smart art & entertainment

For the past two weeks I talked about the impact of games on people and I presented some attitudes designers take when confronted with the topic.   

So what is the moral of the story? Do game designers think about the impact of their work on our society? Do they aim to make the world a better place? Do they start with such noble goal in mind?

The answer is Yes and No.  I picked the three designers above to talk about because they all had interesting things to say.  Many others (I won’t mention names), have simply said that their games have a positive impact on our society because they present the player with accurate knowledge about the world, such as the historical events of World War II.  I, for one, do not believe that presenting our players with accurate information should be considered to have a significant positive impact on our society.  Yes, playing as a soldier in Normandy on D-Day can teach one some important history lessons, but we should not aim for that.  Knowledge and information should not be our goal.  If all else fails, it is a great place to land on, but I don’t think we should start with that.

I consider that the three designers above rank higher than average when it comes to how they assume responsibility towards society.  They all recognize the potential impact of games on society and I believe most of their games have a positive impact on people who play them. Civilization inspired many people to study history, Bioshock raised many questions about how safe is it to be convinced that something is “right”, and Fable and Black and White made us realize that we are all naturally good. 

However, none of the three designers seems to start by thinking of the impact of their games. Their achievements seem to come naturally from their talents. Their beliefs seem to come out through their games unintentionally.  It may be that at the early stages of design they are not even aware of what their players will learn through their games. 

I believe there are two things to be said about this:

1.       Congratulations to Sid Mayer, Kenneth Levine and Peter Molyneux for making games which have a positive impact on people while still keeping their mass market appeal.

2.       If Sid Mayer, Kenneth Levine and Peter Molyneux are able to make games which have a positive impact on society without thinking about it, it makes one wonder: what if they started with that goal in mind.

The question to be asked is: What if designers started with a goal of having a positive impact on society, and then build a game on top of it, how much more impactful the game could be?

I agree that it is dangerous to think that you have your head screwed on straight to teach others something valuable; however, isn’t that in itself a lesson worth teaching?  Shouldn’t we at least aim to make the player question his convictions about that? What about racism, poverty, sexism, drug use, politics and wars? What about environment? We don’t need to tell people what is good and what is evil. However, we could show them at least how little they know about good and evil, by making them reassess their values, and question their beliefs.  Watch the movie Crash if you haven’t seen it already and you will see a perfect example in cinematography of what I am talking about.  The movie doesn’t try to convince the viewers of any right or wrong, but it lives them baffled and afraid to judge.  Out of all mediums I believe games to be the most impactful form, since games are interactive. Thus we could project powerful experiences through games. We could reduce the crime rate, decrease drug usage and bring world peace. The reason while this sounds so foolish and unrealistic is because we never take the subject seriously enough to try to do something about it.

 The only thing that keeps us away from the Promised Land is ourselves.

Last week I mentioned some of the positive and negative effects which games have on people. I finished my post by asking the question “How many designers make games with that though in mind?” Today I will quickly compile the answers of three legendary designers on the topic: Sid Meier, Kenneth Levine and Peter Molyneux.

On the day Sid Meier received the Lifetime Achievement Award at GDC 08, I had an opportunity to ask him why does he make games. He answered that the reason behind is “the fundamental desire to make a game that hasn’t been written yet.”  In another question regarding the impact of our medium on people, he commented that it is an issue of raising importance:

“One of the issues that will become more important is … what do we do with this industry art form. Is it just about sitting in front of the computer? Is there some positive effect that we can have? Do we take responsibility to recognize that? ”

During another GDC session Kenneth Levine was asked if he set out to create a moralistic message in Bioshock?  His answer was: “No, I didn’t set out for that. We set out to make this great game with the story and the message presented itself.” He later commented:

“Frankly, I don’t really know if I have anything to teach people. And I’m not that confident in what I learned and what I could teach. I’m more interested in asking a lot of questions, because that’s how I go through life. I look at this thing and I get confused and I say well, I don’t know… Is it this, or is it that? Because once you know the answers, and you think you know the answers, that’s when you are in real trouble. The responsibility of great literature, great fiction, is to ask those questions, not to present you with the answer. … Otherwise lots of creepy stuff can come out of the teaching, when the teacher doesn’t have his head screwed on straight. And I’m not so confident that my head is screwed on straight enough to ask those questions. […] I gather my shoes and I wonder where they are made, but I’m still wearing them. They may be made by this little girl in a factory somewhere… what does that make me? I don’t know. But they are still on. Look at your shoes. They are still on.”

Another great opportunity I had at GDC was to have an argument with Peter Molyneux regarding his take on game design. Despite the fame surrounding the legendary designer, I was unable to play any of his games made after Syndicate due to his emphasis on Good vs Evil. Having a good alignment by nature, I found Black & White and Fable unapproachable since these games made me feel like all the choices were premade for me. I always felt that Peter decided exactly how I am going to play when he made me decide between saving the villagers or killing the villagers.  Real life is never that simple, as your choices are hard to understand. Many times you are forced to decide the lesser of two evils, a concept which does not exist in Peter’s universe. I didn’t feel that I learn anything about the world in his games, since the world is not as simple. Additionally, past a certain age, we all know the consequences of making good vs evil choices. So what is it that you Peter is trying to teach us?

Ken gave his opinion on Peter’s games: “Peter when he talks about morality, I don’t think his games are about this is right and this is wrong. His games are more about: Try this, try that, you tell me.” Peter was quick to respond: “That’s exactly right.”  However, since trying either alignment and suffering its consequences does teach one about the alignment itself, I would argue that his games are in fact about teaching good and evil. And Peter’s response only reassured me I am right.

However, when I confronted Peter directly with the question of “what do his games try to teach us” I’ve got a different answer. It took us some time to find a common vocabulary, but I feel that I understood the goal which Peter strives for. My understanding is that his games are not a journey for discovering the world we live in, and they should not be judged as such. His games are aimed to be a journey of self-discovery.

Peter has worked really hard to put an identical amount of content in both available paths in Fable: good and evil. To his surprise, only 85% of people played as good. When creating Fable 2, he decided to help people discover just how good they are. Thus, his aim in Fable 2 is to make it hard for one to be good. When playing as a good hero, the player will be faced with numerous sacrifices which will make one question himself and his dedication to good. We all take for granted that we are good. But how good are we really? We never know until we are placed in an “impossible” situation and discover with great surprise that we are not that good after all.

Next week, I will continue with the conclusion of this survey. Stay tuned.

I type this as I fly over US in my return trip from GDC 2008. I had a really great time, attended numerous lectures and met a lot of interesting people. I also managed to gather new material regarding one of the topics which have interested me for a long time: what do we do about the effects of games on people.

When somebody draws a line on the ground people feel the need to rally and take sides.  Thus when some desperate politician went on air claiming that violence in games has negative effects on children, the rest of the world felt the need express their opinions in the matter.

Many have attempted to demonstrate how violence in games results in violence in real life. Others have shown how the explosion of games in the mass market is the cause behind the lack of family time in today’s society.  Also, the addictive qualities of popular MMOs (such as World of Warcraft) was held accountable for disrupting social lives.  

On the positive side, researchers have found that video games are enhancing numerous skills, such as communication, organization, logic, statistics and creativity. Games can also help bridge the gap between opposite sexes, prolong brain activity into the old age, and inspire many to study (Noah Falstein – GDC 2008).

Chances are that you are not new to any of these arguments. Maybe you are aware of them, and maybe you have made up your mind regarding the topic. Regardless of which side one chooses, almost everybody agrees that interactive digital entertainment is a format like all others (literature, TV, movies, music, art) and as such it almost certainly does have an impact on the people consuming it. This determined me to try to get some insight into the minds of some of the highly acclaimed designers.  I find that we as gamers and game developers are very quick to defend our creations and claim that games don’t have a negative impact more than any other forms of media do. We point our fingers at all the research which shows the wonders of video games.

