I'm surprised the article doesn't mention the Unknown Worlds team, who are using this model to produce Natural Selection 2.
I suppose it's less revolutionary for them, as they have a history of community-funded development. In the original Natural Selection, players could pay a modest fee to enroll in the "Constellation Program." Members received a variety of perks, including early access to beta releases and an in-game/forum icon.
Of course, the NS2 developers have a history of transparency and delivering on releases, and have ready access to a supportive community. So the model works for them. I doubt a new studio would get similar traction. I suspect it will be even harder once a few community-funded games fail, leaving behind a bunch of pre-rendered screens and angry gamers.
IME, the implementation is a train wreck. I have a Visa card through Bank of America, and the first time I ran into the "Verified" prompt, I was positive it was a scam:
The form is in an iframe, so it's not even immediately obvious whether it's encrypted
The iframe contents (for BoA) are hosted at bankofamerica.vbv.cyota.com, not, you know bankofamerica.com or visa.com
The first time it popped up, it prompted me for the last four digits of my social security number, to "activate my account"
Actually, the game uses a reductionist damage system, which only consists of three levels (normal, bloodied, broken). The ill-named "stay up" stat is apparently used to determine whether or not a particular activity causes a character to transition from one damage state to the next.
While it's not as sophisticated as, say, the old Cyberpunk 2020 damage system, it certainly isn't a "Big Red Bar" system.
The last few chapters in the book are a bunch of "wouldn't it be cool if..." game concepts. Actually, much of the rest of the book was filled with critiques of popular games and various bits of "if only they had done it this other way" armchair speculation.
There is a tradition, in chinese and japanese landscape painting, of telling stories by presenting a activity-rich space. While you can stand back and look at an entire painting, it's best experienced by letting the eye wander across the various little vignettes happening in different parts of the work. Here some children are playing with a dog, over there the emporer is receiving guests, etc.
I suspect the way forward with interactive storytelling is to rediscover this notion of spatial storytelling. This requires authors to abandon plot as a mechanism for communicating with the reader. Instead, the author is left with setting and characters.
Traditional authors will balk at this, but as the field matures, I think a new generation of authors will find their voice in creating rich, interactive worlds.
I can't speak for the Blackberry or Sidekick, but I own a Treo 600, and am fairly pleased with it. Once you add enough software, it's a pretty complete device. After ferreting out the right apps, I now can:
Send/receive mail using POP3S/SMTP-TLS, via SnapperMail.
PdaNet in particular is killer if you have a laptop handy, but no internet access. I get about 144k down via Sprint PCS, which isn't screaming, but it's adequate for browsing and downloading small files.
how shallow the plot lines and characters are of many modern games
I suspect this will become a larger problem in the future. The problem of lifeless characters can be solved, somewhat, by improved character animation techniques. But without fundamentally changing the way characters interact with their world, you're just going to end up with hyper-realistic quest-dispensing vending machines.
Bulking up backstory and character dialogue is a short-term patch, but ultimately characters will need to become as emotive as they are well-rendered. And they'll need to interact with the world's social landscape as effortlessly as they interact with the world's collision meshes.
Scott McCloud discusses this at length in Understanding Comics. As a rendering of a person becomes more stylized and abstract, the viewer begins to fill in details on their own. Not only can our minds fill in more details than an artist can draw, but the content we fill the drawing with is our own, which makes the character more accessible.
A similar effect occurs with The Sims. Their reductionist design and behavior allows users to ascribe all sorts of baroque narratives to their simple actions.
Maybe the bumblebee hunter will get some quest that is comparable somehow, but he will still want to do what his friend did.
My thinking was that quests would be dynamically generated. So there would be the implicit assumption that you'd never get the same quest as someone else.
However, that does bring up an interesting point, which is grouping. It would be nice if emergent stories could be tailored to the skill sets of everyone in the group. Maybe our hypothetical bee-hunter has grown to be immune to certain types of toxins. Perhaps, rarely, the game could throw in a key guarded by giant wasps or something.
This technique is used from hell to breakfast in stories, where a character might be revealed to have some vital piece of information, or an unexpected skill, or connections in the right places. You know the drill:
Uber-skilled Death Machine: Rushing the gate is impossible. We'll never get in the castle. Noob: My cousin works in the stables. He could sneak us in.
