What it was all about...

I really enjoyed the time I spent on playing Frontier. People who never really understood what it was all about would probably say something like "Why do you play that ugly old game, my Win* Commander 3000 looks so much better". Frontier is/was about something beyond good looks. Not that Frontier looked bad at the time it was released, anyway...
For me, it was about a certain feeling, a certain atmosphere... A sense of total freedom.. A feeling that you were in an infinite environment without any borders. Millions of star systems to visit, yet no specific directions, orders...
The realism was greatly enhanced by Braben's effort to "simulate the whole galaxy". It was wonderful to see the astronomically correct data. (Except space isn't blue, damn it!! :)


What can be done better...

Naturally, the most obvious part is the purely audio/visual factor which is always limited by the hardware available at the time of release. But as I said, before, graphics were not what made Frontier and it's predecessor such great games, there were other factors.
The overall downside was basically the lack of variation. Even though there were millions or even billions of star systems to visit, they all looked pretty much the same. Bug or feature, settlements could be found hundred of sectors away from Earth (Sol system), yet these settlement were exactly the same as those found in the sectors near Sol even though they were (as I said) many thousand light years away.


My vision for the ultimate Frontier experience...

So what do I want to see in a game of this kind? Well, I'll skip the obvious stuff which is being discussed at other Elite 4 fan-sites and so on.. Instead I will focus on the unique ideas... All right here we go then...

I have a vision of a (simulated) universe where the content is so unique that not even those who wrote the game would have the slightest clue of what's out there. Every star-system would have it's unique environment, animal life or even it's own unique civilizations... You would be able to fly to an unexplored system and land on a habited planet. There you could exit your ship and walk on the ground, you would see an unique environment (in true 3D).. Unique floral life, animals... Wondrous creatures, too strange to come from the average 3D artist.. The creatures all have good AI so they seem to behave very naturally... The 3D models & textures are unique as well. (The textures would be procedural, not bitmaps as for the 3D models... Some kind of NURBS objects perhaps...)

The big question is how?! How can this be done?
The quick answer is: Algorithmic programming. The programmers would only set a (huge) number of rules and define a number of algorithms. It would basically be a life simulator where life can evolve on it's own. Every star-system would get a random 11-figure random number. After that a simulation would be run to "activate" these life-simulations. The algorithm would be carefully balanced, but not force any specific development.

  • First of all the star-system itself would be algorithmically designed. It should reflect the rules we have learned so far (and that we will learn in the near future) about how star systems are formed, wether there will be any planets or how many stars there will be in the system.
  • Next, when the star-systems are defined, there would be a random factor that would decide wether life will even star evolving on suitable planets.
  • If life does start to evolve, there would only be some loose rules of what's possible or not, just a guiding light to get things going. After that the life-object would become small chunks of classes in an object-oriented environment. These "chunks" would be able to "mate" with other chunks an produce offspring. Skills would be inherited from on class to another and so forth.. It's likely that there would have to be some additional rules to help the algorithm to evolve beyond a certain step, but I'll leave that to the metemathicians..
  • The AI (Artificial Intelligence) of "animals" could also be algorithmic with some simple rules and boundaries. To get a complex AI for creatures you don't know anything about would be very difficult of course. Obstacle avoidance and random movement should not be impossible to achieve, however..
  • To make this truly amazing (and truly unachievable for the next 10-15 years or so.. :) Some of the life forms on these habited planets would evolve into intelligent species capable of crating their own technology. I can imagine that the set of rules that would need to be set up here would be incredibly complex. But imagine how cool it would be to meet an advanced species who's capabilities you know nothing about. (It would be like when the Enterprise meets the Borg for the first time, a truly classic Star Trek episode.)
  • All these advance algorithms would of course have to be pre-calculated. When all the conditions were set up, the computers would run the simulation to equal 5-10 billion years of evolution. The "game" itself would only run online with dedicated servers for AI and all the heavy stuff like that. The client would only be one of millions of participant in this huge simulation. (The game would be impossible to run without heavy duty servers. THe client-server model is also very fitting for these kinds of games where the player is one of many participants)

I understand that all this sounds "pretty much impossible" and more or less, at this point, it is.. But games evolve. Compare the simple games such as Pong with todays games like Half Life 2.
     Let me know if these links stop working..
Sooner or later, someone will attempt something like this. If it takes 10 years or 15 years or even 50 years, I cannot say. History shows that most ideas are thought by many people independently, throughout the world. One day, when computing power is far superior to todays, and the graphical limitations of graphics card are history since long, there will be someone ( a 60 years old David Braben? :) that will initiate a project like this. Maybe at first at a smaller scale but nevertheless..



Return to main page! | Go to Daniel Sevo's homepage!