Parser vs. Point & Click (not just KQ)

[Armakuni]

I love both the parser interface and point and click interface.

But as much as I enjoy the parser interface in older games, I really can't see how that would work in modern games. It would seem a bit absurd... having modern graphics, music, all of that... and still having to type in commands.
So I don't mind how that was eventually replaced.

Point & click is in my opinion ideal for an adventure game... that is, the kind of adventure game I love the most, which are games in the style of Lucasarts and Sierra classics.
For a while I suspected it might just be nostalgia, but after having played more modern adventure games, often using direct control and the like... I think I've given it enough time by now to say that for me, it's not only because of nostalgia.

I just enjoy the kind of adventure game you get when you have point & click controls... even though it means you can't have equally cinematic scenes (not without going through more hassle of figuring out how exactly to implement it anyway), but that doesn't bother me one bit.
I've never needed cameras moving around and lots of various 'artistic' camera angles, etc... I can honestly say that I can get just as good of a feeling of nice atmosphere in a game without that.

So for me, in the end, I prefer point & click as that is something even new games can use without appearing really strange.

The reason I like it much better than direct contol... well, there are more than one, but a biggie for me is that constantly manually controlling your character everywhere seems a lot less convenient to me than simple point & click controls - as my ideal adventure games are challenging ones where there's a good chance you'll end up stuck more than once. In situations like that, where you end up walking all over the game world, possibly for quite some time, I think it's a hassle to have to do that using direct controls compared to point & click.

So in that way, it's less of an issue for me in dead easy games like the ones Telltale usually develop... when I never get stuck (for more than a few minutes, anyway)... that greatly reduces the amount of just walking around I have to do.

On a related note, I've always been really happy with the kind of controls Sierra used in their older adventure games, where you had to use the keyboard - I love how their system works in the way where you just tap a key to start walking and then tap it again to stop.
People might argue this means you don't have the same precisioun when controlling your character in these games, and that's true, but other than in a few cases where Sierra decided to make navigating certain areas a challenge on purpose... I never had any trouble controlling any of those games.
And if you did have trouble, you could always use a joystick... well, if you had one... but at least they did give you somewhat of a choice, I guess.
Really nice how they did that, I'm certain I would not replay them as much as I do if the game had direct controls the likes of what you normally get these days.

EDIT - forgot to mention - the most ideal control scheme in adventure games for me is point & click, with hotkeys for the various verbs/actions. I love how this gives you the choice of using the mouse only, and even without using hotkeys, I find it quite convenient... but I love having the option of using hotkeys though I don't always use them.

[Kisai]

Originally Posted by Anakin Skywalker

Is there anyone else here who prefers the parser system to point in click? I don't just mean for KQ, but for adventure games in general.

Wall of text incoming...

In my opinion, and I'm saying this because Kings Quest III was "new" when I first got a PC, it was a natural progression how games have gone so far.

If you roll back the clock. Early games (Japanese games maintained these text parser interfaces longer) went from complete text, where you had to use your imagination to crude visuals where you could approximate what you wanted to do. If the programmers anticipated the same thing, you'd progress in the game, if they didn't, you'd get a remark about trying something else.

The point and click interface was a natural progression and even in foreign games, it was pretty much the advent of the mouse that made the games easier, from a decision tree (like in the lucasarts games) instead of "try X with Y" guessing.

So where the Sierra games had some design issues with P&C, they more than made up with the storyline or jokes. Space Quest actually made it fun to try and see how many different ways to die. But many of the people I've talked to who played both Lucasarts and Sierra games thought the Sierra games relied a little too much on "click everywhere, try everything."

But if I had to redesign the text parser interface for new devices like the iPhone/iPad, I'd probably go with Leisure Suit Larry 7's version where you could add a verb to appropriate... or inappropriate... situations to unlock easter eggs, or see things that you otherwise wouldn't think of. It requires a little more forethought, particularly with slang.