I tend to thing MM games are going to be the killer app for interactive storytelling. If you think about it, there are really only three ways to do interactive content for MM games:
1. Create a fixed set of non-interactive content. Everyone in the game world does the same quests, reads the same dialog, fights the same big bad guy for the same reason. Alternately, a few people get to participate interactively, while the vast majority simply watch (or show up for the epic battle). 2. Employ humans to create custom content for each player. Figure out a way to do this for under $20 per player per month. 3. Programmatically create custom content for each player.
There's actually already quite a bit of work going on in this area. The most interesting approach I'm familiar with is Padraig Cunningham's work with case-base systems. In a nutshell, Cunningham proposes borrowing plot recipes from the russian structuralists. One then plugs existing NPCs into the appropriate roles (wise old man, bearer of gifts, etc). The system scales by allowing one NPC to play multiple roles in multiple simultaneous stories, so long as the roles don't conflict.
So, not unlike real life, your view of an NPC (as a dastardly villain, unexpected benefactor, or innocent in distress) depends on what you're doing when you meet that person.
For more information on Cunningham's work, check out his publications. In particular, the paper entitled "A Multiplayer Case Based Story Engine."
Personally, I'm working on a system for generating plot-rich city histories (via personality and relationship modeling). My goal is to get to a point where one could plop down a new town, set some parameters, and "age" the town n years (and get sensible personal relationships, family trees, interesting local history, street layouts, etc).
This theoretical programmer was mistaken. First off, each screen was created by artists, true, but it was then rendered by computer.
My example is handily defeated; Myst was a poor choice. However, the analogy works with countless SCUMM-style 2D adventure games. I suspect you grasp my point (that this is a matter of a paradigm shift), which makes arguing the analogy sort of pointless.
This involves concepts, ideas, a whole host of fundamental knowledge of "how the world works".
Disagree. Again with the 3D graphics example. Apples are tasty, can be cut into wedges, and rot if they're left out in the sun too long. However, if you're trying to present an exciting chase scene, in which an apple cart gets knocked over, a 3D engine can abstract apples down to a model, texture map, and collision mesh.
I think we can establish similar abstractions for things we might want to present in the course of a story. Can I enumerate the optimal model? Nope. I can point you to people who have tried (Crawford, Bringsjord and Ferucci, Mateus and Stern, Cunningham, various russian structuralists and folktale analysts, etc), but no one has discovered the magic recipe.
Regarding the apple problem, we're obviously dealing with a lot more complexity here. Most toddlers have a workable mental model for how apples fall, but most adults don't understand why people behave the way they do. However, while the problem is more complex, I don't think it's intractable. And it probably doesn't require everyone to have a copy of Cyc on their desktop.
The problem with this is that it's such a big job. If you're going to provide that kind of freedom of action, you're looking at writing thousands of similar but separate stories. And no cheating like many games do, where no matter which branch you take, the impact is slight at best and the story remains mostly unchanged. We're hoping for replay factor here. We want total changes in the storyline. Some cheating this way is okay, because things happen that way sometimes through no act of your own, but still, too much of this and you lose the point of doing it on a large scale in the first place.
I anticipate that, at some point in the playtesting of the original Myst, someone probably said "It would be awesome if I could just walk around, and look at the puzzle from any direction I wanted."
And some patient programmer no doubt explained that such a thing would be impossible. Since each screen was hand-drawn by artists, you'd need thousands of artists spending years to draw pictures of a given room from every possible vantage point. And you'd need to ship gigabytes of image data with the game.
Of course, we now take that level of immersion for granted. And there aren't thousands of starving illustrators locked away in dungeons, drawing views for Doom III. Different views are created in real time by specialized hardware, using complex maths. When you zip through a room in Far Cry, how many thousands of paths could you have taken?
I tend to think the interactive "storytelling" problem will, given time, shake out the same way. We will stop working with stories as individually hand-crafted products (like paintings). We will instead begin to work with them as processes (like renders of a 3d scene).