But what do we do about that? Do we design games with that thought in mind?  Do we aim to have a positive impact on our players, or we just design games and without any thought of the implications, and just hope for the best?

Next week I am going to survey three highly praised game designers and present their answer regarding the topic.

    

Putting all together

In the past few weeks I have described some of the technology available which I believe can be used in developing Next Gen Text interfaces for games. Today, I will briefly cover some of the other parts necessary for one to build a game.

Story

If building a text interface for our game seemed hard, having a computer generated story that works seems downright impossible. The level of complexity between creating immersive drama (at a macro level), and creating coherent sentences to express the drama (at a micro level) is outrageous. It is one thing for one to understand the language, and another thing for him to be able to talk the language. The latter is infinitely more difficult.

Yet the task could be solved through the same solution everything else gets solved: Divide & Conquer. Instead of trying to solving the entire story problem at once, we could divide the problem in two, and solve each individually.

Low Level Story

At the lowest level of a story we have sentences. These combine in clever ways to form paragraphs. For a computer to generate sentences is fairly hard. And generating paragraphs is even harder. One way of avoiding this problem altogether is to make an actual writer write the paragraphs of a story and let the computer only alter them. For example, let the following sentences be part of a game. Please note that the sentences in green color are written by a writer while the red markings indicate modifications made by a computer. Blue color represents user input.

“The torch is all that brightens the cell. An old picture hangs behind the glass door on the right.”

> Smash the glass door with the torch.

“The torch is all that brightens the cell. An old picture hangs on the right. There is broken glass on the right.”

> Use the torch to burn the picture.

“The old picture burns. There is broken glass on the right.”

Notice how the computer does not create a scene from scratch. It only modifies a scene at your request. The beauty of this process is that it allows us to keep the best of both worlds: expressiveness and interactivity. And, while the expressive power of an actual writer is not something new to text adventures, this level of interactivity certainly is. The key point here is that the game designer may have never intended of letting anyone burn the picture. He may have never explicitly coded that possibility. The GAME allowed the player to interact with the scene as if it was real. To underline this important difference consider the same scene with a different course of action:

“The torch is all that brightens the cell. An old picture hangs behind the glass door on the right.”

> Open the glass door and look at the picture.

“You open the glass door, and look at the picture. In the picture you see the portrait of a young lady. “

> Draw glasses around the girl’s eyes.

“You draw glasses around the young lady’s eyes.“

> Color her teeth black.

“You color her teeth with black ink.“

> Look again at the picture.

“In the picture you see the portrait of a young lady with glasses and black teeth.“

If it wasn’t sufficiently obvious that a designer cannot think of all the possible ways the player might want to vandalize his scene, this second example hopes to clear any doubts. As such, it is demonstrated how using the techniques outlined in my previous blogs a Next Gen Text Adventure can allow for unparalleled levels of interactivity.

High Level Story

It is well known that a combination of random paragraphs does not make a story. And many would argue that the story is not much by itself without the storyteller. A story needs to provide certain features in order to make its audience feel it and enjoy it. A story needs anticipation, excitement, surprise, etc. Also these elements do not come in random order, but they carefully build on each other. Traditionally, it takes a lot of skill for a writer/storyteller to be able to raise these types of feelings in us. So how can we expect a computer to do it?

There has been a lot of work done on this topic, some of which I already mentioned in my previous blogs. Both Chris Crawford and Michael Mateas have covered the topic in their research. For more information please check out Chris Crawford on Interactive Storytelling, and the Interaction and Narrative chapter in The Game Design Reader (edited by Katie Salen and Eric Zimmerman). In this section I will VERY briefly discuss Chris Crawford’s work, as I am more familiar with it.

