Adventure Game Studio

Note: Based on some of the feedback here, I made a small addendum to the final version of part one on my blog.  For the rest of the articles in this series, I'm going to post them here first and then modify them based on the discussion before posting the final versions to my blog.

Previous parts:

• Part One: Cursor Confusion (http://xiigames.com/2008/07/19/why-your-game-is-broken-part-one-cursor-confusion/)

Part Two: The Sierra Operating System

In this edition of "Why Your Game is Broken," I'm going to pick on something near and dear to many adventure fans' hearts: the Sierra and Lucas Arts interfaces.  These two interfaces, Sierra much more than Lucas Arts, are used in an overwhelming majority of the games that come out of this community. This is largely because the Sierra interface is used in the default template provided with AGS.

Unfortunately, partially due to the inherent flaws of these interfaces, and partially due to the developers' sloppy use of them, nearly every one of these games are broken.

The Sierra Operating System

Imagine you're using a computer.  Shouldn't be too hard.  Now imagine it's a completely different operating system â€" a new one you've never used before.  There are four items on the desktop: a folder, a video file, an mp3, and a word document.  There's also a little image of a man in the bottom corner.


Let's open the folder.  Move the cursor (with its well defined hotspot) to point at the folder and double click.  Nothing happened… Oh wait, the little guy is walking up there.  That's strange.  Try the movie.  Same thing.  The little guy just walked over there.

Obviously something about this new OS works different from the ones you're used to.  You'll have to learn how to use it.  Oh look, pointing at the top of the screen reveals some buttons, let's see…


"Browse," "Watch," "Listen," "Read"?  Oh my god… I know what this is…  It's the dreaded Sierra Operating System!!! Dun dun duuuuuuuuuun!

In this nightmare OS, you have to change your cursor mode depending on the type of file you want to interact with.  Instead of simply double clicking on any type of file to open it, you now have to change your cursor to the appropriate mode depending on the filetype you want to open.  Want to watch that video?  Mouse to the top of the screen, click the "Watch" button, and notice your mouse turns into an eyeball.  Now go and click on the video.  Want to open the word document next?  First mouse up to the top of the screen again and click the "Read" button.  Alternatively, right-click to cycle through all the modes until the cursor turns into a pair of glasses.  Oops!  Clicked past it?  That's alright, just keep right-clicking, it'll come up eventually.

Sound horrible?

Well it is!  But probably not for the reason you're thinking.  The Sierra interface, and for the same reasons, the Lucas Arts interface would make terrible computer operating systems.  After posting the first "Why Your Game is Broken" article in the AGS forums, a number of people mentioned how generally annoying the whole Sierra system is.  I've also seen a few other threads and discussions about the problems with the Sierra and Lucas Arts interfaces, but none of them have gotten to the real heart of why these interfaces are broken.  But before I get to the real issue, let's talk about the most obvious problem with these two interfaces.

The obvious problem: they're annoying!

The Sierra and Lucas Arts interfaces are annoying.  Anyone who can manage to take off their nostalgia-tinted glasses should be able to see that these interfaces are irritating.  Sierra has you mousing up to the top of the screen, selecting a mouse mode, then mousing back down to the object you want to use that mouse mode on and clicking a second time.  Arguably more annoying, the Lucas Arts interface has you mousing down to the bottom of the screen and choosing from a list of around 8 (the number changes depending on game) verbs and then mousing back up. 

We may be used to mousing away from the action to select menu items in any software, or to select tools in graphics programs like Photoshop, but when I'm trying to identify with a character and lose myself in a story, making me fight with the interface is going to take me right out of the game.

The standard Sierra implementation also allows you to right-click to cycle through the modes, but as illustrated in the Sierra OS hypothetical above, this often leads you to miss the mode you were looking for and have to click through a second time.

One recent game, Ben There, Dan That (http://www.adventuregamestudio.co.uk/games.php?action=detail&id=1056), gets rid of the buttons at the top of the screen and relies solely on right-clicking to change the cursor mode.  I've seen other games do this as well, but it always results in a frustration.  I hope from the few comments in the game's thread about this issue, that they've realized their error and will remedy it in their next game.

It could be argued that a well-implemented verb-coin GUI, which lists your available interactions when you click on an object, is a far more elegant solution than the Sierra and Lucas Arts interfaces since it doesn't require your mouse to go flying all over the screen when you want to try multiple interactions on an item.  However, the Sierra interface tops the verb-coin interface in one area: mode permanence.  If I click "Use" on one item and it doesn't work, I can quickly go and try "Use" on another object without changing modes.  With the verb-coin, you need to select "Use" with a second click each time you want to try an object.  Lucas Arts interfaces sometimes lack this mode permanence as well, which is another mark against it in my book.

An interface with mode permanence â€" left-click on anything to use the previously selected mode, right-click on an object to bring up a verb-coin and change the mode that you want to use on the object â€" might be a better solution. This would require fewer clicks than an interface without mode permanence (verb-coin), and require less mouse movement than an interface with fixed mode buttons (Sierra/Lucas Arts).  I haven't tried this out, nor do I know of a game right now that uses it, but it might be a workable interface.  Just throwing that out there.

Another way people can improve the Sierra/Lucas Arts interfaces' annoyance issues, also mentioned in the forum thread responding to the first article in this series, is to add keyboard shortcuts.  These can be useful to players who don't mind playing with one hand on the keyboard rather than relaxing and using just the mouse (no jokes here as to what your other hand may be up to).  However, this fix still has problems.  There are no "Use" "Walk" "Examine" "Inventory" or "Talk" buttons on my keyboard last time I checked, so the developer will have to decide which buttons would be the most intuitive.  Do you use (U)se or (I)nteract?  (L)ook or (E)xamine?  Giving the player more options like also using the 1, 2, 3… keys to represent the verbs can help, but regardless, until the shortcut keys become ingrained into the player's brain so that he can hit the right one without thinking, they're still acting as a barrier to immersion.

I have yet to play a game with either the Lucas Arts or Sierra standard interface that I didn't feel was annoying.  I don't like having to fight with the interface, and I don't like having these things get in the way of my immersion in the game.   I think that some games manage to be great despite using these interfaces, but I almost universally think that these games could have been more enjoyable with a better designed interface.

However, I still think that the whole "annoying" thing is a minor issue with these interfaces when compared to the second issue which is my real sticking point.

The real problem: They're completely superfluous

I'll admit that much of that last section dwelled on opinion and personal preference.  Some people may, for some reason, love either of these interfaces despite (or because of?) these issues.  However, what I'm going to discuss next is not opinion.  It's the hard truth.  If your game is using one of these interfaces, I can almost guarantee that your game is broken.  It's broken because your interface is completely superfluous.

What do I mean superfluous? 

One of the defenses that I always hear for using the Sierra or Lucas Arts multiple-verb interface over, say, the simple left-click to interact/walk, right-click to examine interface is that having multiple verbs gives the player more control over the player's actions.  Let's see how this argument holds up.

I'm playing a game with the Sierra interface.  I have four verbs, Walk, Look, Use, and Talk.  My character is in a room.  There's an elevator with a button and a man standing in the corner.  We know "Walk" doesn't do much.  It's usually only useful for helping us see the other end of a long scrolling room or for moving through an exit into the next room.  "Look" doesn't usually play into puzzles unless you need to examine something closely before you can do something with it.  It's usually only used for helping the player understand his surroundings and adding to immersion.  So the only real verbs that we have for gameplay are "Use" and "Talk."

Let's try "Use" on the man standing near to the elevator.

"I'm not touching him!  That would be sexual harassment!" quips my quirky character.

Oh, right.  I'm supposed to use "Talk" on him.

Let's try "Talk" on the button next to the elevator.

"I don't think buttons are very good conversationalists."

Thanks, quirky character!  I've now learned some valuable information.  I suppose I should try to "Use" the button instead.

With very little exception, every single game that I have ever played using this interface has used the "Talk" on people or other characters and "Use" on everything else.   Let's show that in a handy Venn diagram:


If the set of things that I need to use "Talk to" on is completely separate from the things that I need to use "Use" on, and trying to use the other verb just results in a useless "I can't do that" comment, your game is broken.  There's no puzzle in it.  No thinking required.  Person equals talk, thing equals use.  That's it.  It doesn't make me feel more "in control" of my character.  If anything, I feel less in control, since my character systematically refuses all my commands that don't fall in line with the above diagram.

If you're going to make two separate verbs, there had better be a reason for them.  That is, the set of things that I can use one verb on had damn well better overlap the set of things that I can use the other on otherwise there's no reason to ask me, the player, to specify which verb I want to use.

Think back to that Operating System example at the start of this article.  Remember how terrible it sounded?  Each file had only one obvious use, but I still had to change the cursor mode to tell the computer how to use each file.  Your game works this same way as this system.  It's pointless.

If, on the other hand, you design your game with the peculiarities of the interface in mind, the Venn diagram would look more like this:


Considerable overlap means that there is good reason for me to have different cursor modes.  Imagine if that Sierra Operating System had only one file on the desktop called "Resonance data" and using the "Watch" mode on it would open up the movie files contained therein.  Using the "Read" mode would open up any design documents, or maybe code.  "Listen" would play any mp3s contained within the data.  Now, this OS has good reason to have these cursor modes.  (It's still annoying… but at least it's justified.)

Games using the Sierra interface that fit the second Venn diagram are not broken.  At least not for the reasons I'm discussing now.

However, 90% of the games that I've tried that use the Sierra interface (and that's a conservative estimate) more closely resembles the first Venn diagram than the second.  And if the sets are distinct, then there is absolutely no reason for you to inconvenience me with the previously discussed annoyances.  Combine the separate verbs into one cursor and be done with it.  This has two big benefits.  One, I don't need to right click through an extra cursor mode, and two, you don't have to write so many "I'm not gonna do that, ha ha" messages.

One game that I worked on, Spooks (http://www.adventuregamestudio.co.uk/games.php?action=detail&id=688), fits into this 90%.  I can only remember one time in the game  where you have to use the hand cursor on a character, and that was before you realized he was a character.  After that, you must always use talk on him.  Every other object/character fits into the two distinct "Use" and "Talk to" groups.  I didn't realize it back then, but I do now: Spooks is broken.

Let's see if the Lucas Arts interface does any better:


Nope.  Pretty much all the items are distinct in most games.  Small objects can be picked up.  Big objects can be used.  Characters can be spoken to.  Things with hinges can be opened and closed, and things that are movable can be pushed and pulled.  There may be some overlap here and there (a mechanism that you can use and also push aside to reveal the plug behind it), but for the most part it's all completely superfluous.  Even though the open/close and push/pull circles overlap, you can usually only use one at any given time on an object, for example, open on a closed door, close on an open one.

Again, this doesn't require any real thought by the player.  It doesn't give me more control.  It just makes my character refuse my commands more often.  All you're doing is adding this overcomplicated barrier between me and the game, and if I have to struggle with this slow and unnecessary process just to give commands to my character, I'm not getting immersed, I'm being cut off.

This is just the Sierra Operating System with even more verbs.

So, combine the superfluous verbs into one cursor mode.  Let the computer figure out what I'm trying to do.  It'll usually be right.  If I clicked on a closed door, quite likely I was trying to open it.  If I clicked on a crowbar lying on the ground, yeah, I was probably trying to pick it up.  Clicked on Jerry the bellhop?  Fair chance I was trying to talk to him.  The system works and it works without wedging an interface between me and the character I'm trying to connect with.

As such, I am a strong advocate of the two button system.  A left click interacts with whatever I clicked on in whatever way is likely my intention.  If I click where there is nothing to interact with, that's a walk command.  A right click examines whatever I click on.  This system doesn't require taking my mouse away from the action, it doesn't put a GUI between me and the story, it has mode permanence, and it doesn't require any keyboard shortcuts or other tweaks to make it bearable.  Moreover it is fast and easy for me to communicate my intentions to my on-screen character.

If I were to convert your game over to this simple system, I would have eliminated pretty much every issue we discussed in the section about the annoyances of the Sierra and Lucas Arts interfaces AND if your game matched that first Venn diagram (and seriously at least 90% of them do) I would have managed to do it without sacrificing any of the gameplay or puzzles that you had in your game.