This will require a lot of work to define the theoretical underpinnings of this process, and to create the tools people will use to create these potential-rich environments. It will also require a fundamental shift in the way authors/creators work. This will be hard. It will be resisted, both by technical types (who view the problem as insoluble) and by creative types (who view the loss of authorial control as anathema).
But it will happen, because the promise is so compelling. Perhaps because the experience of getting wrapped up in a computer game is much the same as getting caught up in a good book.
A story is a meaning applied to events after they have occured. A game is a game, like sports or a board game. You can only make a story out of it after events have been completed.
I will never understand why this old chestnut appears every time there's a discussion of interactive storytelling.
By your definition, fiction is impossible. When the author sits down with a blank sheet of paper, he should be stuck, since there are no past events for him to relate.
But of course, we know this isn't the case. Even school children are capable of inventing fanciful, novel stories. The path to interactive storytelling is collaboration between the player and the computer to produce a narrative which is both interesting to the player and dramatically compelling. The narrative is a product of this process, it is not the process itself.
Perhaps we need to come up with a new term. When people say "interactive story" they obviously don't mean "give me a book I can shove into my Playstation." They mean "give me a piece of software that allows me to have an experience similar to reading a good book, but also provides me agency, and allows me to control how I experience events."
The Bat does a pretty good job of this. In addition to various other processing rules, you can launch your own apps when a specific rule condition is hit.
I wrote/hooked a small program that gets called whenever I get new mail on my work account. It:
Checks to see if the current time is outside working hours (or during lunch). If so it..
Strips out HTML tags from HTML-only messages
Removes attachments
Removes MS Exchange LDAP nastiness from From: addresses
Re-sends the email to my mobile phone address
The app gets called with a command line parameter pointing to a file containing the message. So there's not much magic required to write your own extensions.
Version 1.01 was available through phone and mail orders, and then on a shareware CD-Rom. Players needed to purchase the CD-Rom and then call id to receive an alphanumeric code that would "unlock" the shareware CD-Rom and install the full version of Quake. Purchasers could also then order a full version of the Quake on an "unlocked" CD-Rom for an additional amount.
I recall much wailing and gnashing of teeth when my group of friends pooled our money to buy a copy.
Because it's a large system that will have to integrate with numerous airline systems from god knows how many vendors. And it will need to be maintained and patched. And it's a potential single point of failure (from a software standpoint; obviously they could stripe it across as much hardware as needed).
Even if CAPPS is only connected to ticketing and passenger information, a bug could result in a pretty nasty transportation snarl. Suppose airlines are unable to issue boarding passes for an hour, or an unusually large number of people were flagged for screening.
For any of these total-information-awareness type systems, one has to ask "what happens when some part of the patchwork breaks?" Even the most diehard "I have nothing to hide from my government" type understands that multi-hour flight delays are bad.
My box:
CPU: AMD Athlon(tm) XP/MP/4 1532 MHz
GPU: NVIDIA GeForce4 Ti 4200
Score: 1294
Now, if I go to the Futuremark Project Search, enter my exact system, and change out my GF4 for a Radeon 9700, the top results are in the 4500 range. Which would suggest that I could get a 200% performance boost by buying an ATI card. However, game benchmarks suggest the 9700 would buy me a much smaller gain (5-50%, depending on the game/resolution/options). IMO, 3dmark places a bit too much emphasis on the newest DX9 features to be useful as a general performance benchmark.
No doubt about that. But the current powerful oil companies would not be very excited about that unless they could ensure that THEY would be the powerful hydrogen companies as well.
Considering their expertise in seismic analysis, drilling, and natural gas processing, I think that's pretty much a given.
The Bat! is a lovely MUA, but its continued development is very strongly shaped by Stefan Tanurkov's vision of a good email client. This means it has a variety of non-configurable "features" which are unlikely to die soon:
The program will attempt to guess what reply number you're on in an email exchange, and replace the Re: in the subject line with Re[3]:. IMO, this is mostly meaningless clutter.
The Bat! will not append quoting characters to blank lines in quoted/forwarded text. Sounds great, until you want to forward a multi-paragraph article or code block.
Various other odd little anomalies.
Now, I still think it's the best option for mail out there, but I would kill for fifteen minutes alone with the source code, so I could correct some of the irritating little features.
