Alexandra Freeman
 |
|
The Big Al website
Whilst I was a graduate student I had learnt html and basic web authoring skills, and with my partner John Alden had set up web pages for various organisations, and had taught html and web authoring to undergraduate classes. As soon as I started work on the 'Walking with...' programmes I found myself automatically thinking about the possibilities of the accompanying website. Whilst I was working mainly on Walking with Beasts, I also helped reconstruct behaviour for the dinosaurs in the Christmas special 'The Ballad of Big Al' and the accompanying 'Science Of' programme.
I immediately had several ideas for features of the website, including a database of questions and answers which would help viewers instantly find the answer to the perennial question 'How do they know that?' and several games. These games would not only be fun to play, but would also have an educational basis, demonstrating how palaeontologists interpret fossils and basic biological principles.
The producer of 'Big Al', Tim Haines, introduced me to the Editor of the Science website team, Carmen Pryce, and we worked together to produce a treatment for commission. Unfortunately, the programme was only a 30 minute 'special' and so it was decided that only a small budget would be allocated to its accompanying website.
The Big Al Game
However, on a couple of long train journeys, my partner John Alden and I had already put together an almost complete version of the 'Big Al game', in which players can try their hand at being the star of the programme, an Allosaurus, and move around a map (representing the habitat in which the programme was set), trying to eat the right food - and avoid being eaten. As the player eats, their dinosaur grows, and moves through different levels as it does so. At the beginning of each level, the player has to make a palaeontological decision like those I had to make as researcher for the programme, such as 'does a mother Allosaurus stay with her young?' The decision the player makes controls what happens during that level. Points are awarded for surviving, and in the final level, for mating successfully! I actually decided to add a little extra challenge for any players who managed more than three successful matings - not that I imagined many people would! On the fourth mating attempt, after successful courtship, the female Allosaurus asks a multiple choice palaeontological question!
Since the game was already almost ready for release, and seemed popular with the production team at least, it was incorporated into the Walking with Dinosaurs website. It turned out to be immensely popular! Some players on the high score table after the first month of release had made over a thousand moves, and it generated half a million hits per week, even 6 months after release. In fact, John decided voluntarily to rewrite the game to run under mod perl (rather than perl) in order to make it more efficient and reduce the load it placed on the BBC web servers.
Play the Big Al game here.
I also produced a spreadsheet of 100 or so questions and answers about dinosaurs, fossils, and Big Al in particular for the website. Unfortunately there was not the budget available to turn these into a database which could be queried and searched as I'd hoped, but many of them were grouped into a series of plain web pages for the site. Although we had not been able to implement many of my initial ideas for Big Al, Carmen and I were very hopeful of being able to use some of them to produce a large website for the second series Walking with Beasts, working with Assistant Producer Duncan Thomson, who had worked on the original Walking with Dinosaurs site.
Learning with Beasts
During the spring and summer of 2000, after the release of the Big Al game and site, I started to develop my ideas for the Walking with Beasts website. I had had the idea that children's interest in dinosaurs - especially their vital statistics - could actually be harnessed to help them learn about maths. The problem that I had with maths at school was that it seemed 'pointless', and yet when I found a subject I was interested in (biology), I discovered that maths was actually a tool - a way of finding the answers to questions that really did interest you. So I thought that by posing questions that children clearly were interested in, and that were real (rather than the very contrived questions I was familiar with from textbooks), then I might be able to inspire children to want to learn maths. In fact, by looking through KeyStage textbooks I found that palaeontology could, with a little imagination, be used as the inspiration for almost the whole of KeyStage 3 maths: Calculating the angle that a sabre-tooth cat has to open its mouth to bite, calculating the length of an Andrewsarchus by scaling up from its skull in the right proportions etc.
This opened up the possibility of a whole new area of the Walking with Beasts website - one devoted to education. Again, Executive Producer Tim Haines was immensely supportive and we approached the Education Department for funding to develop this area of the site. There, they were just developing the idea of 'Learning Journeys' - using the web to produce a linear series of pages taking the user through a number of steps, at the end of which they would have a deeper understanding of a particular subject. So, using this idea, I developed outlines for a number of Learning Journeys which could accompany Walking with Beasts: on evolution, the extinction of the dinosaurs, the ice ages, and 'be a palaeontologist' - a step-by-step guide to how to reconstruct animals. For each of these sections I designed at least one 'Online Interactive Experiment' (or OLIE). I wanted the users to be able to try things out for themselves, as they had done in playing the Big Al game, and see the consequences. Experimentation seemed the most fun (and hence the best) way to learn! On this basis, I finally secured funding to develop these areas of the website, although later down the line the Ice Ages section and its OLIE were dropped.