I could do this because your game was broken.

Choosing an appropriate interface

You chose to use the Sierra interface in your game, either because you like that interface or because it came prepackaged with AGS and was the easiest option.  Or maybe you managed to tack on the Lucas Arts interface because you're aiming for the nostalgic value of those old classics.  However, an interface should be tailored to the game, not chosen because of nostalgia value or ease of implementation.  I really like Photoshop's interface, but I wouldn't want to slap it on Firefox, that just wouldn't work.  I really like cover flow in iTunes because it's fun to flip my albums using it.  However, I think it makes a shitty file browser in OSX Leopard.

Your game is broken because you chose an interface for reasons other than how it would lend itself to the gameplay that you designed.

Before you choose an interface for your game, you need to know if it's going to be appropriate for your game.  I think a lot of us amateurs just pick up AGS and go, "Yay, I can make games.  Ok, let's see, default template… It has an inventory!  I'm going to have my character pick stuff up!"  I mean, working within your limitations is an important skill for an amateur game developer, but don't impose limitations on your work that aren't there.  Inventories are fine, but you don't have to use one.  Lots of great AGS games don't have inventory items.  Likewise, you don't have to have multiple verbs just because they're there. 

You don't have to use the interface that comes with AGS.  You don't even have to use an interface that you've seen on any game before.  Figure out the kind of game you want to make, and then figure out the best way that you can let the player interface with it.  Nanobots (http://www.adventuregamestudio.co.uk/games.php?action=detail&id=1040), for example, has a very unique gameplay style and an interface that, I hope, lends itself to that gameplay.  Anna (http://www.adventuregamestudio.co.uk/games.php?action=detail&id=607) also has a very non-standard interface that I think works really well for that type of game.

Before making your game, take a long look at your design, be it on paper or in your head.  What is the best way for the player to play this game without even realizing that he's communicating with the character through an interface?  That's the interface you should be using.  If it isn't, your game is broken.

Very nice article that I wish a lot more people would read, if you know what I mean!  :)

Quote from: Vince Twelve on Sat 19/07/2008 01:13:39In general, you want the hotspot to be clearly pointed out and usually in the upper left hand corner of the cursor.  This is familiar and comfortable to computer users and is usually what I try first in the absence of a clearly defined hotspot. 

See, this is the problem right here. If your adventure game needs to have a cursor with a one pixel mark (so you know exactly where you are pointing), your game is already broken. If you're pointing at a man and you need to differentiate between his nose and his moustache and you can't make out where your cursor is hitting him, you need LARGER hotspots. If those hotspots would collide, you need to increase the size of your art. If you won't put the effort into the game and decide the player needs to click pixel perfect with the cursor on his moustache, then I won't be playing your game. Point is, a cursor shouldn't have to have a tiny clear marker over what part of the cursor that's triggering an interaction. I know you already pointed out the importance of large hotspots, but it can't be stressed enough.

Quote from: Vince Twelve on Sat 19/07/2008 01:13:39
And always, always have an arrow or some pointy thing helping to show you the exact pixel.  This goes double for inventory items. 

I disagree with this. Again, people must design their games so that pixel perfect cursor clicks aren't neccesary. Look at how Sam and Max solved it; by creating a huge red frame around the inventory items when they hovered a hotspot and hand actions for walking and talking etc. that also changed when over a hotspot. If two hotspots are so close together and small that you NEED pixel perfect clicking, there's a huge flaw in your design.


When we're talking cursors and inventory, I'd also like to share my dislike for games whose inventory items disappear from your cursor if you try to combine it with an object on screen. It's unnecessary and time consuming for the player if he wants to try the item combined with the game world, and it bugs the hell out of me. Don't do it people! That goes for you too, Telltale!

Nice article Vince, I know it's given me things to think about on the project I'm working on now :)  In fact, I immediately opened up my GUI and looked for for problems!

Quote from: LeChuck on Sat 19/07/2008 01:48:07
See, this is the problem right here. If your adventure game needs to have a cursor with a one pixel mark (so you know exactly where you are pointing), your game is already broken. If you're pointing at a man and you need to differentiate between his nose and his moustache and you can't make out where your cursor is hitting him, you need LARGER hotspots. If those hotspots would collide, you need to increase the size of your art.

I agree with LeChuck here :)  I don't think anyone enjoys pixel hunts; bad icons can make it worse, but having a pixel hunt there in the first place is already bad design.  If you make hotspots large enough, I think you can get away with icons that are not exactly defined.

As an example, here is a portion of the GUI I'm currently working on:
(http://crabshackgallery.com/misc/guiicons.gif)
After reading the article, I took a look and wondered if these were good icons.  I think the hotspots for the walk (the middle of the X), examine (the middle of the magnifying glass), and talk (the end of the balloon tail) are intuitive, even though they are not all in the same locations or have pixel-precision... perhaps others might not agree, though? :-\
The interact button, on the other hand, is more ambiguous and I'll try to come up with something better (already your advice is paying dividends ;))

Thanks for the article!

I agree with LeChuck. I haven't really paid too much attention where the hotspot of the cursor is. I haven't felt the need when I have a cursor that changes it's appearance over hotspot and have an overlay text that tells me what is the name of the hotspot I'm pointing. But when there is no @OVERHOTSPOT@ or changing mouse cursors I have really frustrated while playing some games. I usually end up clicking blindy everywhere because I don't know what is a hotspot and what is not...

Yup, I definitely should have added in something about making the room hotspots easy to click on.  In the example, Fluxworld, half of the problem is the size of the dinky little guys.  Really, three things should be considered in every game:

1) Well-defined cursor hotspots
2) Mouse over indication
3) Large and easily spotted room hotspots.

In the absence of number 1, numbers 2 and 3 will suffice.  I'd still rather see all three.

Mash: I think all of those are fine except for the interact one which would be something that I'd probably complain about!

With walk cursors, rules I discussed above aren't all that important really, since you're rarely need the character to walk exactly to a specific pixel.  Yours is well defined, and having it in the middle isn't a big deal.

The examine cursor being in the middle of the circle means that you won't be able to see exactly where you're clicking, but the icon is small, so you always know it's in the neighborhood.  Having large room hotspots will negate any problem I might have had to moan about.

The talk cursor is well defined.  The upper-left consistency isn't a hard rule, just a matter of comfort and personal taste.

But yeah, I would advise some kind of redesign on the interact cursor.  Maybe flip the image horizontally and vertically to put the small cog in the top-right corner, then fill in the center of the large cog and make the hole in the small cog the hotspot.  Just an idea.


Thanks for the feedback.  I should make an amendment about the big hotspot thing.  Thanks LeChuck.

I don't think visibility for the hotspot of a cursor is necessary unless the cursor in itself is awkwardly shaped or does not have an intuitive hotspot location (like at the center, or in a place on the image where it would make sense).  Examples Vince gave above, like the huge mirror, are a good example of icons not designed with the player in mind.  Sometimes it hurts your game more to be arty than it does to be unoriginal; sometimes, grandiose and pretty icons are harder to use and more unwieldy than just using a crosshair, so it definitely pays to consider function while making cursor designs.

I gazed over and read most of the article (I spent a lot of time skipping to the pretty pictures! :=) and I rather enjoyed it. I'd say you definitely make a fair argument. I hate not knowing where the hotspot of my cursors is at. If I ever decide to make a game myself I'll definitely keep this one in mind!

Vince: Could you give an example of some good mouse cursors? (or anybody else if you know any)
Only games that have well established cursors that come to my mind are those that don't use a different cursor for different actions, like the simple cross in old Lucasarts games. I'm not nitpicking here, but I am truly interested if someone really has succeeded with all those different kind of cursors.

Quote from: ProgZmax on Sat 19/07/2008 05:25:42
Sometimes it hurts your game more to be arty than it does to be unoriginal; sometimes, grandiose and pretty icons are harder to use and more unwieldy than just using a crosshair, so it definitely pays to consider function while making cursor designs.

Definitely.  I couldn't agree with this statement more!!!  Our voices are as one.   ;)

And I completely agree with your statement about this being needed more for oddly shaped cursors.  A pair of scissors has a natural point.  A pencil could be a good cursor.  A basketball... not so much.  If your cursor is a basketball, you need a little arrowy thing.

Quote from: zabnat on Sat 19/07/2008 12:40:37
Vince: Could you give an example of some good mouse cursors? (or anybody else if you know any)

That's a good idea!

In Nanobots, Erin started with a the same bunch of cursors you see in the final except without the arrows.  After some pressure from me, she changed them to the nicely standardized set you see here:

(http://www.xiigames.com/misc/nanocss.PNG)

The text on the right of the cursors only appears when you mouse over something usable.  For example, the shelf on the wall is analyzable, so the analyze text pops up while mousing over it.  But the shelf wouldn't be realistically pushable because it's mounted on the wall, so the push text won't pop up.  This slightly and subtly simplifies the huge number of things the player has to mess around with right off the bat.

I could name lots of other games with good cursors.  Like I said, the majority of games that really break these rules are first games or low-quality in most other areas as well. 

One recent game that I remembered that had nice cursors was Ben There Done That.  Each cursor has a little wedge in the top-right corner.  Very simple, very clean, and the cursors are attractive.  The one weird thing about BTDT's cursors, though, is that the inventory items all have the little wedge up there too when they're sitting in the inventory.  The authors should have used one of the few methods available to make the wedges invisible in the inventory, but visible when they become cursors.  A very minor complaint though!  I'd rather have the wedges in both places than not at all.

Well, can't agree with you more. It wasn't so hard trying to change the cursors to some of your choice though. Sierra had big cursors less pixel hunting and no hotspot diplay label, thus you could pick bigger area of the screen. But this isn't designed for all games. It's not Chris problem if people can or didn't had the time to recognise that bloody default sierra interface can't or isn;t completely suitable for their creation/s.

These were the initial cursors I made for Heartland Deluxe:

(http://homepage.eircom.net/~limpingfish/Cursors.png)

Only the first three made it into the final game, but you can see that I didn't stray far from the traditional "pointer"-type design.

I dropped the ball on the inventory item hotspots, though. :-\

Ideally, when an inventory item was chosen, it would have been displayed in the center of the first cursor, and the hotspot would be pretty clear.

Unfortunately, I lacked the...um...technical ability to do so...

I specifically remember loving the slick cursors for Heartland, even in it's original OROW version.  Though, yeah, having the inventory cursors match the others would have been nice.  If AGS allowed two sprite slots to be defined for each inventory item, one for the inventory, one for the cursor, it would be easier for people to make sure they had well-defined hotspots on their inventory cursors.

Found this tracker entry. (http://www.adventuregamestudio.co.uk/tracker.php?action=detail&id=216)  Wish I could bump it.

Some very good points here. I've seen several games with very good ideas behind them that had their gameplay lessened (or, in some cases, ruined) by a poor interface. Not seeing where you click, well, that's not good. My personal quibble is keyboard shortcuts.

This isn't just AGS, either. I can think of a few commercial games that suffer from the same problems. And there's always the interface Hall of Shame (http://homepage.mac.com/bradster/iarchitect/shame.htm).

Quote from: Radiant on Sun 20/07/2008 11:57:41
My personal quibble is keyboard shortcuts.

Oh, I'm gettin' to that one!

Quote from: Vince Twelve on Sat 19/07/2008 01:13:39
Inventory cursors too!

With some very simple scripting, you can use one big attractive sprite to represent the item in your inventory and one smaller, iconic sprite (preferably with a big arrow pointing to the upper left hand corner) as the cursor.

Is there a tech forum post with how to script this? The article was very well written and extremely relevant. I think I link would improve it further.

Also, I would go so far as to say the default AGS cursor system is broken in itself. Because of the right click = cycle function, it takes ages to do anything, and if you miss a cursor when cycling you have to go round again (unless they are all displayed in the GUIs, but that makes cycling pointless anyway...). The first thing I did was swap this for a Broken Sword/BASS system of one cursor, with right and left click functions.

Hah, cursor pointer thing, this is something I told Grundislav after every Ben Jordan game....
Many other people too.

