Arne ClaassenThe way things stay editable in Canvas of Kings gives me an idea for #edcgame. I never liked how construction was immediate in #cityskylines and I did like that road building in #simcity took time. So how about a builder mode, in which you lay out what you want. All editing is non-destructive and it shows you the projected time and cost and then once you commit, demolition occurs and construction happens in game time.
Absolutely love Canvas of Kings, a gorgeous hand-drawn aesthetic map making tool. Lots of cool ideas in there for building the Apple Maps-esque UI for #edcgame
While I really like the emergent design that comes from the WFC game loop, I started this projected because I needed to hone my 3D programming skills in order to move #edcgame forward. And going with the free form village builder would yield a lot more usable skills for returning to the city builder
Had an idea for a cozy game leveraging the graphical tools I’m building for #edcgame :
I absolutely adore the crazy, cliff hugging villages of Liguria and I can see myself loosing many hours in a townscaper/monterona-esque builder, creating colorful villages on procedurally generated/eroded coastal cliffs
Plan is to put the simulation on hiatus for the winter break and devote the time to graphics, hopefully finishing the virtual world tutorial to #godotengine port #edcgame
Having built my simulation with sqlite as the original data backend, I just kept nodding my head while reading @ajmmertens's "Why it is time to start thinking of games as databases" https://ajmmertens.medium.com/why-it-is-time-to-start-thinking-of-games-as-databases-e7971da33ac3
I've realized that even if I ran sqlite as in-memory only, performance is going to be an issue and #flecs seems to be better suited. Plus I get graph traversal which is still a pain with an RDBMS. #edcgame
I've spent the last couple of nights wrapping my head around modeling in #flecs. Having based my original simulation model on a relational DB schema, it actually fits better into an ECS with relationships than an object model, given the Object-Relational impedance mismatch.
I think this model will get refactored a number of times as I go through this and get my ECS footing, but #flecs feels like the perfect match for agent-based simulation modeling. #edcgame
Started on a spike porting a subset of the current simulation to #Flecs, see how I get along with the paradigm. Been deciding between #Flecs and #ArchECS and have settled on Flecs for now. We'll see how the .NET binding treat me and this way making a later switch to #swift would be less painful than if I went pure C# for the ECS. #edcgame