I was very keen that the 'educational' areas of the website would not be branded as such in any way (nothing puts me off more than thinking something is deemed to be 'good for me'!), and that the OLIEs would be embedded in pages surrounded by light text, leading the users on through the site. It would work in different ways for each section.
What Killed the Dinosaurs?
For 'What Killed the Dinosaurs' I was lucky to be able to draw on the expertise of my co-researcher on Beasts, Paul Chambers, who had studied the evidence for the extinction for his PhD thesis and had collected a great deal of evidence together which had not previously been published in any way. In fact the evidence had some startling lessons to teach for children just starting out in science. For example, all the data for land animals and freshwater animals comes only from one small site in the USA. So, is it really accurate to say '80% of a group became extinct' when that is based on such a small sample? These are ideal questions for classroom discussion, and on a topic which fascinates everyone! So, I structured the whole site as an online experiment:
Firstly the user is given the basic background and is posed the question. Then they have to choose their first line of enquiry - they have three possible sources of evidence. They can examine the geochemistry of rocks (in which case they are taken to a page explaining the results of their analysis and helping them interpret it). They can search the world for volcanoes and craters dating from the extinction (and are taken to a page with some candidates and interpretation of their findings). Finally they can mount a fossil-hunting expedition. In order to do this they are given a map of the world and a list of the kinds of animal and plant fossils they might look for. As they move their mouse over these options, the world map shows 'X's over the spots where such fossils are found. This gives the user an idea of how representative a sample each kind of fossil is likely to produce. Finally, when they click on a fossil type, they are taken to a page giving the results of their expedition and help with interpretation of these. At the bottom of the page they can click to add this data to their 'rucksack' and as they collect the data from different fossil types this 'rucksack' page gets longer, showing the graphs of all the data they have collected. This allows them to compare the effects of the extinction on the different groups.
Once the user had explored two pieces of evidence, then a new option would become highlighted - 'Theories'. This section outlines the major theories for why the dinosaurs became extinct (such as the meteor impact, global climate change etc), and then works out what evidence each explanation would have left. The users are then encouraged to compare the evidence they had found in part one (collected in their 'rucksack') with the predictions of each theory. Once two theories had been explored, the final part would become highlighted - 'Vote and Discuss'. This section allows the users to vote on the theory they thought best fitted with the evidence they had collected, and could then join in an online discussion on the subject. The fact that they could only reach the discussion by passing through at least two 'evidence' and two 'theories' sites would hopefully mean that the discussion was 'informed', and any disagreement over the facts could easily be checked by a quick return to the evidence sections.
Together with my partner John Alden, I put together a working 'What Killed the Dinosaurs' site over Christmas 2000, only missing some graphics which would need to be specially drawn. This was finally reversioned and released onto the Walking with Beasts site just before launch in November 2001. For release, there was a lot of discussion over whether the block on moving from 'evidence' to 'theories' to 'vote' should be kept, and eventually it was decided that this would be too much of a discouragement and so the current site is freely available without this feature.
Evolution
Evolution is a subject which is very misunderstood - a fact which was brought home to me especially hard when I was involved in the 'live chat' after the transmission of The Ballad of Big Al. Since the concept of evolution underlies the whole series of Walking with Beasts I felt that this was an excellent opportunity to construct a Learning Journey with interactive elements which would help people understand it and the simple underlying concept of natural selection.
One way in which I had always explained the inevitability of natural selection (under certain conditions) was to get people to imagine a bag of differently coloured sweets. If I eat all the red and black ones (as I always do!), and leave the yellow and green ones, then soon the bag will consist of almost all yellow and green ones. Now, imagine that the sweets had babies. We know that in the real world children tend to look like their parents (they 'inherit' characteristics), so imagine that the baby sweets were the same colour as their parents. It is now clear that inevitably, as long as I keep eating more red and black ones than green and yellow ones, that there will always be more yellow and green ones in the bag. Depending on how fast I eat them and how fast they reproduce, I might even cause the red and black ones to disappear.
Using this as a basis, I designed a simple OLIE in which the user could 'shoot' coloured sweets (or fish, as we eventually changed them to) which moved around the screen whilst a clock counted down. This little game would be embedded in the page of text about evolution, and would have three levels.