You should probably have negotiated fixed SLA terms with definable metrics prior to signing the service contract. It's strange to be thinking in datacenter mode when you're wiring up your house, but you sort of have to.
I'm not just pointing the finger here, I went through this same thing two years ago with Verio. I was paying $300/mo. for "business" DSL, and had to deal with constant downtime, awful/slow tech support, and a billing office that kept reminding me that my contract included "no specific uptime guarantee" (but assured me I wouldn't be charged for the days of downtime I kept experiencing).
After my contract was up, I switched providers, moved to ISDN, and signed a contract with a fixed uptime guarantee. The few outtages I had after that were corrected within the day.
First off, obviously, Chris has done some amazing stuff with the Erasmatron. However, as a product, I don't think it has a viable future. It has some pretty significant shortcomings, and it would take an incredible amount of work to bring the engine up to modern standards. That said, Chris's documentation of his development is, quite simply, the best text out there in the field of interactive storytelling.
If you're interested in this sort of thing, you'd can get a good feel for the existing work in the field from:
InteractiveStory.net - Michael Mateas and Andrew Stern's interactive drama/believable agent project, and obligatory huge page o'links.
Erasmatron@Robotwisdom - Jorn Barger's excellent thumbnail sketch of Crawford's writings. In most cases, Jorn's synopsis is hyperlinked to the related page on erazmatazz
I'd also recommend the following papers (try CiteSeer before heading down to the library):
Selmer Bringsjord and David Ferrucci, Artificial Intelligence and Literary Creativity: Inside the Mind of Brutus, A Storytelling Machine, August 30, 1999.
Nicolas Szilas, Interactive Drama on Computer: Beyond Linear Narrative, 1999.
Antonio Furtado, Angelo Ciarlini, Plots of Narratives Over Temporal Databases, 1997.
Barbara Hayes-Roth, Robert van Gent, Story-marking with improvisational puppets, 1997.
W. Scott Neal Reilly, A methodology for building believable social agents, 1997.
IMHO, interactive storytelling is one of the most interesting cross-discipline computational problems out there.
If this could be made to work reliably, consider what it would do for sites like Penny Arcade or Something Awful. Sites like these are expensive to run because of the cost of bandwidth. If you decentralize the distribution, you could spread the bandwidth cost between all the users. Since most users pay a flat rate for their internet connection, in most cases the distributed cost would be $0. And since he's no longer paying for hosting, Lowtax could afford a better haircut.
Does Ebay have the word 'auctions' in its name? Does Yahoo use the word 'directory'? You can name your product something off the wall, and people will pick up on it.
Yahoo stands for Yet Another Hierarchical Officious Oracle. Remarkably, prior to becoming a household term, Yahoo actually had to have a descriptive name for people to know what the hell the site did.
Slashdot Mirror
I'm surprised the article doesn't mention the Unknown Worlds team, who are using this model to produce Natural Selection 2.
I suppose it's less revolutionary for them, as they have a history of community-funded development. In the original Natural Selection, players could pay a modest fee to enroll in the "Constellation Program." Members received a variety of perks, including early access to beta releases and an in-game/forum icon.
Of course, the NS2 developers have a history of transparency and delivering on releases, and have ready access to a supportive community. So the model works for them. I doubt a new studio would get similar traction. I suspect it will be even harder once a few community-funded games fail, leaving behind a bunch of pre-rendered screens and angry gamers.
IME, the implementation is a train wreck. I have a Visa card through Bank of America, and the first time I ran into the "Verified" prompt, I was positive it was a scam:
Actually, the game uses a reductionist damage system, which only consists of three levels (normal, bloodied, broken). The ill-named "stay up" stat is apparently used to determine whether or not a particular activity causes a character to transition from one damage state to the next.
While it's not as sophisticated as, say, the old Cyberpunk 2020 damage system, it certainly isn't a "Big Red Bar" system.
Maybe cool from a vinctiveness standpoint, but not for notoriety. Who would you brag to?
The rest of your warez group, who all want to see Steam fail spectacularly?
I suspect he would have loved to.