This is something AGSers tend to ignore, not sure why.

QuoteAlso, I would go so far as to say the default AGS cursor system is broken in itself.

I'd hardly call a system broken that does the job its meant to.  Inelegant, perhaps, but hardly broken.

Quote from: ProgZmax on Sun 20/07/2008 18:31:35

QuoteAlso, I would go so far as to say the default AGS cursor system is broken in itself.

I'd hardly call a system broken that does the job its meant to.  Inelegant, perhaps, but hardly broken.

OK, a little harsh. But, using 'broken' in the same manner as Vince I'd say it is. It does the job yes, but as mentioned in the article clicking lots of times to search for something renders a game broken, partly because this takes a lot of time. The default system takes a lot of time too, and can be very frustrating - that was the point I was making.

Quote from: markbilly on Sun 20/07/2008 13:41:22

Quote from: Vince Twelve on Sat 19/07/2008 01:13:39
Inventory cursors too!

With some very simple scripting, you can use one big attractive sprite to represent the item in your inventory and one smaller, iconic sprite (preferably with a big arrow pointing to the upper left hand corner) as the cursor.

Is there a tech forum post with how to script this? The article was very well written and extremely relevant. I think I link would improve it further.

Also, I would go so far as to say the default AGS cursor system is broken in itself. Because of the right click = cycle function, it takes ages to do anything, and if you miss a cursor when cycling you have to go round again (unless they are all displayed in the GUIs, but that makes cycling pointless anyway...). The first thing I did was swap this for a Broken Sword/BASS system of one cursor, with right and left click functions.

I absolutely agree with this. I actually hate both the LucasArts-style interface and the late Sierra-style interface, because they both force you either to take ages cycling through cursors or to move your mouse all over the screen to select a new one. Charlie Foxtrot and the Galaxy of Tomorrow is one of my absolute favorite AGS games, but the interface drove me crazy - there were billions of different interactions to try, but I had to move the mouse to the top of the screen, select the cursor I wanted, and then move it back to the object I wanted to interact with, even if I was interacting with the same object more than once. I'm having fun with Ben There, Dan That, but I don't even have the option of going directly to an interaction mode - I have to cycle through all of them, and if I miss the one I want, I have to go all the way through again. These are both really good games, but I would like them a lot better if the interface was more streamlined.

My game has a simple right-or-left-click interface (like Beneath a Steel Sky's) exactly because of this. If you want to make a game with a hundred different cursor modes, make it so that clicking on a hotspot opens a GUI with the other options, like in 7 Days a Sacrifice or The Longest Journey. THESE ARE GOOD SOLUTIONS TO INTERFACE PROBLEMS! There's really no excuse to keep using these terrible interfaces when you have seen and played games with actual decent, well-functioning, usable interfaces.

Quote from: ProgZmax on Sun 20/07/2008 18:31:35
I'd hardly call a system broken that does the job its meant to.  Inelegant, perhaps, but hardly broken.

Indeed, though maybe also due for retirement? Seeing as AGS supports 1024x768 now (though 800x600 didn't exactly result in an influx of higher resolution games), maybe an alternate set of "default" icons could prove handy?

Or keep both types, and split them between resolutions.

Quote from: Makeout Patrol on Sun 20/07/2008 19:51:07
I actually hate both the LucasArts-style interface and the late Sierra-style interface, because they both force you either to take ages cycling through cursors...

This is my preferred interface, actually. It's a breeze once you get the feel over what cursor comes after another, but until then it's a bit messy. It will always be messy if the cursors dont make sense.

I know a lot of people prefer keyboard shortcuts to change the cursor in games like Sam and Max, and seeing as I'm making a similar interface, which keys usually go with what interactions?

u* = use
w = walk
l = look
? = active inventory

This sound about right? I'm figuring people want to be able to open the inventory with the i button, so what button should I script to select the active inventory (if any)? Suggestions? Or am I all wrong in my assumptions?

* This is all AGS' de-l33tifying script... sigh, do we really need one? It also corrupts my homepage urls...

Quote from: LeChuck on Mon 21/07/2008 01:25:54

Quote from: Makeout Patrol on Sun 20/07/2008 19:51:07
I actually hate both the LucasArts-style interface and the late Sierra-style interface, because they both force you either to take ages cycling through cursors...

This is my preferred interface, actually. It's a breeze once you get the feel over what cursor comes after another, but until then it's a bit messy. It will always be messy if the cursors dont make sense.

I know a lot of people prefer keyboard shortcuts to change the cursor in games like Sam and Max, and seeing as I'm making a similar interface, which keys usually go with what interactions?

you* = use
w = walk
l = look
? = active inventory

This sound about right? I'm figuring people want to be able to open the inventory with the i button, so what button should I script to select the active inventory (if any)? Suggestions? Or am I all wrong in my assumptions?

* This is all AGS' de-l33tifying script... sigh, do we really need one? It also corrupts my homepage urls...

I Have No Mouth and I Must Scream went this route, and it was an improvement, though still not optimal in my opinion. If you're doing this, I'd advise you to implement more than one binding for each mode - maybe the u / w / l / i keys for people that are used to that, as well as 1, 2, 3, and 4, so that I can just switch with one hand and without needing to look at the keyboard.

I still maintain that having a GUI pop up when the player right-clicks on a hotspot with buttons for all of the available interactions right there is the best way to do this, however.

Quote from: Makeout Patrol on Mon 21/07/2008 01:36:11
I Have No Mouth and I Must Scream went this route, and it was an improvement, though still not optimal in my opinion. If you're doing this, I'd advise you to implement more than one binding for each mode - maybe the you / w / l / i keys for people that are used to that, as well as 1, 2, 3, and 4, so that I can just switch with one hand and without needing to look at the keyboard.

Actually, now that I think about it, I think both having multiple  keyboard shortcuts and an optional GUI that will pop up in one of the corners showing all the actions is the best way to go.

I'd still like to know what letter people associate with the active inventory being used, though.

I can see that I'd better get to work on Part Two which is all about the inherent flaws in both the Sierra and Lucas Arts standard interfaces before you guys bring up everything I was going to discuss!

Quote from: LeChuck on Mon 21/07/2008 02:51:14
Actually, now that I think about it, I think both having multiple  keyboard shortcuts and an optional GUI that will pop up in one of the corners showing all the actions is the best way to go.

I agree.
The best part is that in AGS, it'll take all of five minutes to code. Just have your on_key_press code call whichever interface_click function fits best.

Quote
I'd still like to know what letter people associate with the active inventory being used, though.

(U)se, and have (T)ouch be the regular hand icon.

Quote from: Vince Twelve on Mon 21/07/2008 03:11:19
I can see that I'd better get to work on Part Two which is all about the inherent flaws in both the Sierra and Lucas Arts standard interfaces before you guys bring up everything I was going to discuss!

Cool. Actually I don't think they're all that bad, but perhaps that's because I grew up with them. It is quite important to note that (unlike many fangames who apparently use the same GUI) both Sierra and Lucas have numerous keyboard shortcuts. For instance, pretty much any Sierra game has F5/F7/F9 for save/load/restart, and TAB for inventory; and Lucas games like Maniac Mansion have keyboard controls for inventory scrolling (hint: use the mouse wheel, nowadays!)

Right-mouse-button cursor cycling works with just a few cursors (say, the standard five) but gets progressively more annoying with more of them. Verb coins can also be quite annoying if poorly implemented (click, hold, drag, release is NOT something you want for every interaction, imho)

One solution to the hotspot dilemma can sometimes be beta-testing with someone who hasn't worked on the game, watching them but not giving away the puzzles.  Watching them would allow you to see frustrating problems that might be overlooked by normal beta testing.

When I've needed to use small items or characters (for instance, a 4-6 pixel mouse on the screen) I've usually made sure that the click didn't have to be exactly on the item in question, it could be off by a few px without problems.  I specifically remember this in KQ6, where Alex has to talk to the Ferryboat Captain.  You've only got a second or so to talk to him, and you have to hit just the right spot.  It took getting on my dad's AOL to figure that one out, back in the '90s, I just guessed he didn't want to talk, not that I hadn't managed to click on him in the right spot!

Quote from: LeChuck on Mon 21/07/2008 02:51:14
I'd still like to know what letter people associate with the active inventory being used, though.

You could make the letter "U" be the "touch" the first time it is pressed and "use inventory" the second (and then back to touch the 3rd and so on)

Actually, the Nanobots cursors underwent quite a few changes.

See here:

(http://i8.photobucket.com/albums/a18/The_Ivy/Nanocursors.png)

A. Preliminary ideas for cursors. Notice how the "walk" icon was originally more like a target.

B. I thought it would be cool to put a box around all of them, with the light green sections moving around like a marquee. Then Vince was like, "Where are the hotspots?" And I was like, "Top left corner?" And he was like, "There's nothing telling the player that." And I was like, "Oh."

I still kind of like the "Loading" one though. :)

C. I tried to go with something a bit more robot-y so I added the green corners to everything. Vince was like, "That's better, but look at the cursor for 'Go to.' Is the hotspot the arrow, or the green corner?" Back to the drawing board.

D. Finally, we had a nice clean set of cursors that were, above all, consistent. They didn't keep the player guessing. I'd actually be quite happy if no one had noticed the cursors during the game.

E. I included this to show you how I changed Strongbot's GUI; it was functional before, but waaaay too small to click on. I guarantee I would have had a few complaints if I hadn't changed it.

F. The cursor set was by now consistent enough that I could add new items without having to worry about their design (like this inventory item from the tutorial).

G. By the way, we had to make cursors for every possible combination of chemicals, even the ones that didn't do anything in the game. I bet you didn't see many of these, eh? We also had to make Chembot "pouring" animations for all of these, and design a system that would let you mix up to three colours at once in the beaker. While I was hitting my head against my keyboard for not thinking of that in advance, Vince swept in like a ninja and solved the whole thing. I'm surprised that he still speaks to me. ;)

Part one (http://xiigames.com/2008/07/19/why-your-game-is-broken-part-one-cursor-confusion/) has been moved to my blog (http://www.xiigames.com/) with a little addendum based on some of the good feedback from this thread.

I put Part two into the first post (http://www.adventuregamestudio.co.uk/yabb/index.php?topic=35142.msg460095#msg460095) in this thread.  Instead of posting it on my blog at the same time, I'm going to modify it based on the discussion here (if there's any to be had) and then post the final version to my blog later.

So please read and give me feedback!  Down with the Sierra and Lucas Arts interfaces!

I dunno what to say. I really dislike the 2 button system. And prefer the Sierra/Old LucasArts GUI to a verbcoin, mostly because the clunkiness of the verbcoin (1 click, move to desired action, another click seemed counter-intuitive to me, and the verbcoin usually blocked whatever I was working on). But perhaps that is personal preference. In a perfect world, my use, pick up and talk to would all work, and would all be well catered for. Or someone would come up with an awesome new GUI that would easily allow you to do all possible things you could think of doing to all possible objects on screen (a sort of Lure of the Temptress, or those LEGEND 1/2 text adventures with some magical easy to use/navigate GUI).

Don't worry, I'm working on it.

I do have to say that all the points brought up are good. But, they are all opinionated! You can go ahead and say the Sierra and Lucas system make broken games, but only if the creators swing them the wrong way.
If you don’t like them, fine, but I think they are the most successful interfaces ever made (not so much Lucas). If applied correctly the Sierra interface can form some of the most clever, challenging and logical puzzles to be found. The verbcoin, bah, leaves no challenge and can be tedious most of the times. That’s my little opinion and nostalgic view on favourites (and yeah, not much of an argument, I have more, but its not why I write).
The real issue (and its painful playing most of these games) is plain and simple. NONE HAVE BEEN UTILISED PROPERLY. Every game I come across using most any interface has never filled in all the gaps.

Quote from: Vince Twelve on Sat 19/07/2008 01:13:39

If the set of things that I need to use "Talk to" on is completely separate from the things that I need to use "Use" on, and trying to use the other verb just results in a useless "I can't do that" comment, your game is broken.  There's no puzzle in it.  No thinking required.  Person equals talk, thing equals use.  That's it.  It doesn't make me feel more "in control" of my character.  If anything, I feel less in control, since my character systematically refuses all my commands that don't fall in line with the above diagram.