In the first level, the sweets appeared in a range of colours, on a plain coloured background. The sweets would 'reproduce' (visually, by splitting into two), and the baby sweets would be identical in colour to their parents. Instructions would encourage the users to attempt to shoot sweets of a particular colour. I hoped that the coloured background would help by making those most similar to that colour more difficult to see and so cause all player (whether they were trying to or not) have some bias in their shooting. When the clock reached 0, the user would be given a score (number of sweets shot) and the difference between the average colour of the sweets before the game started and after their shooting spree. This would show the effect of their selection, demonstrating how natural selection occurs in the real world. There would then be a 'next level' button which would load the next page of text with the game embedded in it again.
In the second level, the sweets would 'reproduce', but this time the baby sweets would not quite be identical in colour to their parents. Instead they could be very slightly different, as we see much more commonly in the real world (the text on the page would offer links to pages explaining the underlying principles of inheritance and reproduction). In this game, the population of sweets could eventually become much closer to the colour of the background than any of the original parent sweets were - eventually making them almost impossible to see and shoot. Again, at the end of the game a summary would be given, and text would explain that the selection has cause the population to evolve.
Finally, in the third level, the sweets behave as they did in level 2, but now the background colour changes constantly. This makes some sweets easy to see one minute and hard the next. So it is much more difficult to keep up the same selection criterion. Again a summary is given, and text explains that this represents the constantly changing environment. Things do not simply evolve to a 'perfect state' and then stop, because the world around them (especially the other animals and plants) is constantly changing.
So, I hoped that this series of text pages, and their embedded interactive 'game' would help people understand the simplicity of the concept of natural selection and hence evolution. With one eye on the National Curriculum, though, it was clear that schools concentrate more on the historical story of the 'Darwin's theory of evolution' (which is actually incorrect, as Darwin came up with the theory of natural selection to explain evolution, which had already been observed and accepted). Therefore I created a series of accompanying pages explaining the historical story. My plan was that the first page of the 'Learning Journey' would provide a very simple paragraph's introduction to evolution and natural selection and options to follow either the 'interactive route' or the 'historical route'. There would be opportunities to link across from one to the other, and also to a few extra 'in depth' pages about the underlying genetics which is not necessary to understand the basic principles of the subject but provide extra information (which is also relevant to the national curriculum).
The interactive demonstration of natural selection provided couple of challenges. It had to be programmed in Flash, and my partner John Alden again provided the underlying programming. The animation work was done by animator Matt West whom we hired for several months to work on the Beasts site. There were difficulties in making the fish (which had replaced my initial idea of sweets) appear to swim in a random pattern, whilst actually using animation loops in order to coordinate their tail movements. The next challenge was to cause them to reproduce realistically, and for a 'new fish' to appear just in the right place and at the right time to coincide with the end of the 'breeding cycle' in the animation. Finally, there was a challenge in how to represent the population of fish at the beginning and end of each level. An 'average' (i.e. mean) fish colour could be very unrepresentative (since the mathematical average of two colours can be a third, completely different colour). We struggled for some time to come up with a solution to this problem, attempting to represent the range in many different ways. Since colour is multidimensional (it has three colour components plus a lightness to darkness component) it would be impossible to represent visually as a graph. In order to minimise this problem we decided to reduce the range of colours possible, restricting it to two colour components plus a lightness component. This helped us represent the change over the game, although further work is probably necessary to perfect it.
The Evolution site is online here.
Be a Palaeontologist
Every child (and many an adult!) seems to want my job - the detective work of building up a picture of an extinct animal from the evidence. I wanted to be able to harness this interest and allow them to try out some of the techniques themselves, on the web. I split the section into three parts:
How do we know what they looked like? Many people seem to imagine that fossils come out of the ground looking like the one in Jurassic Park - just blow away the sand and there are all the bones, layed out in the shape of the animal! Unfortunately this is very rarely the case. A couple of paragraphs of text would explain that putting together tiny fragments of fossilised bone was rather like doing a 3 dimensional jigsaw with no picture, most of the pieces missing, and possibly the odd extra piece from another jigsaw! Embedded in this page would be a 'skeletal jigsaw' - a chance for users to try this for themselves. They would be presented with a pile of bones, and they had to put them together to form a complete skeleton - against the clock. I designed several levels, suitable for different ages and levels of skill. In the simplest there was an outline of the animal, the few sections of skeleton were very large, and there was the option to show the outline of the skeleton for very young children. In the middle level the skeleton was split into smaller sections, and in the top level the bones were essentially completely independent. Then (with an anatomist friend of mine in mind!), there was the 'superchallenge' - the individual bones of two animals mixed together, and no outline to help you! Anyone who can do it in under 15 minutes is doing extremely well! Once the skeleton had been put together successfully, the animal 'came to life' and walked off the screen.