Mr. Crawford purposes the eradication of the word “storyline” and introduction of the word “storyworld” . The implication is that the writer has to focus on writing multiple scenes which all belong to the world of the story. The scenes are categorized by similarities, outcomes, flow, etc. The scenes are linked at runtime by the computer based on the player’s actions. Thus, the task of having the computer generate plausible stories, is reduced to the task of having the computer link certain scenes.

Mr. Crawford then purposes the introduction of a Drama Manager or a game component which watches over the level of entertainment experienced by a player. This is achieved by ranking every scene with a certain entertainment score. If the player’s actions take him into scenes which constantly have a lower entertainment score, the Drama Manager steps in to spice things up a bit, by introducing some new scenes/events. Great trick!

For example:

Scenario: The king is fighting a war in a far away land. The player’s actions don’t seem to have much impact on the war and he doesn’t go to the battlefield.

Drama Manger Alert: The player felt insignificant for too long.

Action Taken: End the war, bring the king home and give the player a chance to mess things up a bit directly with the king.

So Chris Crawford is a genius, but everybody already knew that. To find out a lot more detail on this read Chris Crawford on Interactive Storytelling.

Puzzles

The puzzles are an interesting part of a Next Gen Text Adventure. And they are interesting as they cannot fully be play tested. Due to the nature of our game, with such a high level of interactivity, it is impossible to think of all the possible ways of solving a puzzle. It is also very hard to make sure that the player CAN solve a puzzle. Going back to our previous example, if the player burns the picture before he looks at it and the picture contained a clue of how one could open the door, the player may not be able to get out of the cage anymore.

This may call for the need of a Puzzle Watcher. Every puzzle is designed with at least one solution. The Puzzle Watcher is watching over to ensure that all the components needed to solve a puzzle are still existent. If they are broken, the Puzzle Watcher can step in and erase the puzzle, or introduce the missing elements in the game.

Other than that, the puzzles are very “normal” or “traditional.”

Conclusion

And this concludes my Next Gen Text Adventures series of blogs. Next Gen Text Adventures are ready to come out. The technology is here for us (and has been here for a while). However, before we are going to see one on store shelves, two important things must happen. First of all, we have to rid ourselves of our backwards thinking. Thinking in new ways is hard, but is the only way to evolve. Secondly, Next Gen Text Adventures need to prove themselves as commercially viable products. So keep an eye on Storytron and when it comes out next month, make sure to buy a copy. I assure you that Storytron has more cutting edge technology in it that the Madden series did for the past decade.

 

Understanding of the combination of words in a sentence

In order for the computer to understand the relations between the words in a sentence, it needs to perform two major steps:

1. To be able to parse the sentence in a structure from which it is easier to derive logical relations from.
2. To be able to derive logical relations between the words in the structure constructed in Step1.

I will give a quick overview of both these steps.

 

Parsing:

In order to have a sentence in an understandable form, we need to be able to parse every word in the sentence in a logical tree, based on a Context Free Grammar for the natural language the sentence is written in (in our case, English). The following is an example of a tree generated for the sentence “I shot the sheriff”:

 

Input Sentence ->I shot the sheriff

No. of parsers : 1

[ S2

    [ NP1

        [ NPRO1 <I> ]

    ]

    [ VP2

        [ V-with-NP-complement1 <shot> ]

        [ NP2

            [ Det1 <the> ]

            [ Nominal1

                [ N1 <sheriff> ]

            ]

        ]

    ]

]

 

In a simple scenario as the one above, it may be difficult to notice the advantages of having parsed sentences into trees. However, as sentences become more complicated, parsing becomes a necessity:

 

Input Sentence ->which bandits does the sheriff want to shoot

No. of parsers : 1

[ S4

    [ WhNP2

        [ WhDet1 <which> ]

        [ Nominal1

            [ N2 <bandits> ]

        ]

    ]

    [ AUX1 <does> ]

    [ NP1

        [ Det1 <the> ]

        [ Nominal2

            [ N1 <sheriff> ]

        ]

    ]

    [ VP4

        [ V-with-VPto-complement1 <want> ]

        [ VPto1

            [ Vto1

                [ To1 <to> ]

                [ Vinf1 <shoot> ]

            ]

        ]

    ]

]

 

This can be further complicated by addition of other forms of speech such as prepositional phrases, adjectives etc.