The last few chapters in the book are a bunch of "wouldn't it be cool if..." game concepts. Actually, much of the rest of the book was filled with critiques of popular games and various bits of "if only they had done it this other way" armchair speculation.
There is a tradition, in chinese and japanese landscape painting, of telling stories by presenting a activity-rich space. While you can stand back and look at an entire painting, it's best experienced by letting the eye wander across the various little vignettes happening in different parts of the work. Here some children are playing with a dog, over there the emporer is receiving guests, etc.
I suspect the way forward with interactive storytelling is to rediscover this notion of spatial storytelling. This requires authors to abandon plot as a mechanism for communicating with the reader. Instead, the author is left with setting and characters.
Traditional authors will balk at this, but as the field matures, I think a new generation of authors will find their voice in creating rich, interactive worlds.
- Send/receive mail using POP3S/SMTP-TLS, via SnapperMail.
- SSH via Mocha Pocket Telnet.
- Play MP3s stored on my SD card, via pTunes.
- Use it as a wireless modem for my laptop, via PdaNet.
- Play various time-wasters from PopCap.
PdaNet in particular is killer if you have a laptop handy, but no internet access. I get about 144k down via Sprint PCS, which isn't screaming, but it's adequate for browsing and downloading small files.how shallow the plot lines and characters are of many modern games
I suspect this will become a larger problem in the future. The problem of lifeless characters can be solved, somewhat, by improved character animation techniques. But without fundamentally changing the way characters interact with their world, you're just going to end up with hyper-realistic quest-dispensing vending machines.
Bulking up backstory and character dialogue is a short-term patch, but ultimately characters will need to become as emotive as they are well-rendered. And they'll need to interact with the world's social landscape as effortlessly as they interact with the world's collision meshes.
Scott McCloud discusses this at length in Understanding Comics. As a rendering of a person becomes more stylized and abstract, the viewer begins to fill in details on their own. Not only can our minds fill in more details than an artist can draw, but the content we fill the drawing with is our own, which makes the character more accessible.
A similar effect occurs with The Sims. Their reductionist design and behavior allows users to ascribe all sorts of baroque narratives to their simple actions.
Maybe the bumblebee hunter will get some quest that is comparable somehow, but he will still want to do what his friend did.
My thinking was that quests would be dynamically generated. So there would be the implicit assumption that you'd never get the same quest as someone else.
However, that does bring up an interesting point, which is grouping. It would be nice if emergent stories could be tailored to the skill sets of everyone in the group. Maybe our hypothetical bee-hunter has grown to be immune to certain types of toxins. Perhaps, rarely, the game could throw in a key guarded by giant wasps or something.
This technique is used from hell to breakfast in stories, where a character might be revealed to have some vital piece of information, or an unexpected skill, or connections in the right places. You know the drill:
Uber-skilled Death Machine: Rushing the gate is impossible. We'll never get in the castle.
Noob: My cousin works in the stables. He could sneak us in.
I tend to thing MM games are going to be the killer app for interactive storytelling. If you think about it, there are really only three ways to do interactive content for MM games:
1. Create a fixed set of non-interactive content. Everyone in the game world does the same quests, reads the same dialog, fights the same big bad guy for the same reason. Alternately, a few people get to participate interactively, while the vast majority simply watch (or show up for the epic battle).
2. Employ humans to create custom content for each player. Figure out a way to do this for under $20 per player per month.
3. Programmatically create custom content for each player.
There's actually already quite a bit of work going on in this area. The most interesting approach I'm familiar with is Padraig Cunningham's work with case-base systems. In a nutshell, Cunningham proposes borrowing plot recipes from the russian structuralists. One then plugs existing NPCs into the appropriate roles (wise old man, bearer of gifts, etc). The system scales by allowing one NPC to play multiple roles in multiple simultaneous stories, so long as the roles don't conflict.
So, not unlike real life, your view of an NPC (as a dastardly villain, unexpected benefactor, or innocent in distress) depends on what you're doing when you meet that person.
For more information on Cunningham's work, check out his publications. In particular, the paper entitled "A Multiplayer Case Based Story Engine."