I chose four skeletons of contrasting animals for the game: an early whale, a giant bird, a mammoth, and an australopithecine (a bipedal ape). These, I hoped, would introduce children to the similarities between the skeletons of animals which looked very different from the outside, and which could be used in classes teaching the National Curriculum. In order to add to the educational side of the jigsaw, I then added a feature where each bone (or section of skeleton) could be dragged over a button in order to pop up the name of the bone. This could be used as a clue to help people build the skeletons, and be a fun way to learn the names of bones necessary for the National Curriculum. This jigsaw was very well received by all the advisors I sent it to, many of whom planned to use it in their undergraduate teaching!
The skeletal jigsaw presented a few technical problems as well. Firstly, file size. I was extremely concerned that the site should be as lightweight as possible in order to increase accessibility. I was also very much against reliance on plug-ins such as Flash. However, like the natural selection experiment, this jigsaw could only be made interactive using Flash. In fact, Flash 5 produced a relatively small file size (we used streaming to allow people to start assembling their first animal whilst the others downloaded), however, I wanted to be able to provide a Flash 4 version, and this produced a ridiculously large file size. So, we cut the jigsaw up into separate files for each level. You can try these versions here. I also wanted to be able to provide a version for those without the Flash plug-in, and in fact the easiest way of doing this seemed to be to provide the skeletons as simple images which could be printed out, cut into suitably sized pieces and used as a real jigsaw. I suggested that it could be made more realistic by hiding the pieces around a room, and perhaps even destroying a few! I was very keen to encourage 'offline' activities as well as people sitting at their computer (and probably paying a phone bill!) You can see these here.
How do we know how they moved? The second major part of my job, once we've worked with the model makers to build the animals, is to work with the animators to make them move. Part of the evidence that helps us do this comes from fossilised footprints. The idea for the interactive experiment here came to me when I was working on the Ballad of Big Al. There is a famous set of fossilised footprints (called the Paluxy River trackway) which seems to represent a dinosaur hunt. Initially, looking at this trackway it is difficult to interpret, but my idea was that it should be possible to make the user's cursor turn into a dinosaur (either the predator or the prey), and for it to leave appropriate footprints as it moved across the screen (calculated by the speed and direction of the cursor's movement). By making one half of the screen a representation of the fossilised trackway and allowing the user to experiment by moving their cursor over the second half to try to replicate the tracks I hoped that people would get a feeling for how researchers have been able to interpret what happened. I also wanted to be able to provide a 'toolkit' (including a 'ruler' which measured the distance between two mouse clicks, and a calculator) along with some instructions and the equations which allow researchers to calculate the speed an animal was moving from its footprints. The calculator could have different 'levels' - from calculating the speed automatically given the measurement from one footprint to another, to doing none of the calculation automatically, making the user do all the data entry in the right order. This would make it useful for maths teaching.
The footprints experiment was technically difficult. Only Flash 5 had the mathematical capabilities to make the necessary calculations. We decided to get an external company to produce it, and Duncan Thomson took responsibility to adapt the idea for Walking with Beasts. You can try it (if you have Flash 5) here. My accompanying text is here.
How do we know what colour they were? This is one of the most common questions we are asked. It is also a subject very close to my heart! I remember being set a biology essay on the different uses of colour in animals at school, and my teacher explaining afterwards the principle of 'counter coloration' to me. It was such a great concept that all I could say was 'cool!' - and my teacher replied 'That's what I said when somebody told me'. Those little lightbulb moments are so satisfying that I wanted to be able to pass that on through a simple interactive demonstration.
Animals which use camouflage (either to hide from their predators, or from their prey) have several ways in which they do it. Firstly, they try to match their basic colour to the most common colour in their environment. Secondly, they try to break up their obvious outline with dark markings which mimic the shadows in their environment. Finally, they use counter coloration: when a basically cylindrical body is lit by sunlight from above, their belly is always in shadow, and so looks darker than their back. In order to counter this effect, animals usually have paler bellies, which therefore makes them appear to be a uniform colour when standing in normal light.
I wanted to create an online experiment where people could try out different colours, markings and shading patterns on animals in different habitats to see the effects for themselves. This could be achieved easily without the use of a Flash plugin by simply creating a grid of web pages showing all the possible combinations. If it were also programmed in Flash, it would be possible to add a 'test your camouflage' button showing an animation of a stalk in which your predator or prey animal was put to the test! Unfortunately there was not time to complete this! Try it for yourself here.