And this is why I say these are good points that have been brought up. I don’t agree that the interfaces are flawed. I believe that they have not been master by any amatures out there yet. Ya! The games are broken! But please, someone out there prove that the Sierra and Lucas systems can be used to make a great amature adventure game!

For more Email: advenreview@ymail.com

Hmm, lots of things to think about :)

One thing I've always disliked about the default Sierra system is having to open a new window to get to the inventory and having to press "OK" to select an item.  That's at least 3 clicks, 5 if you accidentally select the wrong item at first and have to go back to the regular cursor in order to select another.  Far too many!

On the other hand, I personally don't have a problem with the Sierra cursor interface (assuming that there aren't too many extra cursors to cycle through), and I don't believe it is inherently flawed--it's just that few games take advantage of having separate cursors for Talk and Interact.  I agree that the Broken Sword interface is the best one if that's the case (that's the one I've used for my games so far).

That said, I still think it could be improved.  One of the things I like about the BS interface (which was also mentioned) is the default walk when you're not clicking on a hotspot, and I was thinking that this could be combined with the Sierra style to make a more comfortable interface.  You would have Examine, Interact, and Talk cursors that would cycle normally, but when you click anywhere that's not a hotspot, the cursor would briefly turn into the walk cursor (just to indicate that the walking action is beginning), then default back to whichever cursor mode you were on before.  It would eliminate the walk cursor, making cycling much faster and thus less of an annoyance if you accidentally miss a mode.  I could swear some games do this, but I can't remember which ones (possibly The Longest Journey).  I can't think of any drawbacks to that interface at the moment (assuming, of course, that separate interact/talk actions ARE needed).

Keep the articles comin', Vince!  Reading this one made me think of the default-walk-thingy I just described, which is what I will probably use in my next game unless I can think of something better--these articles have been very timely, I must say ;)

Hmmm...yeah, I guess a major point is that those games are broken only if more than 1 cursor is not used on most objects. But if there is more than 1 interaction on a certain object, and I can't do them due to lack of options, I'd be annoyed. Like if I clicked the ACTION button of the 2 button system on a person, and he was pushed off a cliff instead of me getting a chance to talk to him.

I have to agree with most of what you're writing here. Let me add a suggestion: For those games where one or two items can give meaningful responses to more than one interaction, I think a verb coin would be appropriate -- as long as this verb coin appeared only when more than one interaction was possible (and was disabled if one of these interactions no longer were valid, of course).

I'm not really convinced this is too much of a problem.  I think the main reason people like to have multiple actions/verbs in their games is because it provides the player with an illusion of depth.  And I would say most of us fall for this illusion.  If a game simply had interact on the left mouse button and look on the right button, then one might come away from completing the game with the sense that it was a bit too easy.

I'm not saying that we should riddle our games in verbs and cursors to create a false illusion of complexity, but I'm tempted to think there is something satisfying in trying lots of different combinations of actions; the more you are clicking on the mouse, the more you feel as though you are doing something rather than simply watching an interactive movie.

In the real world every action has a reaction, but in an adventure game world it would be boring if could only select an action that would be guaranteed to have a reaction that would progress the game.  One of the beauties of the adventure game is the trial and error - the red herring.  If you only have a choice of two buttons for every object or person then you are severely decreasing the scope for this kind of gameplay device (and the longevity of the game itself).

Quote from: Stupot on Fri 25/07/2008 16:39:25
I'm not saying that we should riddle our games in verbs and cursors to create a false illusion of complexity, but I'm tempted to think there is something satisfying in trying lots of different combinations of actions; the more you are clicking on the mouse, the more you feel as though you are doing something rather than simply watching an interactive movie.

The problem I see with having more verbs/cursors is that in order to keep immersion up there has to be a unique response for almost every interaction, or else you feel even more restricted than if the options weren't there at all since you see you CAN do them, but the game's not LETTING you.  However, if the varied responses are present, than this could be a big boost to the game--it all comes back to choosing the right interface for your game :)

Quote from: MashPotato on Fri 25/07/2008 17:16:14

Quote from: Stupot on Fri 25/07/2008 16:39:25
I'm not saying that we should riddle our games in verbs and cursors to create a false illusion of complexity, but I'm tempted to think there is something satisfying in trying lots of different combinations of actions; the more you are clicking on the mouse, the more you feel as though you are doing something rather than simply watching an interactive movie.

The problem I see with having more verbs/cursors is that in order to keep immersion up there has to be a unique response for almost every interaction, or else you feel even more restricted than if the options weren't there at all since you see you CAN do them, but the game's not LETTING you.  However, if the varied responses are present, than this could be a big boost to the game--it all comes back to choosing the right interface for your game :)

Absolutely.  If the author is willing to put time into writing the extra speech then, as you say, it can really add to the immersion of the player and the depth of the story.

Vince, I like how you mentioned that "Anyone who can manage to take off their nostalgia-tinted glasses should be able to see that these interfaces are irritating." Sadly, a lot of adventure games are still made for adventure gamers, which is good for the fans, bad to move such a rich genre forward and return it into the mainstream.

In this day and age, the Sierra and LucasArts interfaces need to be abandoned. They were practically made-up right before everyone else (like your mum) started using a computer with a mouse and a GUI. Sure, we had the Macintosh OS since 1984, but sadly it never caught up to the masses. Windows98 was probably the most widely spread OS, and came out only 10 years ago. Not to mention that the Sierra/Lucas interface is pretty poorly made (and does not deserve to be called designed except maybe the verb coin in Full Throttle/CMI, which was, in a way, borrowed for The Sims, one of the most popular games ever made.).

Another important thing that needs to be mentioned is that the poor interface just leads to lousy game design. I often see a lot of puzzles where the solution is to figure out which verb needs to be used next. So a lot times adventure games become a guess-the-interface/guess-what-the-designer-was-thinking-but-is-not-really-clever-nor-logical disaster. There are no other game genres out there where you have to use one single action one time in the middle of the game, ever, because it's too obscure and frustrating to do that. Instead, the genre should focus more on teaching the basics first (what are all the possible actions the player can do and when they are appropriate to use them),  then allowing to solve problems not through guessing and doing things in a proper order but from truly using the player's intelligence. Designing the puzzles needs to be something completely independent of the interface. After all, the gui is just the means for how the player can interact with your game. It can make your game more enjoyable and fun, but it can't really make your game's design better.

I agree wholeheartedly about the Sierra system, but not so much about the Lucasarts system, the reason being keyboard shortcuts. When I first played Monkey Island and Fate of Atlantis, yes, the interface was annoying and took too much time to move my mouse back and forth. So that's why I always had one hand on the keyboard and one hand on the mouse, and it make the GUI experience streamlined, and close to perfect, in my opinion. My left hand would pick up the keyboard shortcuts instantaneously, allowing me to switch between cursor modes without ever having to touch the GUI at the bottom, save for perhaps inventory items. It was extremely intuitive, in my opinion.

About the 2-interaction system (left click interact, right click look, or vice versa), I used to always disagree with it, seeing how it limits the players choices to only two interactions, thus dumbing down the game and guessing at what the player was trying to achieve with use. Recently though, I've come to agree with what you've said with your article, in that it's basically a more efficient version of the current Sierra system of Look-Talk-Interact, because Talk and Interact are 98% of the time used separately in games.

I dislike verb-coins for the reason of the "Holding down to bring it up" aspect. If I have to wait a second for the GUI to pop up, that's too long for me. However, verb-coins that do pop up instantaneously (ala 7 Days A Skeptic), I love. Those seem to be close to the most improved version of the Sierra interface.

p.s. Vince, did you get my PM?  :)

Stupot, I could agree with you if it weren't for that adventure games have at least two other aspects to add to the trial and error: inventory items and advanced nested dialogues.

I agree that complexity in a game can be good and make it feel more realistic, but I really don't see how it makes a game better in any way that I have the opportunity to speak to the grass only to find that it surprisingly enough doesn't talk back.

QuoteThe Sierra and Lucas Arts interfaces are annoying.  Anyone who can manage to take off their nostalgia-tinted glasses should be able to see that these interfaces are irritating.

Blanket statements like this really don't make for good supporting arguments because they simply aren't true for everyone (myself included).  I happen to like the multiple-verb functionality of the Lucasarts interface because it gives the player a clear idea of the scope of their abilities (instead of a single click that does everything necessary depending on context, which I also happen to find to be very boring).  I do think the original interface (Maniac Mansion, for instance) had redundant verbs, though.  Similarly, I think the Sierra system works and isn't 'broken' (neither system really is), though with a Sierra structure you have to pay attention to cursor design instead of having obvious words to indicate options available.

QuoteOne of the defenses that I always hear for using the Sierra or Lucas Arts multiple-verb interface over, say, the simple left-click to interact/walk, right-click to examine interface is that having multiple verbs gives the player more control over the player's actions.

I think you're missing the most important aspect of more options:  more potential.  The more options you provide, the more the potential for control rises (in proportion to the potential for confusion).  The problem you should address is how so few people seize this potential and make use of it instead of just having some verbs that may only be used once in a game.  Still, I don't really think this is a problem because my goal as a designer is more to provide an interactive story experience than to present the player with a ton of redundant acts they can perform, and my games have been fairly modest in that area up until now.

One of the games I'm making doesn't even HAVE mouse support or a user interface, just keyboard controls and a tiny, tiny inventory display, and in spite of the simplicity of using arrow keys to get around and two keys to pick up/use/talk or use an item on yourself I guarantee people will complain that it's either too complicated or that they'd rather have a GUI.  I mean we can argue back and forth on this subject all day long, but in the end you have to defer on the side of the masses, who clearly seem to prefer more over less, even when the 'more' has very little content.

The thing about multiple-verb interfaces is that the more verbs you have more you need to make those verbs have a coherent influence on the game world. Therefore every verb has to have an influence on every object/person in that world; or else you left with lots of "You can't **** the ****" messages, which in my opinion, are incredibly annoying. Or lots of blank interactions. We know you can't talk to grass, but to use the talk verb on it with no feedback is just as annoying as being told you can't talk to the grass.

I agree with Vince that the Sierra and LucasArts verb GUIs are only still used for nostalgia purposes. You want to make a game that plays the same as early adventure games, so you use the same interface. No big deal. But by doing so, you're exposing yourself to all the foibles inherent to those interfaces. I don't think I'd label them "broken", but certainly "primitive", and therefore limiting.

In my opinion, the "look" verb interaction is a leftover from text adventures, when the game world was described rather than shown. Think about it:

A modern graphic adventure. There's a character sprite on the screen. He's wearing a trenchcoat and hat, and hiding in the shadows. You select the "look" cursor/verb, click on him, and the game says "That guy looks shifty." Well, I knew that from looking at him myself, so "looking" at him in-game was a waste of time. An oversimplification, maybe, but if you continue to distill the look verb down to the core of it's function, you may find that you could do away with it altogether.

I agree with everything in the new article, with one exception. I absolutely agree that the Sierra and LucasArts interfaces are irritating as hell, and I think that anybody who doesn't have fond memories of the original games that used them will agree. (I myself have a soft spot for the early Sierra text-input interface, which is clearly terrible, but which was in some of the first games I ever played.) Seriously, I know this is a 'blanket statement,' but even if you're playing a game where multiple interactions are necessary, wouldn't you prefer an interface that doesn't require you to move your mouse all over the screen? I know that a lot of people have a soft spot for these interfaces, but it has to be said - there are much better ways of doing it. These interfaces were implemented in some great games, but they were designed back in the day when games were made by two programmers sitting together in a garage, without any consultation with usability experts.