Personally, I'm working on a system for generating plot-rich city histories (via personality and relationship modeling). My goal is to get to a point where one could plop down a new town, set some parameters, and "age" the town n years (and get sensible personal relationships, family trees, interesting local history, street layouts, etc).
This theoretical programmer was mistaken. First off, each screen was created by artists, true, but it was then rendered by computer.
My example is handily defeated; Myst was a poor choice. However, the analogy works with countless SCUMM-style 2D adventure games. I suspect you grasp my point (that this is a matter of a paradigm shift), which makes arguing the analogy sort of pointless.
This involves concepts, ideas, a whole host of fundamental knowledge of "how the world works".
Disagree. Again with the 3D graphics example. Apples are tasty, can be cut into wedges, and rot if they're left out in the sun too long. However, if you're trying to present an exciting chase scene, in which an apple cart gets knocked over, a 3D engine can abstract apples down to a model, texture map, and collision mesh.
I think we can establish similar abstractions for things we might want to present in the course of a story. Can I enumerate the optimal model? Nope. I can point you to people who have tried (Crawford, Bringsjord and Ferucci, Mateus and Stern, Cunningham, various russian structuralists and folktale analysts, etc), but no one has discovered the magic recipe.
Regarding the apple problem, we're obviously dealing with a lot more complexity here. Most toddlers have a workable mental model for how apples fall, but most adults don't understand why people behave the way they do. However, while the problem is more complex, I don't think it's intractable. And it probably doesn't require everyone to have a copy of Cyc on their desktop.
The problem with this is that it's such a big job. If you're going to provide that kind of freedom of action, you're looking at writing thousands of similar but separate stories. And no cheating like many games do, where no matter which branch you take, the impact is slight at best and the story remains mostly unchanged. We're hoping for replay factor here. We want total changes in the storyline. Some cheating this way is okay, because things happen that way sometimes through no act of your own, but still, too much of this and you lose the point of doing it on a large scale in the first place.
I anticipate that, at some point in the playtesting of the original Myst, someone probably said "It would be awesome if I could just walk around, and look at the puzzle from any direction I wanted."
And some patient programmer no doubt explained that such a thing would be impossible. Since each screen was hand-drawn by artists, you'd need thousands of artists spending years to draw pictures of a given room from every possible vantage point. And you'd need to ship gigabytes of image data with the game.
Of course, we now take that level of immersion for granted. And there aren't thousands of starving illustrators locked away in dungeons, drawing views for Doom III. Different views are created in real time by specialized hardware, using complex maths. When you zip through a room in Far Cry, how many thousands of paths could you have taken?
I tend to think the interactive "storytelling" problem will, given time, shake out the same way. We will stop working with stories as individually hand-crafted products (like paintings). We will instead begin to work with them as processes (like renders of a 3d scene).
This will require a lot of work to define the theoretical underpinnings of this process, and to create the tools people will use to create these potential-rich environments. It will also require a fundamental shift in the way authors/creators work. This will be hard. It will be resisted, both by technical types (who view the problem as insoluble) and by creative types (who view the loss of authorial control as anathema).
But it will happen, because the promise is so compelling. Perhaps because the experience of getting wrapped up in a computer game is much the same as getting caught up in a good book.
A story is a meaning applied to events after they have occured. A game is a game, like sports or a board game. You can only make a story out of it after events have been completed.
I will never understand why this old chestnut appears every time there's a discussion of interactive storytelling.
By your definition, fiction is impossible. When the author sits down with a blank sheet of paper, he should be stuck, since there are no past events for him to relate.
But of course, we know this isn't the case. Even school children are capable of inventing fanciful, novel stories. The path to interactive storytelling is collaboration between the player and the computer to produce a narrative which is both interesting to the player and dramatically compelling. The narrative is a product of this process, it is not the process itself.
Perhaps we need to come up with a new term. When people say "interactive story" they obviously don't mean "give me a book I can shove into my Playstation." They mean "give me a piece of software that allows me to have an experience similar to reading a good book, but also provides me agency, and allows me to control how I experience events."
- Checks to see if the current time is outside working hours (or during lunch). If so it..