But the Text parser style that the AGI engine games had was rather unintelligent and required fast typing skills, rendering some game scenes difficult to time. In KQIII you also had the magic book which difficult to read (I'll admit, I played this game when I was like 8 and the font it in the manual was too difficult to read and took WAY TOO LONG to type in.)


The interface design we have to work against now is the touch, and natural interface (Kinect), and motion controls (wiimote, acceleration, gyroscopes, etc) So if you want to design a game now, you have to think about how to make it work with all of these input types.

For mouse/keyboard and wiimote, you always have a cursor, so you can do the traditional Sierra P&C or Lucasarts verb tree, and have more opportunity to add jokes and comments to doing the wrong thing. The strongbad games worked particularly well this way. This is not however an appropriate control scheme for touch and natural interfaces.

For PS3/Xbox360 and similar joysticks, you have to assume the player is going to move the character with the joystick and "do action" at various locations, so you don't really want the P&C control scheme (look at KQ5 NES to see why.)

For touch, particularly multi-touch, you probably want to actually make a variant on the P&C control scheme where the multitouch gesture is the "verb" but fallback to a menu of verbs if the player is having trouble. So you could either use one-finger to select and the second finger to "verb" or you could do something like "pinch to pickup", "give/use to/on X by dragging" and "look by spreading fingers apart", conversation trees can use the other traditional P&C methods by clicking on the object during the conversation or "bioware-like" where the conversation always has a "good" or "snark" option but ultimately progresses if the right flags were set anyway.

IMO, the iPad is probably very promising for the revival of P&C types of adventure games, as it's one of the few games that don't require a "joystick with many buttons" to play. The iPad could do the "text parser" interface as well, but instead of typing in one letter at a time, use the predictive text system. It might be interesting, but I wouldn't want to design a game completely dependent on the text parser. IN FACT, what would be an interesting way to revive the "text parser" would be by leveraging speech recognition instead that instead falls back to text parsing. So someone could say "brick and window" instead of typing it.

Again, the natural progression of game interfaces. Speech recognition has been around since at least 1996 or so (It was available in Windows 95), but never really been used due to patents, inaccuracy and slow processing. It wouldn't be unreasonable to have a configuration panel when the game first loads with:
Select a play style
[ ] *ABC microphone icon* - The game will listen for commands in addition to...
[ ] *ABC keyboard icon* - The game assumes a keyboard is plugged in and will accept keyboard commands (PC, PS3, Xbox360 or Wii can all do this) in addition to...
(O) *one-finger icon/mouse icon* - The game will assume a pointing device like a mouse or wiimote with one button is being used (can be used with joysticks too)
(.) *two-finger icon* - The game will accept multitouch commands
(.) *Joystick icon* - The game assumes the character will be moved with a joystick with at least one button (and present standard configuration of buttons Xbox360, PS3, Wii Classic, Wiimote+nunchuck, Wiimote, etc, depending on the platform)

Not all options have to be available, just those that the platform being played on can support, but the game can tune the instructions for puzzles more appropriately if the player is at least asked which way they'd prefer to play. More jokes can be hidden in the text-parser keyboard mode without having to neuter the gameplay for P&C mode then.

So an iOS device may only have the multitouch, microphone and keyboard option, where as the Xbox360 might have the joystick, microphone (headset/kinect icon), and keyboard (keyboard accessory) icon. The Wii would default to the pointer option but could be switched to joystick. As the Wii and PS3 do not come stock with microphone accessories it would be safer to assume it's not available (though there are some games that use one) but there is the possibility of using USB/Bluetooth accessories. All options are available to the PC/Mac, however multitouch can't be used unless the input device has it.

But I would like to see some creative return to the text parser. Kings Quest might not be the right game series to do that with (Space Quest or Leisure Suit Larry would.) The Quest for Glory Series had a lot of ways to play the game that were easily explored with the text parser and not quite as easily done with the P&C interface since if you imported your character from at least QFG2 or 3, you could be a Paladin with all the magic options, solving them either way. The QFG games had inherent replay-ability just from this alone.