People are responding by claiming that a two-option interface limits the puzzles. I disagree, for the following reasons:
1) As Vince Twelve points out, more than two options are almost always useless. There are only two options already; all that these extra cursor modes add to the game is a bunch of mouse movement and mode-switching. Babar, you say that you'd be irritated if the 'interact' command resulted in pushing a character over a cliff. When would that ever happen? Why would anyone ever implement that? How does that even make sense? If I'm 'interacting' with a character, I think it's pretty clear that I'm trying to talk to them.
2) If your puzzles are solved by choosing the correct interpretation with an object, your puzzles are terrible. Aren't puzzles supposed to be solved by abstract problem-solving? Do we not have inventory items to use? Are we incapable of coming up with puzzles that are more involved than "you have to talk to this man, not USE him"?

Very few games have used the multiple interactions to great effect, though they are out there. A lot of the appeal of Charlie Foxtrot and Ben There, Dan That came from the messages when you used the 'wrong' interactions on objects. I still didn't like the interfaces, but if they'd had a verb-coin GUI I don't think I would have been too heartbroken about the superfluous modes; the creators were using them, even if they weren't using them 'in-game,' if that makes any sense.

I actually love the one-click interface. It's the one I've got in the game I'm finishing up right now.

I also agree that the Longest Journey-style verb coin GUI is the best option if you truly do need more than one interaction, though this one is with reservations. You propose that a good verb-coin interface could be created if you selected a mode that then 'stuck' when you left-clicked. If you're doing this, it's absolutely essential that left-clicking always 'walks' when you're not clicking on a hotspot - there's a reason that 'walk' is the default cursor mode in most interfaces. If you're going this route, however, and this also applies to the two-button interface, you need to make it abundantly obvious to the player that the mouse is over a hotspot. You can do this by changing the mouse graphic (like in The Longest Journey) or by displaying a string with the hotspot's name next to the cursor (like in Beneath a Steel Sky). I also think that a good verb-coin GUI would just not show interactions that won't logically have any effect on the hotspot.

So aside from that, this little article basically says what I've always been thinking.

Quote from: Makeout Patrol on Fri 25/07/2008 20:02:52
Babar, you say that you'd be irritated if the 'interact' command resulted in pushing a character over a cliff. When would that ever happen? Why would anyone ever implement that? How does that even make sense? If I'm 'interacting' with a character, I think it's pretty clear that I'm trying to talk to them.

Not to the designer of the game. The designer of the game views that specific character as an obstacle in the game, and the solution to the puzzle is to push him off the cliff. I just got there. I don't know anything about a puzzle, so I attempt to talk to the character. He gets pushed off the cliff. Just an example. Hell, flip it over. I wanted to push the character off the cliff (a fairly simple and easy way to solve the puzzle), but I am only able to talk to him.
See, what matters here is not the GUI, but the level of detail in the game. With good detail, either one of those systems would be 'unbroken', because each of those interactions would have an effect. A good game would allow me to do something like that. A bad game would not even give me a response to attempting that action. An 'efficient' and irritating game would not even consider it, because there was only 2 buttons to click, and one of them did whatever the programmer decided was the way the puzzles in the game would be solved.
Similar stuff has happened to me in some 2-click games (I am remembering KQ7 and Beneath a Steel Sky).

Quote from: Makeout Patrol
I also agree that the Longest Journey-style verb coin GUI is the best option if you truly do need more than one interaction,

Wow, TLJ had a verb coin?  My porous memory has failed me yet again ;)

Quote from: LimpingFish on Fri 25/07/2008 19:48:17
In my opinion, the "look" verb interaction is a leftover from text adventures, when the game world was described rather than shown. Think about it:

A modern graphic adventure. There's a character sprite on the screen. He's wearing a trenchcoat and hat, and hiding in the shadows. You select the "look" cursor/verb, click on him, and the game says "That guy looks shifty." Well, I knew that from looking at him myself, so "looking" at him in-game was a waste of time. An oversimplification, maybe, but if you continue to distill the look verb down to the core of it's function, you may find that you could do away with it altogether.

I think when the Examine interaction is superfluous (and I agree that it can be in some situations), it's often to do with writing.  Describing things exactly as they appear does seem a bit redundant, but including some background information, or adding a tinge of the character's personality to the description can provide more depth.  For example, instead of saying "That guy looks shifty", the character might add "I wonder if he's got something to hide", "I think I recognize him from a line-up" if he's a detective, or "I'd better stay away from him" if he's a timid character (well, nothing that lame, but you get my point ;)).  It's perhaps one of the reasons why the Examine option has been eliminated in many Myst-style first-person adventure games, where the character has no real personality (since the character is the player) and the graphics are detailed enough that the appearance of things are self-evident.

I agree, but it's a conscious design decision to use the look/examine interaction in such a way, and more or less reinforces what I was saying about nostalgia.  We're used to being able to use the look/examine interaction to gain extra information not visually available on-screen, but this information was not otherwise available because of the limits of the graphics used.

But, regardless of graphical advances, it continues to be used in this way, because, right or wrong, current adventure games more or less adhere to a decades old template.

Quote from: Babar on Fri 25/07/2008 20:29:19

Quote from: Makeout Patrol on Fri 25/07/2008 20:02:52
Babar, you say that you'd be irritated if the 'interact' command resulted in pushing a character over a cliff. When would that ever happen? Why would anyone ever implement that? How does that even make sense? If I'm 'interacting' with a character, I think it's pretty clear that I'm trying to talk to them.

Not to the designer of the game. The designer of the game views that specific character as an obstacle in the game, and the solution to the puzzle is to push him off the cliff. I just got there. I don't know anything about a puzzle, so I attempt to talk to the character. He gets pushed off the cliff. Just an example. Hell, flip it over. I wanted to push the character off the cliff (a fairly simple and easy way to solve the puzzle), but I am only able to talk to him.
See, what matters here is not the GUI, but the level of detail in the game. With good detail, either one of those systems would be 'unbroken', because each of those interactions would have an effect. A good game would allow me to do something like that. A bad game would not even give me a response to attempting that action. An 'efficient' and irritating game would not even consider it, because there was only 2 buttons to click, and one of them did whatever the programmer decided was the way the puzzles in the game would be solved.
Similar stuff has happened to me in some 2-click games (I am remembering KQ7 and Beneath a Steel Sky).

All right, I agree, in that case it would be better to have a verb-coin GUI. I still think that there's pretty clearly only one way that you can 'interact' with any object in the vast majority of adventure games, though.

Quote from: MashPotato on Fri 25/07/2008 20:32:09

Quote from: LimpingFish on Fri 25/07/2008 19:48:17
In my opinion, the "look" verb interaction is a leftover from text adventures, when the game world was described rather than shown. Think about it:

A modern graphic adventure. There's a character sprite on the screen. He's wearing a trenchcoat and hat, and hiding in the shadows. You select the "look" cursor/verb, click on him, and the game says "That guy looks shifty." Well, I knew that from looking at him myself, so "looking" at him in-game was a waste of time. An oversimplification, maybe, but if you continue to distill the look verb down to the core of it's function, you may find that you could do away with it altogether.

I think when the Examine interaction is superfluous (and I agree that it can be in some situations), it's often to do with writing.  Describing things exactly as they appear does seem a bit redundant, but including some background information, or adding a tinge of the character's personality to the description can provide more depth.  For example, instead of saying "That guy looks shifty", the character might add "I wonder if he's got something to hide", "I think I recognize him from a line-up" if he's a detective, or "I'd better stay away from him" if he's a timid character (well, nothing that lame, but you get my point ;)).  It's perhaps one of the reasons why the Examine option has been eliminated in many Myst-style first-person adventure games, where the character has no real personality (since the character is the player) and the graphics are detailed enough that the appearance of things are self-evident.

I completely disagree that 'look' is outdated. Sure, it can be useless if you're just providing information that the player already knows ("It's a door!" "Her shirt is yellow!"), but most games provide some important information through the 'look' command ("This door leads to the backyard"). Additionally, although "I can't talk to that poster!" doesn't add anything to the game, "It's a poster for 'Casablanca'... I know, that's the one where Gregory Peck falls in love with Audrey Hepburn, the runaway princess, while they're in Rome!" tells us both about the character that owns the poster and the character that's narrating, and really does serve to add depth to the game world.

While I don't deny a 2 buttons interface would be better for the 90% of adventure games that do not use the LucasArts and Sierra GUIs at their full potential, I don't think the Sierra and LucasArts GUIs are not at fault here, what is at fault is the game makers. Ideally, all verb interactions should result with either game progression, jokes or comments that tell something about your protagonist's personality or the situation.

We should encourage people to raise the game experience so it fits their GUI rather than lower the GUI so it fits their game experience, to work harder to come up with funny or insightful reactions to useless interactions, to open the door to new puzzle possibilities by coming up with new verbs, something most LucasArts and most Sierra designers, with their text parser roots, weren't affraid to do.

QuoteWe know you can't talk to grass, but to use the talk verb on it with no feedback is just as annoying as being told you can't talk to the grass.

I disagree with this because if you get no response then you immediately know there's nothing more going on.  Normal people don't talk to grass, and reducing the number of button presses to a single click on everything wouldn't improve this if you 'wanted' to interact with the grass in this way.  Simply put, I'd rather just have no response at all than a canned 'you can't do that' response, because being told 'you can't do that' is annoying and a waste of time, and in my view at least, breaks immersion. 

Quote from: blueskirt on Fri 25/07/2008 23:19:02
While I don't deny a 2 buttons interface would be better for the 90% of adventure games that do not use the LucasArts and Sierra GUIs at their full potential, I don't think the Sierra and LucasArts GUIs are not at fault here, what is at fault is the game makers. Ideally, all verb interactions should result with either game progression, jokes or comments that tell something about your protagonist's personality or the situation.

We should encourage people to raise the game experience so it fits their GUI rather than lower the GUI so it fits their game experience, to work harder to come up with funny or insightful reactions to useless interactions, to open the door to new puzzle possibilities by coming up with new verbs, something most LucasArts and most Sierra designers, with their text parser roots, weren't affraid to do.

I see your point, but if you're changing the game to fit the GUI, you're doing it wrong. The GUI must be designed to fit the game, and not vice versa.

First off, I just want to say thanks for all the great feedback so far!  I love this discussion! 

Secondly, I do agree that the flaws here are mostly the game makers' fault for under-utilizing the potential of the interfaces rather than an inherrent flaw with the interface itself.  I tried to get this across in the article.  However, since the Sierra interface I was discussing uses mostly "Talk" and "Use" for the puzzles, I think it is really hard to design a game that uses these two verbs in enough ways as to justify them.  I ballparked the number 90% for games that don't fully utilize this interface, but really I was just trying to cover my back.  If anyone can name one game, classic Sierra, commercial, amateur, whatever, that fully utilizes this interface to it's potential and list a few puzzles from that game where you had to use both verbs on the same item, or use the verbs in clever ways, let me know.  Otherwise I'm still going to put some of the blame on the interface.

Thirdly, definitely the word "broken" is an exaggeration to be more eye catching or controversial or whatever.  The games are hardly broken if I can still play through them.  I'm just trying to point out some fundemental flaws that are propagated through numerous AGS games, and calling them broken is kind of my catch phrase.  Hope no one takes it to harshly.

I think those replies cover a number of the complaints people had.  I'll still do some direct responses here.

Za_Uvek:

Quote from: Za_Uvek on Fri 25/07/2008 14:49:55If applied correctly the Sierra interface can form some of the most clever, challenging and logical puzzles to be found. The verbcoin, bah, leaves no challenge and can be tedious most of the times.

See, but I still can't think of any examples of when the Sierra interface was applied correctly to form challenging and logical puzzles that couldn't otherwise be implemented in a less frustrating manner.  Furthermore, I'm not sure how you can say that the verbcoin leaves no challenge when it's fundamentally the same as Sierra, only changing the order of your command input.  Most Verbcoins that I've seen have "Look" "Use" "Talk" and "Inventory"  This is exactly the same as Sierra except that you select the verb after the item instead of the other way around.  What impact does this have on puzzles?

Note also that in the article, I'm not advocating for the verb-coin, just showing it as an alternative with it's own drawbacks.