In general, our brains apply the process of parsing automatically to sentences, such that we don’t even notice when we do it. We only become aware of this process, when we try learning a foreign language. Parsing is a crucial process for any written or verbal communication. Therefore, computers must know how to do it too. (SHOCKER!)

The parser used to generate the above sentences is a simple chart parser I built in one afternoon. Due to the wealth of research in the area, the technology is very simple to build by anyone interested. And while the accuracy of my parser is not yet high enough to be used in a game, Eugene Charniak reported ten years ago that lexicalized statistical parsers were achieving an accuracy of 88%. Considering that parsing has an upper bound of 95% (people agree on only 95% of the cases on an interpretation of a sentence), 88% is a really good achievement. And 88% ten years ago suggests that we are very close to human parsing accuracy today.

 

NOTE:

Furthermore, the parsers referred to in Eugene’s research (Statistical techniques for natural language parsing) are tested on large text corpuses. A text corpus is a large collection of texts from different fields. For example, the Wall Street Journal corpus is the collection of all Wall Street Journal articles ever written. This means that the parser has to deal with many words which do not belong to its lexicon (dictionary), phrases and words that come in and out of style over time, and multiple domains (weather, market, scientific research, politics etc.). Additionally the average sentence in the Wall Street journal is 23 words in length.

The point I am trying to illustrate here is that we currently have the technology to do OVERKILL in parsing for a text adventure. Having a limited lexicon of 2000 words, and a limited domain, anyone with basic computer linguistics knowledge should be able to build a parser that achieves human parsing accuracy.

 

Building logical relations

Once the sentence is accessible in a logical parse tree, and the meaning of every word is well understood, we can proceed at building logical relations between words and meanings.

A very simple way to perform this is by using First Order Logic on the words in parsed tree:

Using simple first order logic we can infer the meaning of most sentences. Higher Order Logic is rarely required in the average sentence; and again, for the purposes of a game, it can most certainly be avoided.

 

Conclusion:

Currently the technology is available to everyone interested in performing Natural Language Understanding on a limited domain. When dealing with the confinements of a game, we can ensure that all words and sentences are well understood by a computer. Thus, we can now build text interfaces to interact with game worlds in a very natural way. Next week, I will tie everything together, and show how a Next Gen Text Adventure could be built using the technologies described in the past few weeks.

 

In order to build a Next Gen Text Adventure, the computer needs to understand any command given by the user. This means that our game needs to understand the “meaning of every word” in the sentence specified by the user, as well as the “meaning of the combination” of words. Today I will cover the understanding of every word.

Understanding of every word in a sentence.

Choosing a sublanguage:

Natural language has a very large number of words. For example, English reaches about half a million words without counting scientific terms. To add meaning to every word in a natural language seems an impossible task. While I do not like to use the word impossible, I think it is safe to say that with the technology currently available to us, the task does seem very difficult.

Yet, to make a Next Gen Text Adventure one does not need every word in English. As Chris Crawford points out, if we look at the top 2200 most common words in English, we will find words such as “handkerchief,” “photography,” and “ray.” One can certainly build an adventure with a sublanguage of English which does not use these words. Thus, instead of trying to construct a dictionary for our adventure with every word in English, one could focus on the top 2000 words. It is a hard task, but certainly not impossible. If you have doubts, consider for a moment that every book has an upper bound on the number of words it uses. We could safely say that everything is communicated in a sublanguage of English. For example, King James’ Version of the Bible contains 12,143 different words.