Finally, I wanted to link from this Learning Journey to a page giving ways in which users could take their interest further. This ranged from the 'museum search' (more details here) to short courses with the Open University, and full undergraduate courses at UK universities. Only a few universities run courses in palaeontology (it is mostly taught as part of a geology course), and so I wanted to list those that did with links to their course websites. I also managed (though at rather short notice) to make contact with the Open University (thanks to Education Executive Paul Manners in Bristol), and make a link to their 'history of life' short course. In the future (when the third part of the 'Walking with..' trilogy is produced) we both hoped that it might be possible to tailor the course more closely to the trilogy and use some of the footage to enhance the course material - but that involves negotiations with BBC Worldwide and long term planning.
When 'Be a Palaeontologist' was finally developed it was decided that the title was confusing - that many people may not know what a 'palaeontologist' was, and so it was renamed 'Fossil Fun'. Duncan Thomson (who is a geology graduate) also added sections on the formation of fossils, using Flash. The Fossil Fun site is online here.
Material directly supporting the series
Alongside the extra educational material, there were of course several areas of the site which would be directly relevant to the television series. I wanted people who had thought 'how do they know that?', or 'how did they do that?' to be able to come to the website and have their question answered directly.
'The Making Of' & 'Evidence'
In parallel with the interactive TV series (more details of which are available here, we planned to have sections of the website called 'making of' and 'evidence' - two of the alternative streams available on the interactive TV series. The 'making of' section was simply a collection of pages (illustrated with photos) describing the production process. I was hoping to collect little biographies of people who worked on the series in different capacities to help those who were interested in working in one of the different areas know what kind of background and experience they would need. This is another frequently asked question! I also hoped that it would introduce people to some of the lesser known jobs, such as production management which rarely feature in 'making of' documentaries or websites. Many people never consider the problems inherent in flying lifesize mammoths made with real animal hair around the world.
The 'evidence' section I hoped would be released sequentially, along with transmission of the different episodes. For each episode I wrote a paragraph of information about the environment, and each of the Beasts featured, describing the evidence we used to reconstruct how they looked and behaved. I then wrote a series of questions and answers which I felt covered just about everything that anyone might want to know about the series! A list of these questions would be available at the bottom of the 'evidence' paragraph for each animal, but I also wanted to make sure that they were easily and directly accessible to the viewers. So, having created the questions and answers in a spreadsheet I included a column for 'keywords', in which for each question I entered every word I could think of which people might enter in a free text query if they were trying to ask that question. This, I hoped, would allow the very simple query engine written by the software engineering team to give the effect of a free text search. The engine would take each word entered by the user and compare it with the 'keywords' column, and then return the 5 questions which had the highest number of matches. For actual questions and answers we could use dhtml to list the questions, but not reveal the answers underneath them unless a question was clicked on. This would allow the user to scan the questions, but also not wait to load a new (graphics heavy) page with the answer. For browsers with no dhtml support (and printing), the answers would simply appear underneath each question. One further option I was keen to include was links to short video extracts from the many interviews with scientists we did for the interactive TV version (which were only available to those with digital satellite).
The 'evidence' section of the Beasts website is here.
'Fact Files', 'Mammals' Family Tree' and 'The Changing World'
There would then be a level of information which was slightly deeper - that would give people answers to questions they may not even have thought of. There were to be Fact Files for each Beast (as on the Walking with Dinosaurs site), and I also wanted to have an interactive Family Tree of mammals to illustrate how the different Beasts were related to each other and to modern animals - something that we don't yet have for dinosaurs. In fact, towards the end of our research on the series, some spectacular new research on the relationships between different living mammals was published. This meant that our tree is probably still one of the only accurate representations of this new view available to the public, on the web. The first page of the Family Tree gave an outline diagram showing how the main 6 groups of mammals (monotremes, marsupials and the 4 placental groups) were related to each other and when they split. Underneath the diagram was a brief paragraph giving more details on the groups, and I was keen to use dynamic html so that the paragraph changed as the user rolled over each branch of the tree in order to be relevant to that particular group. For those using browsers not capable of displaying dhtml, the paragraphs would simply be listed underneath each other below the diagram, which would be a suitable alternative. The user could then click on a branch to be taken to a more detailed view of the relationships within that group. Again there would be dhtml paragraphs underneath the diagram, and in some cases the possibility to click and go into more detail again. The Family Tree is here.