MashPotato:

Quote
Keep the articles comin', Vince!  Reading this one made me think of the default-walk-thingy I just described, which is what I will probably use in my next game unless I can think of something better--these articles have been very timely, I must say Wink

Thanks Mash!  I like that "default-walk-thingy".  Anything that makes cycling the cursors less of a headache!  As it stands now, it's definitely unneccessary to have it as it's own cursor except in cases where there are hotspots all over the place and it becomes hard to click anywhere but on a hotspot.  I had this problem in one room of Resonance where there's a big car in the middle of the screen, and you have to use to choose strategic places to click to walk to the other side of it.  Still trying to figure out how to remedy this.

Stupot

QuoteI'm tempted to think there is something satisfying in trying lots of different combinations of actions; the more you are clicking on the mouse, the more you feel as though you are doing something rather than simply watching an interactive movie.

I see what you're saying here, but I have to believe that there are better ways to add complexity to a game than making them try multiple combinations.  With the Sierra system you get "Ooh, it's a person!  I know, I'll try to Talk to him!  Haha!  Solved that puzzle!"  And with more verbs (LA) you get "Ooh, it's an item!  Let's systematically try every verb on it!  Aha!  The developer meant for me to Push the thing!  Solved that puzzle!"  Neither of those satisfy me, but maybe some players enjoy those small victories. 

Edmundo and JBurger  I quite agreed with your posts!  Thanks!

ProgZMax  What happened to our unified voice?  ;)  I apologize for the blanket statements.  Part of it was me trying to ruffle some feathers in order to generate discussion.  Both of these system's faults usually lie in the under utilization by the designer.  Really, if AGS didn't exist to bring adventure design to the masses, I probably wouldn't have any problem with these interfaces.  However, I do think that both interfaces are really hard to design for because properly utilizing all those verbs takes care and planning.

I'd still like to hear more examples of games that do utilize the full potential of these interfaces.

LimpingFish and ProgZMax

QuoteWe know you can't talk to grass, but to use the talk verb on it with no feedback is just as annoying as being told you can't talk to the grass.

Ha!  I agree!  The grass probably shouldn't be a hotspot at all, actually, unless there's some puzzle in which you need to use it.  And ProgZ' point regarding this is interesting.  On one hand, I don't like not getting feedback for my actions, but on the other hand, having no response helps me filter useless "I can't do that" statements from useful clues.

One thing that I did in Nanobots was to remove mouseover indications for things that couldn't be interacted with using the active verb.  So if you were using Chembot's "Mix" ability, the mouseover text only appears when the cursor is over something with liquid.  Clicking anywhere else results in nothing.  Better, I thought, than having her say "I can't mix that, dummy!" all the time.  Similarly, you can try to push most things on the desk, including bots, but things that are nailed down, like the shelf connected to the back wall, don't produce a mouseover effect.  That shelf only produces a mouseover for Brainbot's "Analyze" action, because you can gain some useful knowledge from doing that.

The problem with this is that, for example with the grass, if I mouseover it with the talk cursor and get no mouseover indication, or try to click on it and get no response as ProgZ suggested, I assume it is not interactable.  So when a later puzzle involves using the scissors on the grass to get some grass clippings, I don't think of it, because I'm under the assumption that the grass is not interactable.

Makeout Patrol A good post.  Thanks.

Babar

QuoteThe designer of the game views that specific character as an obstacle in the game, and the solution to the puzzle is to push him off the cliff. I just got there. I don't know anything about a puzzle, so I attempt to talk to the character. He gets pushed off the cliff. Just an example. Hell, flip it over. I wanted to push the character off the cliff (a fairly simple and easy way to solve the puzzle), but I am only able to talk to him.

An excellent point!  And this is a situation where having the choices would make the game excellent.  Two ways to solve the puzzle, talk to the guy, or just go the easy, evil route.  Here, if the developer were so inclined, clicking on the character with the two-button system could make the character think to himself 'How do I want to go about this?' and then give some options (via a dialog choice or something) "Threaten him."  "Sweet talk him."  "Push him off the cliff." etc.  However, having these choices worked into the verb system of a Sierra/Lucas Arts interface would be much more elegant than having to choose the choices from a list.  Use the hand to push him, the talk to talk to him, or give him that cake from your inventory to try to sweet talk him.  That would be an example of a well-implemented Sierra interface.  Now just keep that up through the rest of the game...

Also, having a contextual verb coin would work here well.  A contextual verb coin doesn't have "Talk" "Look" "Use" on it, it has a different set of commands for each object you click on.  So, a potted plant might have "Look" "Break off a leaf"  "Push to the left"  "Push to the right" "Hide behind".  Not all of those would be needed in the game, but adding them can add that complexity and sense that you still have to figure out what you're going to do with it yourself.  And it's a choice that requires more thought than "Hmm... should I use 'Use' or 'Talk' on this plant... Duh."

blueskirt  I also see your point, but I have to agree with Makeout Patrol.  Perhaps it's just a semantics argument, but I think the interface should be designed around the gameplay, not the other way around.


Also, the necessity (or lack thereof) of the "look" verb is an interesting discussion.  It likely depends on the way you design your game.  I'm kind of wishing that I had left it out of Resonance because there's so much to write!  ;D

Agreed, I must have verbcoin mixed up with something else and it has no impact on the puzzles.
However, I will always see the Sierra interface as challenging and logical. Frustrating, no, really hard for some, yes!

QuoteWhat happened to our unified voice?

There are some things I just can't agree to! :)

QuoteThe problem with this is that, for example with the grass, if I mouseover it with the talk cursor and get no mouseover indication, or try to click on it and get no response as ProgZ suggested, I assume it is not interactable.  So when a later puzzle involves using the scissors on the grass to get some grass clippings, I don't think of it, because I'm under the assumption that the grass is not interactable.

You don't actually assume this because talking alone doesn't yield results, do you?  If it's clear you can use the grass in some way, I typically make the use/interact/examine command do 'something' to indicate there's value there.  If something has no value at all, I typically just give it no other action than maybe examine, since so many people seem to enjoy looking all over game screens at things.  This actually seems to be a very common 'need' in games of this nature, like 'find the item' games, which rely heavily on examining everything.  People seem to like to be able to look at (and get a response for) virtually all of the items in a room, and it's an easy way to extend gameplay without hurting anyone.  Basically, by simply 'cutting off' a redundant action on an item, I'm effectively telling people 'this action is useless on this thing' without any delay, and without some kind of annoying statement that either insults their intelligence (I can think of a few recent games which do this) or annoys them.

Quote from: Vince Twelve on Sat 26/07/2008 03:43:22
If anyone can name one game, classic Sierra, commercial, amateur, whatever, that fully utilizes this interface to it's potential and list a few puzzles from that game where you had to use both verbs on the same item, or use the verbs in clever ways, let me know. 

Me Go Store one through three.  ;D

Hey Vince!

I actually agree 100% with your arguments here--it actually felt like I was listening to myself talk, since the points you made in both this and in Part One were all things I've noticed myself while playing games.  I got a good laugh out of your Venn Diagrams--I've often thought that exact same thing about the Talk/Use functionality in games like Gabriel Knight, King's Quest, Quest for Glory etc.

One thing I would change before you post this on your blog: I wouldn't draw specific examples from the indy-game community.  Perhaps it's just that I myself tend to be a bit more sensitive to indy game-makers because they're...well, indy game-makers.  A lot of them (I suppose I should say "us" even though I have yet to produce a finished AGS game for the community) are your average work-in-an-office or going-to-college variety of person and might not be used to criticism of their admittedly amateur games.

So, I would either be vague about which games you mean (instead of "Ben There, Dan That", say "one indy game I recently played"), or use examples that are commercial rather than free-indy--such as the aforementioned Gabriel Knight, King's Quest etc.  There's plenty of commercial games to draw from, whose designers are used to negative response and know the risk they take when they produce the game, that you don't need to rely on free indy games as examples.

Just a thought, take it or leave it.

Vince, must say that I agree with what you actually do say(concerning part 2).

Verb coin is good but the delay time is.. terrible. But concerning game interfaces they must be created in order to fit the purposes of the damn games. Example(this is not a self promo): In the Hitchhiker's Guide to the Galaxy, many things are defined by the fact how long does it take you to do things. And opening your inventory to select items can be stupid and takes time as well. That's why I've implemented the quick slots. Also the game was initially designed to have a sierra style gameplay. But since this would make the game frustrating and very very difficult I decided to go with a very simple verb coin system based on the Reactor's 09 improved verb coin system, which I call Reactor Flipped. Since left click opens the GUI. However based on what you say about verb coins disadvantage that sometimes confuses the player, talk interaction is only available when you talk to characters. Also the game DOES use sierra interface(Dark rooms), but then its very important and the game depends on the interface being like that. There's no bloody pops up but one which is in the top left corner and can only be active if you really want. Not accidentally.

"The Legend of Kyrandia" had what I would consider a very helpful interface. If my memory serves me correctly, clicking on something would automatically interact with it. If you clicked on something you could pick up, like a gem, the gem would become your cursor. You then had the choice of leaving the gem by clicking on the ground, or placing it in a slot in your inventory. The inventory itself was always "open" by being docked at the bottom of the screen. You could even sort things according to your whim.

(http://i8.photobucket.com/albums/a18/The_Ivy/Kyrandia1ingame.jpg)
Spot the big freaking red gem.

As far as I recall, clicking on something you couldn't pick up would yield something like a look interaction, and the game would basically tell you whether or not the item was useful. Clicking on a person would lead to a conversation, although I think these were mostly done as cutscenes rather than interactive dialog trees. Not a bad way of doing things, in my opinion. :)

I mostly agree about Kyrandia, since it's one of my favorite adventure games (and certainly one of the most polished-looking Amiga adventure games made).  Since so much of the gameplay revolved around massive inventory management, the pick up and manually place items approach probably made the most sense to them, since by the time you reach the caves you're dropping items everywhere to pick up new stuff to come back and get the items later!

  I would've liked it if the right click auto-tossed the inventory item in your hand back into the inventory, though; that would've made things a bit quicker if you picked up the wrong bauble from the inventory or just wanted to quickly toss in something you got from the scene :).  All in all, I did like the interface, though the gameplay was fairly uncomplicated and wouldn't have really benefitted from a set of verbs.  It was far more inventory-related than anything else.

I should say how glad I am that you´re doing these posts, Vince. I actually have an HCI degree (though nowadays I´m more involved in user studies than in usability testing), and I think interface design is an underappreciated aspect of game making.

The first article is pretty uncontroversial, basic stuff, and I pretty much agree. In the second, though, I think you make a number of problematic statements.

First of all, the "Sierra Operating System": One of the important things about making a UI is to tailor it to your application. An adventure game is not an operating system, and the design that is right for Windows is not right for King´s Quest, or vice versa. Your little example is therefore only confusing and misleading.

One of the ways in which games are different from OS-es (and most other apps) is that many of the traditional concerns of usability (efficiency, no unnecessary complexity, obviousness, learnability) either do not apply, or only do so with modifications. Most tasks in most computer games are made much more difficult than they need to be; usually the computer could do most of the work for you, but where´s the fun in that? (The usual joke is that the optimal usability solution for any computer game is a button marked "WIN".)

Back in the 1980s, Eric Solomon wrote a book on game design where he described adventure games as games where the object of the game was to figure out the interface. Obviously, he was describing text parser games, where discovering the set of useful verbs is indeed a big part of the gameplay. If the UI was obscure, this was at least in part deliberate. The trend in graphic adventures since then has been towards simplicity of UI, probably for the better, but we still need to maintain the balance inherent in the fact that adventure games still are about trying to figure out what you can do.

In analysing a game UI, we need to take into consideration not just all the usual usability aspects (when they apply), but also how well it supports play. That is, ways in which it makes the game more fun. (For example, a "highlight all hotspots" feature, found in some adventure games including Simon the Sorcerer, is helpful from a usability POV, but arguably reduces the fun-factor of playing.)

