ThymineC wrote:On a side note, is there any value in having pocket universes being "unstable", in the sense that they can be dynamically created and destroyed?
Regarding the actual value of it, might be pointless if you can connect to an infinite amount of universes. With systems, I would rather avoid it, since creation of systems would have to be carefully regulated or you'd have ridiculously dense areas. Would be cool if you always have the option to check out a new universe, but it's not necessary.
There might be an infinite number of universes throughout the whole graph, but that wouldn't render the creation of new universes pointless; within any given "region" of the multiverse, there will be a finite number of universes, and the creation of new universes that then link with other universes may create new, shorter trade routes and thus alter the economy considered in this local region. It may even shift the balance of power between factions. The destruction of universes can also have major repercussions, potentially eliminating trade routes or shifting the balance of power between factions in different ways. I discuss the destruction of universes in more detail later in this post.
You can easily balance the system to avoid it forming ridiculously dense areas - Josh is currently coding connectivity to work by generating the minumum spanning trees between all systems in a region (using Prim's or Kruskals or something like that) and then merging these spanning trees to produce the final interconnectivity between systems. You can easily pre-generate a set of nodes that are initially all disconnected from one another, and then determine the connectivity between them by increasing or decreasing the number of spanning trees that are generated and merged. Within this set might be nodes that are not connected to any part of the main graph at all - but if we assume that this connectivity algorithm is ongoing
, then at some later time the node might become connected with the main graph. Similarly, some currently connected node may be disconnected from one or all of its neighbouring nodes if we assume non-monotonicity in the connectedness of the graph - this can be done by keeping a rolling buffer of spanning trees that are, at each moment, determining the overall graph connectedness as the merged result of all of them. This is how you would think of the system abstractly
, a node that was previously disconnected but at some point got connected could be thought of as a universe popping into existence - likewise, a node getting disconnected represents a universe popping out of existence.
Special thought needs to be given to the destruction
of pocket universes. I raised the idea of unstable universes a while ago in Wormholes (in-depth)
, to which Hardenberg replied:
Hardenberg wrote:MEEEEEP. Lousy gameplay alert. Since you as the player have no immediate way of telling when the "pocket universe" will collapse, this one equals an arbitrary game over coming from nowhere. A more viable solution would be to make the wormhole unstable, but generate a new "exit" when the first one collapses. (Yeah, I know: EVE again). That way, you can get "lost", but it's not an instakill. It might take a bit of traveling to get back to familiar shores, though.
My reply was:
ThymineC wrote:Who said the player wouldn't be able to tell when the pocket universe would collapse? Donnie Darko knew when his would. The player can perform a kind of analysis to determine the stability of the universe before or immediately after they enter. The suggestion you offer could work as well, sure.
A player will have the means to analyse how close a system is to collapse at any time and prepare for that.
Questions: If any of your assets are in a pocket system when it collapses, you will lose them. This will include fixed assets like factory installations, stations and other facilities. Would this be a desirable or annoying gameplay feature? How often should pocket systems be collapsing? Is there any way an agent can act to "anchor" a pocket universe, to lower the probability of a universe collapsing? Should it be possible to lower this probability to zero?
What happens if the player is in a universe as it collapses? I reckon the most obvious outcome is that they die - but then how is death handled? I have my own thoughts for this, and so do many others in Death in LT
Kekasi wrote:The issue is dealing with edge cases where universe/system might lose all connections.
This is very easily solved - just make it so that any universes that lose connectivity with the rest of the multiverse get destroyed. That is, FOR ALL X
, IF it becomes impossible to navigate the current topology of the multiverse to get from Universe X
to every other universe, THEN Universe X
gets destroyed.* This means that if a universe gets destroyed, any universes in a "cul-de-sac" who are only linked to the rest of the multiverse via that universe will also get destroyed. The multiverse remains a fully-connected graph at all times.
Question: If an agent can anchor a pocket universe in such a way that it has zero probability of natural collapse, should it still respect the topology of the multiverse graph? If it's part of a cul-de-sac and an unanchored universe between it and the rest of the graph collapses, what happens to the anchored one?
*∀x∀x: Universe. ¬RouteExists(x, y) → ShouldBeDestroyed(x)