn, and you can have enter/space which does the most obvious thing. Then as they become more advanced players and are prepared to learn more keys, the verb-noun is still there as a faster shortcut.
For all the difficulty in the UIs of roguelikes, I find that compared to any other game I can do so much more so much faster because of the many many shortcuts. Giving these up to aid new players when you can simplify things for them without the sacrifice is just unnessecary.
Yes I've always thought ui to be the big let down in roguelikes.
Ascii graphics are fine and in fact preferred by myself but navigating all those menus via the keyboard, ugh... drives me nuts and certainly keeps me from trying a lot of new roguelikes and *bands.
I've always thought a roguelike designed to use a windowing toolkit, as opposed to the tacked on ones, could be a big success.
Isn't that obvious? Roguelikes (especially Dwarf Fortress) are all but unplayable due to how unbelievably bad their UIs are, and Minecraft sells three million copies to the people who would like to play DF, but have no time to learn the controls. On that topic, if DF had a UI that was a joy to use (and not something that makes you scream in rage whenever you want to change a few jobs around), I bet it would sell a million copies easily.
Proper abstractions help too. As Monkey Island 2 has demonstrated, you don't actually need 12 verbs, you can get by with 9 just as well. The difference between zapping wands and quaffing potions is pointless. You're not going to quaff wands and zap potions, and when the player clicks on a potion, you can guess three times what he wants to do.
Mouse + Context menu works a lot better than learning difficult acronyms by heart.
"Is user interface design holding roguelikes back?"
No.
The *absence* of user interface design is. Has always been.
IMHO the driving principle behind roguelikes' interfaces has been historically just one: "it must be easy to program".
I'm happy we're finally (slowly, lately) seeing steps on the right direction of the UI - which is "it must be easy to use".
By all means, feel free to fiddle with Unangbard's UI.
Here I kinda don't agree. Though I understand that most people are "mouse junkies" I kinda hate that thing and would use keyboard instead of mouse almost all the time(except not for obious things like fps, rts, 3d modeling gimp/ps). I totally agree that DF interface sucks quote a big time. But for example I find ADOM's interfaces as one of the best, well of course the fact that its one of my most played games has some influence here. Still I think pointing and clicking is seriously overrated. I understand that it can be easier for newbies. But my observation is that when user is experienced(well I'm talking more about apps here) he will use keyboard more, and it will be faster that way. And since time and money is limited one often ends up just with one paradigm fully developed. And thus I take flexibility and performance of experienced user over simplicity for newbies.
Post a Comment