What we want is a labyrinth which expands without limit in the direction the user chooses to go — any imaginable connection between every possible combination can be simulated instantaneouslyix. A solution might be found in a universal puzzle where each piece is different from the others, yet at the same time each piece can at all instances be seamlessly combined with every other piecex
Let's take the most common everyday representation of a labyrinth, a cityplan. For a boundless simulated cityplan we would need software to manage four main libraries. The first two are:
These libraries consist of a number of elements pertaining to the specific function of the library, i.e. the cityplan library contains a number of cityblocks with streetgrids that can interlock with all other members of the library on all four sides (or maybe six sides, or eight, or sixteen, or fractal...).
Similarly the architectural library contains comparable architectural elements that fit together in all sorts of ways to form buildings, cars, fire hydrants and other assorted street furniture.
The library management software picks random elements from both libraries and combines them into streets and buildings as the participant looks around or walks along. Similar to the Hypercard paradigm this system should have a certain memory for those parts of the city that the participant passed through; it wouldn't do to retrace one's steps and find a completely new environment instead of the one that was experienced five minutes ago.
Now that we have, at least theoretically, our boundless virtual city, we need people following a storyline to immerse the participant in. This is where the other two main libraries come in, and these are rather more tricky:
The character library which, analogous to both others, contains elements to create characters with, people to walk the streets and birds to populate the trees with as it were. Of course these characters need a script to guide them along and this is where the last proposed library comes in.
The dramatic library is the most important, and therefore the most difficult to realise practically, since it enables infinite storylines to be created. By writing little pieces of drama, the story can be expanded infinitely; it effectively never ends (the immediate problem that comes up is: If there is no end, where do I start?).
So creating a fractal labyrinth is akin to knowing all possible moves in a chessgame — and thus all possible outcomes — before the game has actually been played. And then think of the number of gamepatterns that each opening situation can hypothetically generate.
Then there is of course the question of fractality with regard to detail: when one enters a house in this city, what is inside? Furniture, glassware, dust between the floorboards if one gets down on one's knees and cares to check? In other words what we are talking about is the granularity or resolution of our VR simulation. I am inclined to think that a solution to this issue can be found along the same lines as the previously discussed library system.
In conclusion one might say that a good polylinear program can forecast every possible choice of every possible user at any time. Virtual Reality should expand faster than it can be anticipated–predicted–divined–desired by the participant, but not so fast it cannot be generated by programmer/creator and soft– and hardware subsequently.
A polylinear labyrinth cum puzzle can be compared to a book in which every chapter can function as the opening one. Therefore every chapter must link logically with the following ones so that the reader can cross from one to the other in any sequence. And that is only a one–dimensional book.
As the reader starts to read the chosen chapter, as he scans the world created by the writer, an expectation curve starts to grow. Based on his empirical database, the reader expects certain events to take place.