They often focus on the "kill a monster and it stays dead" angle, so that seems like a real aspect of the system, but it's not an advance. Other companies have aimed at this "persistent virtual world-building toolkit" target before, from Second Life to Blue Mars to Metaverse to Metaplace. What distinguishes Improbable from these other ideas?
As I understand it, Improbable also simulates one big world, but they map the various simulation calculations to servers in a different way. How exactly? That's their secret sauce.
Philip Rosedale's new startup, High Fidelity, also takes a different approach. See https://highfidelity.com/
Too me it reads like MapReduce for highly place-dependent computations, whatever that looks like. Probably something along the lines of a distributed kd-tree with message passing at borders handled for you as well.
Imagine every object sitting on the same planet (object). Suddenly, everything interacts with everything. You can't know if the object is a bee gently landing on a flower, or a 20 mile radius asteroid crashing into earth. The effect of the latter would be felt everywhere.
So, if every object was sitting on the same planet (object), even their system would probably crash and burn with n objects passing n^2 messages all the time.
The kd-tree idea sounds neat. They would need to adapt it however, to allow objects to move (to pass in and out of other objects' interaction radii).
In this system the server is basically a distributed database and message passing system, and grants different clients authority over specific simulation objects or fields. If a client is the authority it runs the simulation logic for that object. As long as you can scale you message passing and db it seems like most work can be offloaded on clients, so complex simulation is "free". I'd be concerned about cheating in more competitive games, but it should work well for social or creative games and for non-game simulation where you know clients are trustworthy.
Splitting this article into 5 pages to increase page views is really sad & completely unneeded.
Here is the website of the game itself: http://www.bossastudios.com/games/worlds-adrift/
I just googled 'Improbable London'. On their website not much substantive , just a standard 'hello world' post: http://improbable.io/blog/
I'm personally interested in such an approach for an academic tasks ( complex electrical simulations), but I have no idea where to find some substantive discussion of what they're doing here.
kind of sad
Improbable seems to be different in that it's aiming at the gaming market? But apparently they also have science clients? Maybe the difference is in their computational architecture, or cost?
vs
"In their ground floor office on a bland-looking block in Farringdon, a team of about 60 engineers from MIT, Goldman Sachs and Google sit at $40 desks writing code..."
Typical in gamedev :) (e.g. Unity Technologies):
"“Eventually we realized the tech we were working on was bigger than the game,” says Whitehead."