[Olaus Petrus]

Originally Posted by Kisai

The interface design we have to work against now is the touch, and natural interface (Kinect), and motion controls (wiimote, acceleration, gyroscopes, etc) So if you want to design a game now, you have to think about how to make it work with all of these input types.

The thing is that I'm not certain it's possible to design games which work well with all sorts of different controllers. I have played many PC versions of games which were first designed to consoles and often these games have abysmal control schemes, because what works well with consoles works very badly with keyboard and mouse. Only console I own is NES, but I have played many games with my friend's Wii and I just can't figure out any reasonable way these games could work with PC. In similar manner many PC games have control schemes which can't really be converted to work with Wii remote.

[Kisai]

Originally Posted by Olaus Petrus

The thing is that I'm not certain it's possible to design games which work well with all sorts of different controllers. I have played many PC versions of games which were first designed to consoles and often these games have abysmal control schemes, because what works well with consoles works very badly with keyboard and mouse. Only console I own is NES, but I have played many games with my friend's Wii and I just can't figure out any reasonable way these games could work with PC. In similar manner many PC games have control schemes which can't really be converted to work with Wii remote.

For the Wii, the strongbad game is much closer to KQ7's interface (one cursor for everything except when trying an inventory item) and it works fine on the Wii. But you have to take note that the Wiimote is being used as a mouse and makes for a terrible text-input device.

I think alternative interfaces can be done, but they have to be designed in from the beginning, not an after-thought during porting. This way they're alternate ways to solve a puzzle, not "it's too easy to do with a mouse."

Personally I think touch interfaces are not "pointer" interfaces, and can't be used the same way as a mouse. The P&C style sierra games can technically work unchanged on a tablet since they don't rely on "dragging", just clicking. If you wanted to text-parser or speech-parser the game, it would add another way to solve the puzzles, but relying entirely on it would be frustrating for players that don't have a means to enter either quickly.

[exo]

Nothing was more frustrating to me back on parsar games than knowing how to solve the puzzle, but the game wouldn't take what I was typing.

Something as simple as "slay dragon" had to be typed "kill dragon with sword"... And if I remember correctly, just typing "kill dragon" ended in your death without really explaining why. I expect my character, who is carrying a sword, to use their bloody sword when i tell them to kill something.

Looking back, it makes sense - but as an 11 year old it was ridiculous and frustrating to finally pull up a hint and see I was just typing it wrong.

Parsar's major shortfall was limited vocab and typing the exact actions required by the author.

P&C's major shortfall was always pixel hunting for me. I always wished I had a friggin "look" command in a P&C room to identify the important items. Imagine for a second you typed look in an old parsar game and it listed 100 items, but you had to try interacting with every single one of them to find the 1 you can use - that is a pixel hunt to me, just simplified to clicks instead of typed lines.

In the end I'd rather take the pixel hunts and complex inventory puzzles over having to guess what some guy wants me to type. Besides, you can't hide unicorn bridles inside boats and expect the user to look while standing right on top of it in a p&c game. *grumble grumble kq4 grumble*

[Kisai]

Originally Posted by exo

Nothing was more frustrating to me back on parsar games than knowing how to solve the puzzle, but the game wouldn't take what I was typing.

Something as simple as "slay dragon" had to be typed "kill dragon with sword"... And if I remember correctly, just typing "kill dragon" ended in your death without really explaining why. I expect my character, who is carrying a sword, to use their bloody sword when i tell them to kill something.

Looking back, it makes sense - but as an 11 year old it was ridiculous and frustrating to finally pull up a hint and see I was just typing it wrong.

Parsar's major shortfall was limited vocab and typing the exact actions required by the author.

P&C's major shortfall was always pixel hunting for me. I always wished I had a friggin "look" command in a P&C room to identify the important items. Imagine for a second you typed look in an old parsar game and it listed 100 items, but you had to try interacting with every single one of them to find the 1 you can use - that is a pixel hunt to me, just simplified to clicks instead of typed lines.