In order to complement the Family Tree I designed a section called 'The Changing World' which took users through the different time periods portrayed in the series and showed what the world would have looked like and what was happening to the climate and the animals at that time. There would be a map of the world at the time, and alongside it several buttons which would allow users to add maps of the different types of vegetation at that time, the deep and shallow sea currents, and the movements of different animal groups. Underneath the map would be text describing the world at that time, which would change as users clicked on the buttons for each of the different options. When the button showing animal migrations was selected, the map would act as an imagemap, and moving the mouse over the different continents would bring up text describing the evolution of the different mammal groups on each. I was also keen that there should be the option of viewing this section as a Flash animation, where dragging the mouse over a timeline would show the continents move. This Flash version is available here.
I wanted these three major sections (Fact Files, Mammals' Family Tree, and The Changing World) to be closely linked together. Each Fact File would carry a link to the relevant part of the Family Tree (and the 'evidence' section described above), and both the Family Tree and The Changing World would carry links back to the Fact Files in appropriate places. Together these three would form a major resource, for people at all levels of knowledge, and (especially given the new information on the relationships between the different mammalian groups) one which was unique on the internet.
Just for fun...
There were a few extra features of the site which were primarily designed for fun (although I always find they end up being at least a little educational because I always base the games in some kind of reality).
The Beasts Game
Even before we had finished working on the Big Al game and discovered its success I had started thinking about a new version for Beasts. Technically there were a few things that both John and I wanted to improve on. Firstly, although it was fundamentally accessible to everyone since it is simply built out of a series of webpages (without using any plug-ins) this did make it slow to play over a normal modem connection (something that it is easy to forget when developing inside the BBC with its internal server system). We had built the pages to be as graphics light as possible, and John had used many tricks to optimise this, but we decided that any future games should have a text-only interface as well. A third option, a Flash interface, would also decrease the download problem, but would inevitably take a lot more development time and it was not something we could do ourselves.
I decided to base the Beasts game on one of the accompanying 'science of' documentaries following the evolution of primates. I was very disturbed to read peoples' misunderstandings of evolution on forum websites with comments such as 'I don't understand why, if humans evolved from chimpanzees, there are still chimpanzees around which haven't evolved into humans'. I decided that I would build a game in which the player started 50 million years ago in the Eocene forests featured in our first programme, as one of the earliest known primates. As in the Big Al game they could move around a map, avoiding predators and finding things to eat. However, unlike the Big Al game, they would move through time much faster. Whenever they mated successfully, time would jump forwards and they would play one of their successful descendents. Whatever the user did would be stored and contribute to calculations to determine how sensitive each of their senses were, what foods their digestive system was best adapted to, what habitat they were best adapted to, and what their descendents looked like. The tuning of their senses determined over how many squares' range they could detect each of the other animals in their habitat, and I spent many months determining an underlying family tree of 71 different fossil (and eventually living) primate species. As the player moved through time their path along this family tree would be determined by their actions until eventually they ended up in the present day as one of 21 possible living primate species (possibly trying to avoid being shot by humans!) Throughout these 50 million years, the player could encounter any of these 71 primates or any of 120 other animal and plant species which I entered into the database. I also wrote fact files for all these 191, and had to draw pictures for many of them! I hoped that this richer environment and potentially longer gameplay would satisfy many of the fans of Big Al who obviously played the game more than I had expected! Because there were 21 different endpoints, I also designed a 'high scores table' which allowed people to store the different endpoints they had reached, and the table ranked those who had reached the most.
The game engine is rather complicated, with a great number of variables which need tuning for optimum playability. I think that the game as it is currently is not optimised, and that it is too difficult to reach an endpoint. Unfortunately the tuning requires some dedicated players and either me or John Alden to make necessary adjustments. You can try it for yourself here! Site statistics, though, show that the game is very popular, and that an increasing percentage of hits to the Beasts site are to the game.
Multiplayer Beasts Game
One thing I was keen to add to the Beasts game was the possibility of playing in 'multiplayer mode'. Life is, after all, a multiplayer game! It would, I hoped, increase the level of competition for food etc. in the game, and hence encourage players to diverge - to move to habitats less populated by other player (and hence cause their descendents to adapt in different ways). This would mimic the real world to some degree. It would also add an interesting new aspect to the game. Multiplayer capabilites require the useof different technologies, and once I had thought through the gameplay implications, I went to see George Auckland and worked with him and Will Green to look into the possible ways of achieving it. Unfortunately the current BBC computer infrastructure made it impossible to develop a multiplayer version in the short term, but this is definitely something I would like to experiment with in the future.
