IceBlinkEngine.com

So, progress to report (after a rather mind wrecking phase of tearing code apart and then stiching little Frankenstein together again ):

All (non-time driven for starters) movere on the neighbouring maps are fully functional (pathfinding, scrolling, animation). Next up are their barkstrings. Then it's onwards to the time driven movers.

The time-driven movers are quite hard to get right. I hit a nasty bug with correctly re-loading them from a save, but this is fixed now.

Also, they move just fine on their respective areas as long as not crossing areas. When they cross areas though, they still jump from square to square (no scrolling). The code I wrote for this is not applied correctly right now, so I must look deeper into this. Furthermore, the floaties they use for signaling an area transition (when not moving between nearby areas) need work, too (to do: change these to work from prop's pixel position).

Afterwards, I will have to set up a more complex testing ground for them (eg 5x5 connected area system with many of them running their routines) to put their system under fire. Only this way, I will likely find out where more bugs might lurk.

It's tricky work on details, but I think these time-driven movers are right at the core of a living world, so I hope it will be worth it in the end to go for that extra mile.

I was able to make some more progress here and fixed a scenario where the prop scrolling on changing area did not work out. There remain other scenarios to test though and I already see at least another one that's still broken. We will get there !

One more down! For the moment, it looks all fine. But I will have to do more testing, before declaring victory over this particular nemesis. My reward will be diving into Hearkenwold again !

The current batttle plan:
1. Set up large testing ground 5x5 neighbouring maps (populated by all kinds of different movers and complex topography)
2. Try to survive debugging the testing ground , likely finding out more about the performance limits of IB, too
3. Reward myself with Hearkenwold and fresh Halls of Hamlock
4. Try to utilize camera scrolling and extended prop animations on combat screen
5. Start implementing the "square type" system described far above for eg the sea or space in a scifi scenario

Note to self: Maybe extend the floaty info in debug mode for time driven movers - like showing (1) x,y coodinates, area name and departure time of next way point as well as (2) departure time of current waypoint.

I built 25 neighbouring 32x32 areas (ie all seamlessly connected). For the player this appears as one 160x160 area.

Looks like this:

12|13|14|15|16
11|03|04|05|17
10|02|01|06|18
25|07|08|09|19
24|23|22|21|20

Within this construct I populated 4 areas that share a common corner point (eg 01,02, 07 and 08) witch about 20 random movers each. Despite about 80 mover props potentially in the nearby zone, it scrolled rather smoothly, inclduing animations and prop moves (considering we still run turn and square based in the background, of course, so please dont expect something like a 2d action game ). I yet have to see though how chasers and especially time driven props will affect the performance. Also terrain is flat/all walkable right now - adding more obstacles might slow down the pathfinding, too.

Furthermore, there's a bug related to drawing/maybe even calculating props on the diagonally opposite area (like between 01 and 07). I will start with looking into this one...

It must be really fun to watch it all come together. The movement and seamless transitioning sounds good.

It feels indeed very satisfying . Within its limits (hobby project concerning the stuff I add), IB has gone far and vast beyond initial expectations and goals.

But it also has a few bugs which currently haunt me a bit. I did not do much testing with the diagonally neighbouring map situation and this forces me to debug a lot of little glitches currently. As this is a complex setup, it is sometimes hard to track down the cause of a particular bug, especially when it comes to bugs related to drawing graphics (eg framing the screen corretly currently).