OK, on to the specifics of your complaints about the iconic Sierra/LEC UIs. Mostly, you seem to have a problem with the number of distinct actions offered. The debate of one-action vs. multi-verb UIs is much broader than the particular pros and cons of these two interfaces, and has been going on since parser games were first replaced by point-and-click. I am not convinced that that you are right when you argue that in most games you only need to be able to do one action to each object in the world, and I think your Venn diagrams are inaccurate representations of at least most LucasArts games. One example I´ve seen used before is the "Push Sophia" puzzle at the Cairo bazaar in FOA. And even if many objects only have one meaningful associated action, it is not always obvious what that action is (remember the "Idol of Many Hands" drowning puzzle in Monkey Island). Having more verbs available does genuinely increase the number of possible actions, and thereby possible puzzle solutions. (Though obviously doubling the number of verbs does not double the number of possible meaningful actions. The effect is not linear.)

Even if this was not so, I would argue that, when used well, it increases the immersion with the game world, and thereby makes the game more fun. Like you quoted in that post by Lee in Limbo on your blog, a fun adventure game should be about more than just solving puzzles. The ability to "do stuff" that aren´t (just) moves in a puzzle solution is essential to making any non-abstract game fun. In an adventure, this starts off with simple things like being able to walk around, look at things, and try actions just because you want to. Having more verbs creates more opportunities for non-essential stuff to try. That´s what Sierra tended to do with all those "useless" cursor-object combinations.

Now, I agree that many games, especially those that are not by LucasArts or Sierra, in fact only use one action verb per object in almost all cases. And since neither interface is in fact perfect, the other benefits of all those verbs may not be enough to justify them. Often, you can compensate for removing a verb by adding an inventory item or a hotspot. For example, in FOA you could "Use cattle prod on Sophia" or click on "Sophia´s back".

So yes, a game designer should definitely take a look at his or her game, and see if multiple verbs are necessary to the puzzles, will help immersion, or is otherwise a good thing for the game. If not, a right-button: look, left-button: walk/action UI is probably best. If more verbs ARE needed, however, I think that the LucasArts and Sierra UIs are both fairly decent solutions to a pretty tricky problem, and they have the VERY SIGNIFICANT benefit of being standardized and well-known by most players. Most amateurs are not likely to come up with a better design on their own, so you´ll end up with a worse interface that people will have to learn just to play your game.

What´s so great about the Sierra and (especially) the LucasArts UIs? To be continued...

(Sorry for the double post, but the two are logically separate.)

Let me sing the praises of the standard LucasArts UI (I always called it the Scumm Bar, but that seems to be just me), since that's the one I'm more fond of.

OK, so the drawbacks are obvious. It takes up around a third of the screen. It requires moving the mouse back and forth to one corner of the screen any time you want to do something (or learning some not-always-obvious keyboard shortcuts). How often do you ever use Push or Pull? But let's look at the positives!

First of all, it's incredibly intuitive to use. All of the verbs are right there on the screen (well, they took away "walk to" in the later versions, but by that time people were familiar with clicking on any non-hotspot to walk there), and it's obvious that they represent the things you can do. Compare that to, for instance, verb-coins, which are IMPOSSIBLE to use unless you know the secret (the game will simply appear to not respond to anything you try). Try giving a novice gamer an adventure with a LucasArts/style UI versus a verb-coin UI and see which one they find easier.

Secondly, the sentence bar is pure genius. If you're just mousing over things, it shows you the hotspot name, but if you click on a verb, it starts to build a sentence: "Look at ... important-looking-pirates", "Use... rubber chicken with a a pulley in the middle with ... cable" This is a great "fun feature" that encourages exploration and play (and that "Eureka" feeling when you suddenly put the right ideas together), but it also serves as a transition for players coming from a text parser background. ("See? It's just as if you wrote the command yourself")

Thirdly, the inventory is on-screen along with the main game world. This makes the inventory-based gameplay that is the hallmark of LucasArts titles much smoother than it would be with, say, a Sierra interface.

Fourthly, LucasArts put in a bunch of little touches that makes it much more user-friendly. First of all, you have the default action on right-click. Then you have the "cancel-action-while-walking-towards-hotspot-by-clicking-on-something-else" feature (which separates a good LEC-UI clone from a poor one). Then you have the keyboard shortcuts, and the way they still light up the buttons and start building a sentence in the sentence bar. I think that for an experienced player using shortcut keys, the LEC interface is one of the quickest and most efficient there is.

Finally, you have the nice way the interface is darkened out during the cut-scenes, creating a widescreen effect, and the way the dialog options fit in the same space during conversations.

It should also be recognized that LucasArts invented (AFAIK) the idea of hotspot indications on mouseover. And I can personally vouch for the fact that it's remarkably playable even without a mouse, using the keyboard arrows to control the cursor.

So are all of these excellent properties just luck? Was the UI just put together at random by a couple of programmers without any usability considerations and without any thought to design, like Edmundo argues? If you believe that you're on crack. The LucasArts UI, particularly in the form first seen in Monkey Island, was a remarkably elegant design, and a great step forward compared to what had existed prior. It's a little clunky nowadays, and later innovations (like double-click to skip to the next room) improve on it, but it's still a pretty good choice for an adventure game designer, especially one making a game in the same spirit.

Snarky, that was an excellent and well argued post!  Thank you!  As I mentioned at the start of the first article, I'm no HCI expert, only having worked along side several for nine months.  I really appreciate someone with the proper training coming in here to set me straight on some points.

With the operating system, I was trying to demonstrate how annoying I feel the interfaces are by placing them out of the context of games.  I wanted to do this because of the nostalgia-tinted glasses that I mentioned.  However, you are exactly correct.  The usability concerns that apply to a computer, which aim to simplify tasks, do not apply to games, where tasks should be more complex.

QuoteThe usual joke is that the optimal usability solution for any computer game is a button marked "WIN".

So, so true!  I'm considering adding one of those to Resonance!

As I said, I do think that the annoyance that I feel when I play most games with these interfaces takes a back seat to the second issue about misuse, where the OS example still applies.  I've mentioned up thread, but need to make it clearer in the article, that this issue is primarily due to the developer's misuse, which apart from many of the classic Lucas/Sierra games and a handful of amateur games is extremely widespread.  The annoyance issue is subjective and the OS example doesn't apply very well for the reasons you mentioned.  But I maintain that if your game doesn't have some instances where the usage of "Talk" and "Use" are some how shuffled up, diverting from the norm, then the verbs are redundant and should be combined.  This goes for the Lucas interface as well.

I think the OS comparison does apply, if only loosely, here.  The example where I combine all the fake Resonance documents/video/music, into one master file and then use the verbs on that instead of having each of the files separate demonstrates a way to use the interface's potential in a way that your commands have more meaning without being redundant.  Doing otherwise isn't complicating the interface in the interest of creating gameplay.  To me, that's squandering the potential of your interface.

I'm mainly saying that if you're not going to take the special attributes of your chosen interface into consideration when designing your game and your game comes out resembling one of my worst-case Venn diagrams, you would be better off with a less obtrusive GUI which wouldn't sacrifice gameplay at all.  Ok, maybe it would sacrifice some illusion of gameplay, but fake gameplay made up of unnecessary clicking isn't what attracts me to the genre.

And, yeah, the article is made up of a lot of exaggerations and too-strong statements (none of these games are actually "broken") to try and get the point across and spur some discussion.  I know there are some games who use these interfaces to great effect, especially the old classics around which the interfaces were designed.  I mainly want to make sure amateur devs like me are keeping in mind the full potential of their chosen interface when they design their game.

I'm definitely going to use some of these points in the rewrite before posting the final article to the blog.  Thanks a bunch!  Looking forward to hearing what's so great about the interfaces!  (D'oh!  You posted it while I was writing this.  Reading now, post back in a bit)  Edit: Actually, I'll post back tomorrow.  It's way late.  But it's another great post! Thanks!

QuoteI see your point, but if you're changing the game to fit the UI, you're doing it wrong. The UI must be designed to fit the game, and not vice versa.

My bad, I had trouble finding the right words for that sentence, I thought these would works but obviously it came out all wrong. What I meant to say was: we should encourage people to take the high road rather than the low one, to come up with insightful/funny messages that add to the experience, rather than avoid the problem entirely with a 2 buttons interface. While a 2 buttons interface at full potential is better than any UI that isn't used to its full potential, a LucasArts/Sierra UI at full potential is better than a 2 buttons interface at full potential IMO.

Vince: I was kind of tired of all those game articles that focused exclusively on storyline, narration, story structure and characters development. It's good to see articles that kick the adventure gameplay design in a direction which test our positions and open our eyes to the flaw of today's games. May we be 2 buttons UI proponents or multiple verbs UI proponents, it is quite true that a lot of indie adventure games don't bother to use the Sierra/LucasArts GUIs to their full potential. I hope you will keep kicking the adventure game gameplay design in new directions, maybe touching the subject of puzzles, death or the importance of breaking new walls in term of gameplay.

QuoteThe trend in graphic adventures since then has been towards simplicity of UI, probably for the better, but we still need to maintain the balance inherent in the fact that adventure games still are about trying to figure out what you can do.

I completely agree with this, which is why I'm not particularly fond of over-simplified interfaces.

Thanks for your reply, Vince. Although I'm fond of the LucasArts UI, and think that it was a great solution for its time and a respectable one for amateur games today, I definitely don't think it's above criticism. You're doing a great job of that, and I'm glad the discussion is spilling over into larger issues. As for the Sierra interface, I'm not a huge fan of the right-click-to-cycle-through-cursor-modes principle, but it does get the job done quite efficiently (as long as the number of modes is limited; the Gabriel Knight UI was almost a joke). The standard Sierra inventory system is horrendous, however.

I should point out that although it means I'm aware of some of the issues of interface design, an HCI degree doesn't make me an expert. I've never examined these UIs the way a professional would, and as I mentioned, I'm working in a pretty different area of HCI. So I'm just commenting as an interested amateur with an opinion.

To return to the point I was making about inefficient interfaces, I'm just playing Apollo Justice on the DS. It has all sorts of mini games where you have to perform simple tasks (dust for fingerprints, make molds of footprints, use an X-ray machine) in incredibly inefficient ways. The whole point is that it's fun to use the touch screen and see the effects on the screen. Too few adventure games take any joy in the interactions themselves. Some text games play with language, and you have stuff like the magic cursor in Discworld and the "story told through sentence bar" in the fight in the mansion in Monkey Island.

But recently, outside of the DS and other Nintendo platforms, I can pretty much only think of Fahrenheit and a certain Linus Bruckman. That's a shame. I think the Wii has proven that fun interactions are the keystone of fun gameplay. Maybe we're all too jaded about PCs to have much fun with simply a mouse or keyboard, but I hope not.

Edit: I got a bit off topic there and managed to distract myself from some of the points I was going to make.

1) I think the two-button UI is pretty good (it elegantly supports four actions--"walk to", "look at", "action", and "use X on"--with only two buttons, without complicated click patterns, verb-coin gestures or sub-menus) for what it does, but one problem it has is lack of affordance. There is no visual indication of how it works. You could watch someone play it and have no idea how they're making it happen. You pretty much have to read the manual, or figure it out by trial and failure. This is one area where the LucasArts and (to a lesser extent) Sierra UIs have the upper hand. I seem to remember that there was a Star Wars fan game that tried to use little legends on a crosshair cursor to indicate the leftclick/righclick actions, but the meaning of the labels was pretty non-obvious, at least to me.

2) I wanted to single out Reactor 09 as an AGS game that created an extremely slick, original UI (a version of context-sensitive pop-up verb-menus, as I recall it). However, even in that game I think there were one or two occasions (near the end) where the UI screwed with the gameplay. That's the risk you take when you try to innovate.

There are many interesting point mentioned which I was thinking about a lot in the past few days.
First I thought "He is right, actually only look/interact is needed."
But then I started to think about this whole stuff and some issues appeared to me.

1) Accessability
First of all not all people have a right mouse button. A friend once played Monkey Island on his mobile phone in scummVM just with his pen input device -> no right click.
Also on some notebooks with touchpads doing a rightclick is a pain in the ***.

2) Inventory:
For example the BASS inventory gui was driving me crazy. It just didn't work for me. Additionally it limits the actions that could be performed. A torch for example could be switched on but it could also be handed to another person.

3) Limited possiblities:
Usually there are more ways to interact with an item. You could, for example sit down on a chair, but you could also push it somewhere and stand on it to reach something or you can use it to block a door.
It has happened in adventure games that I had a certain idea how I could solve a puzzle and when I tried to interact with the needed object something completely different happened. It's some sort of "well, I managed to solve the puzzle, but I wanted to do it totally differently" which I think is very bad for immersion.