- Strips out HTML tags from HTML-only messages
- Removes attachments
- Removes MS Exchange LDAP nastiness from From: addresses
- Re-sends the email to my mobile phone address
The app gets called with a command line parameter pointing to a file containing the message. So there's not much magic required to write your own extensions.Why, yes I do. So do other people:
I recall much wailing and gnashing of teeth when my group of friends pooled our money to buy a copy.
Even if CAPPS is only connected to ticketing and passenger information, a bug could result in a pretty nasty transportation snarl. Suppose airlines are unable to issue boarding passes for an hour, or an unusually large number of people were flagged for screening.
For any of these total-information-awareness type systems, one has to ask "what happens when some part of the patchwork breaks?" Even the most diehard "I have nothing to hide from my government" type understands that multi-hour flight delays are bad.
My box:
CPU: AMD Athlon(tm) XP/MP/4 1532 MHz
GPU: NVIDIA GeForce4 Ti 4200
Score: 1294
Now, if I go to the Futuremark Project Search, enter my exact system, and change out my GF4 for a Radeon 9700, the top results are in the 4500 range. Which would suggest that I could get a 200% performance boost by buying an ATI card. However, game benchmarks suggest the 9700 would buy me a much smaller gain (5-50%, depending on the game/resolution/options). IMO, 3dmark places a bit too much emphasis on the newest DX9 features to be useful as a general performance benchmark.
So, like Magic Carpet for MSDOS.
No doubt about that. But the current powerful oil companies would not be very excited about that unless they could ensure that THEY would be the powerful hydrogen companies as well.
Considering their expertise in seismic analysis, drilling, and natural gas processing, I think that's pretty much a given.
- The program will attempt to guess what reply number you're on in an email exchange, and replace the Re: in the subject line with Re[3]:. IMO, this is mostly meaningless clutter.
- The Bat! will not append quoting characters to blank lines in quoted/forwarded text. Sounds great, until you want to forward a multi-paragraph article or code block.
- Various other odd little anomalies.
Now, I still think it's the best option for mail out there, but I would kill for fifteen minutes alone with the source code, so I could correct some of the irritating little features.I'm not just pointing the finger here, I went through this same thing two years ago with Verio. I was paying $300/mo. for "business" DSL, and had to deal with constant downtime, awful/slow tech support, and a billing office that kept reminding me that my contract included "no specific uptime guarantee" (but assured me I wouldn't be charged for the days of downtime I kept experiencing).
After my contract was up, I switched providers, moved to ISDN, and signed a contract with a fixed uptime guarantee. The few outtages I had after that were corrected within the day.
If you're interested in this sort of thing, you'd can get a good feel for the existing work in the field from:
- InteractiveStory.net - Michael Mateas and Andrew Stern's interactive drama/believable agent project, and obligatory huge page o'links.
- Oz - The Oz project at CMU
- Erasmatron@Robotwisdom - Jorn Barger's excellent thumbnail sketch of Crawford's writings. In most cases, Jorn's synopsis is hyperlinked to the related page on erazmatazz
I'd also recommend the following papers (try CiteSeer before heading down to the library):Selmer Bringsjord and David Ferrucci, Artificial Intelligence and Literary Creativity: Inside the Mind of Brutus, A Storytelling Machine, August 30, 1999.
Nicolas Szilas, Interactive Drama on Computer: Beyond Linear Narrative, 1999.
Antonio Furtado, Angelo Ciarlini, Plots of Narratives Over Temporal Databases, 1997.
Barbara Hayes-Roth, Robert van Gent, Story-marking with improvisational puppets, 1997.
W. Scott Neal Reilly, A methodology for building believable social agents, 1997.
IMHO, interactive storytelling is one of the most interesting cross-discipline computational problems out there.
If this could be made to work reliably, consider what it would do for sites like Penny Arcade or Something Awful. Sites like these are expensive to run because of the cost of bandwidth. If you decentralize the distribution, you could spread the bandwidth cost between all the users. Since most users pay a flat rate for their internet connection, in most cases the distributed cost would be $0. And since he's no longer paying for hosting, Lowtax could afford a better haircut.
Yahoo stands for Yet Another Hierarchical Officious Oracle. Remarkably, prior to becoming a household term, Yahoo actually had to have a descriptive name for people to know what the hell the site did.