One could ensure the 2000 words limit, by carefully crafting a story using only 500 words and adding three synonyms for every word used. This will capture the words used in most variations in the telling of the story.

Adding meaning to words:

After choosing the words, we need to add meaning to them. While the process of adding meaning is a very hard task, a lot of research has already been done on the topic. While there isn’t yet a perfect solution, there are many proven ways which demonstrate how to add meaning to a specific domain. WordNet is a large database online which contains the meaning of every word in the form of trees which are connected to each other through certain nodes, forming a net.

A similar WordNet could be build for our game. Our net would be smaller in size (only 2000 words) and more domain specific (specific to our adventure). The following image shows an example of how such a net may be built:

Each word is categorized and linked to the other relevant words in the net. Thus, every word in our game would have a meaning attached to it. For example, some actor in our game could be categorized as follows:

[HUMAN-7:            ]
[ instance-of: Human ]
[ name: “Jill”       ]
[ legs: 2            ]
[ mother: Human-23   ]
[ father: Human-111  ]
[ job: Student       ]
[ address: (unknown) ]

A similar method could be used to categorize the verbs.

I am certain there are more efficient AI techniques (of which I am shamefully unaware of) for dealing with this scenario. Yet, I just want to make quick description of how meaning of a subset of English is technically achievable.

In conclusion, the designer has to preselect a limited sublanguage, such that the words all belong to a limited domain, which is easy to categorize.

Next week I will cover how one can approach the understanding of whole sentences.

This is a continuation of my last week’s blog on Next Gen Text Adventure. Before we dive in deep in explanation on how such a game is possible to build, I would like to introduce you to some of the scientific research made on related topics.

Previous Related Work

Next Gen Text Adventures is a term I invented myself. There are already a few terms in circulation which refer to topics very much related to what I am discussing here. Two terms that come to mind are Interactive Storytelling and Interactive Drama. Interactive Storytelling is described by Chris Crawford in his marvelous book Chris Crawford on Interactive Storytelling, while Interactive Drama is explained by Michael Mateas in his Ph.D work, which is available in part in the Interaction and Narrative chapter in The Game Design Reader (edited by Katie Salen and Eric Zimmerman). Both of these works are truly revolutionary and must a must read for anyone designing interactive entertainment. Even though Mr. Crawford does sustains from the beginning of his book that storytelling and games are very different from each other, it is my belief that any game designer out there has a lot to learn from reading his book.

Nonetheless, I chose not to use either term on purpose. Both, “Storytelling” as well as “Drama” are hard to measure or generate by a computer. Many people have done a lot of work researching the topics, while a lot of other people have put in a lot of effort to criticize the research. Just like every other field of art & science, the hard labor is always performed by a select few who believe that impossible is nothing, while other “great” scientists waste their existence criticizing others. (Forgive my off topic bitterness.)

My “Next Gen Text Adventure” is a smaller beast in relation to “Storytelling” or “Drama.” To keep things simple, our initial Next Get Text Adventure, does not guarantee neither Story nor Drama. It is also specific to text, in order to avoid dealing with harder topics such as graphics and sound (which both Storytelling and Drama may or may not include). First I will focus, on how a computer may be able to understand language and generate puzzles.

Computational Linguistics

Computational Linguistics is the field of science which deals with the understanding of language by a computer. Whether you are computer scientist or not, you can probably imagine that the act of understanding is very hard. This has been a problem since the first days of computing and it is still remains an unsolved problem today. Graham Hirst, a distinguished computational linguistics professor at University of Toronto, has mentioned that the past half a century of research on this field has yielded only about 40% progress on the long road to having the natural language understanding problem solved. The deceivingly small ratio is not a good reflection of progress speed, but more accurately a reflection of how hard the problem really is. Furthermore, the good news is that progress is not made in linear steps by exponential. The closer we arrive at a solution, the more parties become interested in the commercial potential of the solution and thus the more money gets dumped into research. Therefore, many computational linguists expect to have multiple practical applications out in 5 years with a complete solution by 2020.