'Build-a-Beast' & 'Battle of the Beasts'
Duncan Thomson suggested that there should also be a game which was less 'biological', and suggested 'Battle of the Beasts' in which players could design a Beast (made out of the body parts of the Beasts in the series), and then fight other homemade Beasts. The first part of this ('Build a Beast' as we called it) matched with an idea I had been mulling for some time about the possibility of mimicking the children's flip-books where each page is divided into three, and you can match the heads, bodies and legs of different characters in many combinations. This effect was possible with many of the Beasts given a little bit of javascript, making it a neat little game with wide accessibility. The fight was slightly more complicated, but I had the idea of using something a friend had been working on in the first year of our doctorates. It may sound like a bit of an overly complex way to structure a game engine for a relatively simple game, but I think that underlying complexity means that the behaviour of the game is less predictable and therefore makes the game much more interesting to replay over and over again.
When one animal meets another, there are a number of possible outcomes of the encounter, depending on how each behaves. This is true in the real world, and in the world of games, and those who study this use Game Theory to predict and model what goes on. The most famous 'game' used in Game Theory is one called 'Prisoners' Dilemma'. The story is that two criminals have been arrested for a crime and the police are questioning them in separate rooms. The police offer them each a deal. If they 'defect' (i.e. give the police evidence that the other person is guilty) then they will be rewarded and the other person will be put away for the crime. However, if they both 'defect' then they will both get locked up, and if they both 'co-operate' with each other by keeping quiet then the police have no evidence and they will both eventually be released. This situation can be represented as a game in which each player gets a different number of points depending on each of the four possible outcomes. These points are represented as a grid, called a 'payoff matrix'. An example of a payoff matrix would be:
| Player 1 | |||
| Co-operate | Defect | ||
| Player 2 | Co-operate | 3 | -1 |
| Defect | 5 | 1 | |
Now, if the game is a one-off then a player doesn't have that much choice. If he chooses to defect then he can score either 5 or 1, whilst if he co-operates then he can only score either 3 or -1. So his obvious choice would be to defect. Imagine the game is played over and over again, though. If both players carry on defecting, they will both carry on scoring 1 each on every move, but if they both started co-operating, they would start scoring 3 each. So, players who start co-operating with each other might actually be able to start building up a higher score than those who just defect all the time - but you run the risk of your opponent throwing in the occasional 'defect', gaining himself 5 points. How would you react to that? Defect yourself on the next move - or carry on co-operating in the hope of continuing your previously fruitful partnership?
Other than us, though, animals aren't conscious (certainly not many of them, anyway) - but they still show complex behaviour. It seems that that behaviour actually results from animals having a more simple 'strategy'. For example, one of the easiest to understand (and a common one), is 'tit for tat'. Basically, it means 'do whatever your opponent did last'. This turns out to be rather a good strategy in many situations. If you start off co-operating, then you get along fine. If someone defects, though, then you're in trouble. There are other common strategies, and in fact which does best depends on which other strategies everyone else is playing.
So, this was the very basic principle underlying Battle of the Beasts. Rather than choosing what they did on each move, players designed a Beast which had a strategy - a rule-book of behaviour. Then they could let it out to fight against others, and see whether their strategy worked. However, firstly I increased the number of possible moves to 6. So rather than simply 'co-operate' or 'defect' the player could choose 'wait' (neutral), 'kick' or 'bite' (two forms of defection), 'duck a kick' or 'dodge a bite' (co-operation of a sort), or 'run away'. This final option is an important addition to the list as it allows some players to be 'sensible' and recognise a partner they are not going to score highly against, and choose another without going through a punishing fight - a good strategy if you can get it right!
The second important difference was the use of a 'scratch space model'. When animals meet each other in the real world, it would of course be useful to them to be able to remember how previous encounters with individuals have gone in the past, just as it is useful for players of Prisoners' Dilemma to remember what their opponent's previous move was. Animals, though, can't just have unlimited memory - their brain is being used for other things, and so they have to optimise what they remember. The scratch space model is a way of representing how this limited memory could work.
Call the 6 possible moves in the game 'external states' - so a player could be in a 'wait state' for example. The player, though, also has an internal state. In the game I allowed the players 6 internal states. What the opponent does changes the internal state of the player, and the internal state of the player then determines what the external state of the player is. So, for example, if the opponent kicks, then the player's internal state changes (perhaps to 'kicking angry!'). This has a relationship with the player's external state - so in the simplest form if the internal state is 'kicking angry' say, then the player's next external state (move) might be 'kick'. However, the relationship doesn't need to be quite that simple - perhaps the player wants to react to more than just the last move. He wants some memory. For example there is a strategy called 'Tit for Two Tats' (or Generous Tit for Tat) which, rather than kick right back, gives its opponent a second chance to carry on mutual co-operation. So, in this case, the player would use one of the 6 internal states as a second level of anger. So, on the first kick he moves to internal state 'angry' (whose corresponding external state is still a non-defection one), but if he receives a second kick whilst in internal state 'angry' he moves to internal state 'very angry' whose external state is 'kick'.