In the end I'd rather take the pixel hunts and complex inventory puzzles over having to guess what some guy wants me to type. Besides, you can't hide unicorn bridles inside boats and expect the user to look while standing right on top of it in a p&c game. *grumble grumble kq4 grumble*

I have to agree on the pixel hunting vs vocab hunting.

I was playing the Laura Bow game the other day, and ran into the vocab hunting several times, even though I clearly recall solving the game like 20 years ago. I think in any "newer" text/voice parsing would have to make some assumptions based on what is already in the inventory.

The P&C, even in modern games, lacks clues. For example in the strongbad game, you end up with several inventory items that are not used in solving puzzles (like the metal detector) but if you're not interested in achievement hunting, you might neglect to use it in the last episode when it is used in a puzzle. Likewise in Sam and Max, you might create the cake right off or you might make something else in the kitchen in the second episode. So you might wander around for a while wondering where the other food item is used instead of the cake. The clue of course is in a conversation loop.

So a way to make the P&C more fun but less frustrating is to first reduce the number of combinations things can be used with. So a room with 10 hotspots could have a hotspot activated and list inventory items for which there is a use or joke for, and cross off items already tried. Once all the items have been tried or there is nothing new to learn, the hotspot will just have a red X or something icon, where the character will respond with something logical or witty "I don't think I can use this item... yet" or "Maybe if I had a (clue)" if the character has missed picking up the (clue) item needed.

To put this in the text parser perspective, following "look around" the game would narrate or list all the hotspot objects in the room, and the player could "use X on Y" using any grammatically accurate phrase. So "use pot on stove" , or "use pot" (while near stove, if the pot is in the inventory) or "use stove" (if the player has the pot) have the same result. The other thing, is that if the player is in the right room, they shouldn't need to move up to the stove to use it. So say there is a stove and a sink, and the player wants to "use pot" and is near neither, the game should suggest "With the stove or the sink? Try getting closer" Thus providing two suggestions. If they're not in a room with either, and neither were described when looking around, place the joke there. "use pot", standing in the bedroom, "you'd love to but the stove is in the kitchen" Or something snarkier like "you pretend to cook using your pretend stove, and eat your pretend meal, yumm it was tasty, but you're still hungry."

At any rate, some of the limitations of the old P&C games had to do with memory limits, since everything that happens in a room is loaded into RAM, and unloaded when the room changes.

In the text parser games, the AGI games only had limited defined variables, so there was only so many combinations that could be programmed into the game. In fact some of the game code has things like said("W,X,Y,Z") || said("Z,X,Y,W") || said("Y,Z,X,W") , where it was looking for different combinations of the same words instead of said(("W"||"X")&&"Y"&&"Z") and only looking for a SVO (Subject, Verb, Object) match. I suppose one good way of improving text parsers would be to incorporate auto-complete, restricting the auto-complete to the object names the player has or can use immediately.

But anyway If Telltale wants to keep the look and feel of the old sierra games, it's better to keep what fans are familiar with (be it the text or P&C) but make it have less limitations.

[thejobloshow]

I liked how it was utilised in Leisure Suit Larry 7.

You clicked an item and had one or two options with how to interact with it including being able to type in a custom verb. It's a lot of work but it changes up and enhances the game for me.

[gamingafter40]

I play a lot of really old text adventures, so I'm used to parser wrestling. But dealing with the parser's grammatical limitations and phrasings the designer did not foresee can be annoying sometimes.

In King's Quest II, I ran into trouble because (SPOILERS!)...

KILL DRACULA WITH STAKE did not work; nor did STAKE DRACULA; nor did USE STAKE AND MALLET ON DRACULA. I finally figured out that a simple KILL DRACULA does the trick, as long as the proper items are in inventory.

[BagginsKQ]

IIRC, in LSL7, if you discovered a useable verb, that verb would then be added to your list of useable verbs (a menu). At which point you could then try each verb out on various things for differetn and often hilarious reactions.