Yet, there are already a lot of applications in the world using various related technologies. If you have dialed the customer service number for any large phone or ISP provider in the past two years you may have been surprised by a computer assistant who seemed to understand you.

Conclusion

In my next blog I will explain how one can use some of the past research done on the topic of Interactivity with the available research in computational linguistics in order to build a Next Gen Text Adventure.

Text adventures are a thing of the past. Or so does everyone think. Including myself.

Text adventures have been dead and buried for over two decades, which technically means they are well decomposed by now. Many never had the chance of playing one, and may never play one either.

What is a text adventure?

Before I go on, rambling about text adventures, allow me to make a quick attempt at of describing what a text adventure is, so that we are all on the same page.

Text adventures are games which rely completely on text. Both, input and output are formed of a series of words (which sometimes form sentences). The following is “screenshot” of one of the most popular adventure games of all time.

These games have been applauded as well as criticized by many. Since I wasn’t even born at the time, I believe it is a little unfit for me to speak on the topic. I am not here to take sides. I simply dug up a little bit of history in order to make sure we are all thinking about the same thing.

Under very gloom observation, text adventures could be seen as a weak combination of two mediums: books & games. The result is a very shallow narrative, with very little interactivity when compared to both. In addition, the game puzzles are made very frustrating due to the weak interface. The game consists more or less of guessing what the designer intended the player to do.

Text adventure supporters see the games in a much brighter light. While the main components such as narrative and interactivity may be weaker than in other mediums, the combination of these components may result in something greater than the sum of its parts. Thus, text adventures are seen by some as an important medium in its own right. The games are able to convey a certain type of entertainment which is rarely found in other places.

Again, I am not here to agree or disagree with either theory. In my ignorance the only text adventure I have played is an entry for the 2002 Interactive Fiction competition: Photograph by Steve Evans. I have enjoyed the game very much and I highly recommend it to anyone interested in this medium.

What is a Next-Gen Text Adventure?

Now, that we are on the same page, try to forget everything I said, and imagine a real “text adventure.” Make an attempt to erase any previous knowledge you have gained so far, and picture what a eight year old would think when confronted with the term “text adventure” for the first time. Imagine a book which listens to what you do, adapts to your needs and reacts to your wishes. Imagine an ASCII world dynamically created just for you, based on who you are and what you do. Impossible? Maybe! But try to picture it nonetheless. Forget what is feasible from a programming perspective. Forget the complexity of such a “text adventure.” Just focus on your experience as a player! Take a couple of minutes, to do this now. I’ll wait for you.

So how was it? If you did the exercise and you pushed your imagination a little bit, I am sure you will agree with me that the game you just pictured, would be the best game ever. We might as well call it a drug, since it gives you exactly what you need, adapting to your wishes and reacting to your actions… a virtual world, made dynamically just for you, ever changing… growing just like you… It is the ultimate game.

Now let’s get back to reality. I’m sure you have lots of questions. Like: What is the point of thinking about such a game? We are ages away from such an achievement.

And I will agree with you. Partially! While you are probably right that a game which listens and adapts to “everything” you do is sometime far in the future, I believe that we currently have the technology to build a game which listens and adapts to “many things” you do. Allow me to explain.

First of all, when I say game, I am strictly referring to “text adventures.” A game that is procedurally generates the graphics (architecture, character models, etc.) of a game is a little hard to chew right now. However, since “text adventures,” have paved the path to graphical games the first time around, it may be worth starting again at the source.

Second of all, letting the player to do “everything,” and trying to adapt to “everything” is again maybe too high of a goal for a first step. However, “everything” within a smaller universe may be doable.

In my next blog I will take a shot at convincing you that a “Next-Gen Text Adventure” may not be as far as you think.