QuoteAlso, having a contextual verb coin would work here well.  A contextual verb coin doesn't have "Talk" "Look" "Use" on it, it has a different set of commands for each object you click on.  So, a potted plant might have "Look" "Break off a leaf"  "Push to the left"  "Push to the right" "Hide behind".  Not all of those would be needed in the game, but adding them can add that complexity and sense that you still have to figure out what you're going to do with it yourself.  And it's a choice that requires more thought than "Hmm... should I use 'Use' or 'Talk' on this plant... Duh."

As far as i remember, in Bernhard's Room you have to talk to a plant ;)

However, the contextual verb coin could be a good idea. I usually don't like verbcoins because they increase time needed to do a certain action, but in this case this could work out very well. However, I think this is very hard to implement properly. If you could (let's take the chair example) move, sitdown and pickup the chair, every chosen option should do something. If you have the option "sit down" and then you get the response "I don't want to sit down" this is quite useless. The more options there are, the more implementation (animations, coding,...) is needed which will result in huge workload.

So actually I don't know what the best solution is, every GUI has its advantages and disadvantages...

Gwargle, I'm always on my way to bed... no time to post today.  Just wanted to say that I also get really annoyed at verb-coins that you have to hold and wait for.  Usually when a game uses one of these, I can't figure out how to play the game, clicking around with no success until I finally read the manual and realize that I need to hold the button down for a while.

Definitely left or right click to bring up a verb coin seems like the way to go to me.  I've never made a game with a verb coin (En Passant never got off the ground...) but I definitely don't think making the player click and hold for a second before each action is the best way to go.  I'd have to experiment with it.

Hi to all,
great topic start and very good opinions.
On my recent project (Blue Moon) I decided to make my own interface and came with something not original but functional, I hope!
Rui Trovatore Pires gave me some lessons on interface design and as I was improving the Blue Moon Demo he kept directing me on the best way.
It consists on a constant GUI placed on centre bottom of the screen that shows the name of whatever the mouse is over,
then,  If you Right-Click it pops a verb GUI, with some level of transparency, on top of the object that consists on some icons :     
                                      - examine
                                      - use
                                      - inventory (allowing you to use an item with the current object)
if a person is the object then a TALK icon will be present;
Right-clicking or moving EGO will turn the verb GUI off.
I've seen it on other games and think it is much more functional than the standart interfaces for the reasons everybody named before.
What do you think of it?

Quote from: ProgZmax on Sun 27/07/2008 22:18:57

QuoteThe trend in graphic adventures since then has been towards simplicity of UI, probably for the better, but we still need to maintain the balance inherent in the fact that adventure games still are about trying to figure out what you can do.

I completely agree with this, which is why I'm not particularly fond of over-simplified interfaces.

I HATE over simplified interfaces,

hate

hate

hate

hate

An interface such as in the new sam & max games is horrible, there if anywere is a perfect example of a broken game.

Yeah, I was playing one of the Sam & Max episodes recently, and it's very weird...it seems to be just a 1 button system? If there is nothing there, you walk to, if you tap twice, you run, if there is something there not interactable you get a comment, if it is interactable you either do something, or pick it up. Right-click only seemed to be there to skip text.

Also...I am just wondering, is it a problem that some GUIs may cover part of the screen? Sure, I got a little peeved at the end of one of the Monkey Islands when they started showing the backgrounds without the GUI, and I realised that the GUI actually hid some of the artwork, but if it is designed properly, would it be a problem? For example, if your game utilised the the entire screen of your...1024x768 resolution. But then, you magically got a 1024x808 resolution, where the extra pixels would be used to show a GUI. Or is it a matter of the GUI being 'intrusive' on the gameworld experience?

Another thing...since inventory being easily accessible is important, and things not covering the screen is also important, would the the pop-up inventory system (like in BASS) be one of the best current ones? Someone here mentioned they didn't like it. One of the Trilby games used an inventory GUI where when you clicked on something with one of the buttons, the inventory popped up, and you could use that. I wasn't too fond of that, because it covered what you were doing, seemed to cause extra clicks and inventory handling (like looking at it, mixing it, using it, etc) didn't seem to be handled well.

Personally, I like the Broken sword UI. Left click: Walk/Interact, Right Click: Examine. Inventory at the top of the screen for easy acsess, with no pop-up windows or verb coins to break immersion. Over-simplified? Perhaps, but I think the biggest puzzle in an adventure game shouldn't be the interface.

For me the best GUI has been that of CMI. It was essentially a extension of my spinal cord. Not because of the verb coin, it didn't pop up even once when I last played it, but the way I played. Having fingers on T and U I essentially had four mouse with a different cursor mouse on each button. I too hate oversimplified interfaces. And I wouldn't want to play the game with just a Win button (well, MI had the win button).

Well, this is certainly a most interesting post.

What I learn from this is:

- design your gui to match your game complexity:

*) if you're only going to script one or two actions per hotspot, don't bother creating 8 buttons with lots of different options

*) if you've got a very complex game where almost any type of interaction actually does something (even if it is just to give a funny comment that makes the player laugh out loud), you might consider having a few more buttons

- Inventory visibility should match how often it's used.

put your inventory on screen if you're using it most of the time, hide it (but make it accessible enough) if you're only going to use it relatively sparse

- allow for multiple types interaction with a gui control

don't be the guy to create only one type of interface and then FORCE the player to use this and only this way to do stuff. Allow the player to choose from a (limited) variety of types:

*) the simplest being able to switch the left- and right- mouse button functions
*) the ability to use keyboard shortcuts
*) the ability to play the game with ONLY your left mouse button (I mean: laptops, PDA's, ... not everyone HAS 2 mouse buttons and a mouse wheel, you know)

- Use the LOOK function properly if you've got it!

Don't just state the obvious, give the player EXTRA information, as that's what it's for.

------------------------------------------------------------

I, for one, found this article and resulting comments very refreshing. It inspires me to improve my own game even more and check and double check so that there are no "you can't do that" or other blatant annoyances in it.

Though I'm more of a "let's give the player all kinds of options and make sure he WANTS to try out ALL of them" kind of guy, I'll probably give gui design some more thought next time around.

If I find the time I'll most certainly rework my current verbcoin module and game template to incorporate these lessons in the extend it's possible. (after all, it's still the choice of the game creator to use or not use any and all functions)

(And if I find a LOAD of free time I'll work on creating other templates which also include these lessons but for different gui types)

Electroshokker, you're right on everything you mentioned except that I think it's actually very safe to use two mouse-buttons in a game. You name laptops and PDA's as examples for environments with only one button but both those devices have two mouse-buttons as well and have had them since decades. So, allowing the player to play your game with one button only and not using them both would actually worsen your design because of over-thinking and -reaction.

Quote from: dkh on Sat 02/08/2008 15:39:31
Electroshokker, you're right on everything you mentioned except that I think it's actually very safe to use two mouse-buttons in a game. You name laptops and PDA's as examples for environments with only one button but both those devices have two mouse-buttons as well and have had them since decades. So, allowing the player to play your game with one button only and not using them both would actually worsen your design because of over-thinking and -reaction.

ah, but my main point is to allow the player to choose the way he interacts with the game.

In those rare cases a player finds himself with only 1 button (like on Mac OS X), the player should be allow to choose the "one button"-mode rather then the default two-button mode.

By allowing customization of the way the interface responds (rather then changing the interface alltogether), the player will feel far less restricted and will always choose the way he/she thinks works best, thus giving a better gaming experience.


Consider this: most modern games all have a default profile defined for keyboard and mouse interaction, but they also allow the player to customize these.

My idea is to do the same for adventure games, but instead of full customization (except maybe for keyboard shortcuts), the player would be presented with a select amount of choices at the start of the game.

(Preferably the game remembers these settings for next time round, it would be very annoying to have to select the default/left/right mouse/keyboard/... mode every time.)

Example, the verbcoin system:

- default mode (left click pops up the verbcoin, release over the action of choice, right click deselects)
- switch right-left mode (right click pops up the verbcoin, release over the action of choice, left click deselects)
- loose click mode (left click on something pops up the verbcoin, leaves it there, click again to choose your option, right click deselects)
-...

Important is not to allow too many different modes, because that too would confuse the player.

In my case, I'd let the player make only two choices to keep things simple:

mouse or mouse+keyboard
2-button mouse or 1-button mouse

(note: the switch left-to-right mode would be invaluable to left-handed people, I'd think. We musn't forget these. Though I'm sure the OS already allows for these to be switched, which is why I wouldn't include it)

Even on Mac OS you can right-click. (http://www.macinstruct.com/node/66)

I absolutely agree with everything you say, but I just think designing your UI with people who have only one mouse-button in mind is a little over the top.

I think a very configurable interface allows people who are disabled to access it: the more configurable, the easier it is for people to work around disabilities. For example, some people have no arm movement but can use mouse pointers via head and eye trackers. For them, they "click" by hovering in the same place, and some of the interfaces describes would make the use of these aids impossible or hard.

bicilotti also found a problem with his colour-based problems in Colorwise recently and worked around it. Purely sound-based puzzles are also exluding some players.

It's certainly a very honorable idea to give every person with every possible preference or disability a chance to enjoy the game equally, but it's just not realistic. At least not for all games or projects.

Also, if you had created a freely configurable interface that allows every armless, blind or deaf person to play the game, shouldn't you also be extra careful not to have a love-interest in game because homosexuals might be offended or, for that matter, exclude any kind of joke that could offend for example a person with a certain disability?

Once your train of thought starts heading into this direction, it ain't coming back so quick, it leads ad absurdum in a split-second.

Quote from: dkh on Sat 02/08/2008 17:09:46
It's certainly a very honorable idea to give every person with every possible preference or disability a chance to enjoy the game equally, but it's just not realistic. At least not for all games or projects.

Also, if you had created a freely configurable interface that allows every armless, blind or deaf person to play the game, shouldn't you also be extra careful not to have a love-interest in game because homosexuals might be offended or, for that matter, exclude any kind of joke that could offend for example a person with a certain disability?

Once your train of thought starts heading into this direction, it ain't coming back so quick, it leads ad absurdum in a split-second.

I think the point of SSH was that with a little effort on the developers part we can let people with disabilities play our games. It's not a crusade, just two or three ideas that can make the difference for some of us.

SSH had an interesting post on the subject in his blog (post that right now I am unable to find, gahhhh!)

I was merely saying that the more people you can enable, the better. I find it offensive to try and turn it into a "It's political correctness gone maaaaad" which in this country usually comes from the Daily Mail-reading reactionary bigot.

Especially with free games, I can't really be bothered to implement a 'works for all, but is rather boring' interface because I'm receiving no compensation for it, and no matter how much effort you put forth people still manage to find something to complain about.  I think games work best when they have a specific audience in mind rather than generic attempts to reach out to everyone, anyway.

In my work in progress, Galactic Battlefare, I implemented a four action verbcoin that interact with every object. The actions changes if you point a character, a object, diferent objects... Every object and hotspot with customized actions.

There four "generic" actions that changes from one to another object...:

-Interact (primal object use)
-Look
-Talk (Talk to, Eat, Drink, and Talk about (like look, but a "personal" description from the character about what knows or think 'bout an object)
-Pick Up, or Hijack, or Figth/kick/break

In the developer video blog, you can look to the system, ¡bout the minute 5 - 5:30
http://www.youtube.com/watch?v=pvrJMTBsXec
(don't worry about the devvlog, is a BIT freak xD)

I was just thinking about this, and ultimately one of the parts in this series should be Webhosting: RapidShare and MediaFire are THE DEVIL!

Seriously I'm sick and tired of people using services like these for their AGS projects when there are plenty of hosts without restrictions on how much non-premiums can download and whatnot. Say for example there are 20 full-length AGS games I wanted to download. Imagine they're all uploaded to RapidShare. Now I'll have to wait 20 days to be able to download them all. If the author had simply found a decent service, I could get them all within a couple of hours max.