That all sounds a bit complicated, perhaps, but the players didn't have to know about the underlying principles - although I did create an 'expert interface' for those who did and wanted fine control over their Beast's strategy. This consisted of 6 6x6 grids in which they could enter the probability of moving from one internal state to another given the opponent's previous move. Then they could control which external state resulted from each of the 6 internal states.
Normally, though, when a player had created their Beast and wanted to send it into the game, it was assigned a default strategy (one of a selection of common ones written by me). The player could then alter this strategy simply by moving 6 sliders up and down (which increased or decreased the probability of the Beast moving from each internal state to another) - they perceived this as 'increasing tendency to bite' etc.
The player then had several choices of Battle. Firstly, he could choose an opponent from a gallery (consisting of a few stock 'trainers' written by me, plus the top 10 high scorers from the public). There would be a fight screen showing the list of moves made by each (up to 10 each), and the resulting score - with one declared the winner. These preliminary battles would help the player fine tune his Beast.
A second option allowed a player to send his Beast to a friend and challenge them to a fight. This was achieved by assigning each saved Beast a serial number, and this was then sent to the opponent (rather like collecting a e-card). The result and score of each Beast was then e-mailed to the challenger.
Finally, the player could send his Beast into Battle - a weekly round-robin competition where all the submitted Beasts played off against each other. Those with the highest average score (mean score per move - so as not to disadvantage those which did not fight as many rounds) entered the high score gallery at the end of the week, and each player was e-mailed daily with the results of the fights his Beast had been involved in.
Because of the underlying complexity, it would be impossible to create a 'perfect' Beast which would always win, keeping the challenge alive each week. Success would depend on which strategies the other Beasts submitted that week played.
The underlying game engine was programmed (one weekend!) by John Alden and Yan Wong, and the Battle of the Beasts functionality was programmed by Piers Kent in the BBC software engineering team, but has yet to be released.
Battle by SMS
As soon as I had devised the game play for Battle of the Beasts I thought that a good extension of it would be to be able to play via SMS. Working with Richard Williams and Nick Aldridge, I designed a gameplay which would work through text messaging. In this, the player could send a series of 10 moves (each had a one-letter code), and these moves were played against one of a gallery of Beasts. The result would be texted back to the player. They could then choose to play that same opponent again (or choose another). This would allow them to try to 'solve' the strategy that their opponent was playing and maximise their score against it in subsequent games. Again, Piers Kent did the necessary programming, and Nick Aldridge the negotiations with the mobile phone companies. I do not know how the game has performed as yet, but until the web version of Battle of the Beasts is released it is unlikely to fulfil its potential. You can try it yourself by reading the instructions here.
Swarm Screensaver
One final idea that I had for the site was a screensaver which was a bit more interesting than the usual rotating selection of images. I thought that we might be able to combine Framestore's excellent graphics with my knowledge of animal behaviour to produce something a bit different. The herding and swarming behaviours of animals are governed by relatively simple rules which can be simulated by a computer. I worked with my partner John Alden again to determine some suitable rules to simulate the behaviour of a shoal of fish (we managed to simulate a swarm of flies too on the way to achieving this!). My idea was that we could create a screensaver which was based on programme 2 of the series (underwater). In the first week you would download the screensaver, which would consist of a shoal of fish swimming around the screen. In the second week, an extension to the screensaver would be released - a shark, which would swim around the screen and attempt to eat the fish when it encountered the shoal. The fish would, of course, react in a suitable way by dispersing and then reshoaling. On the third week you could download a small early whale, Dorudon, which would eat the shark or the fish, and in the fourth, the huge whale Basilosaurus. Apart from the fish, who would be confined to screen, the animals would swim off screen and only be occasional visitors (very occasional in the case of Basilosaurus!). Whilst John and I managed to perfect the behavioural programming, and created a test screensaver of a shoal of (rather geometric!) fish, we were unable to find an external company who could help finish it by adding the graphics for a reasonable cost. The test screensaver is popular with friends, though, and hopefully I can find a use for it on a future project when there will be more time to complete its development.
Throughout the whole website I was keen to include printable versions of the different areas, to keep page weight (and numbers of individual images) low, and minimise the reliance on plug-ins - all of which I hoped would increase the accessibility and usefulness of the site.