June 2016 Update

[Lumituisku]

Some new civvies and map objects, pic by Jartsev

[Lumituisku]

 

[Lumituisku]

http://www.steelbeasts.com/topic/10355-sb-pro-pe-40-discussion-thread/?page=63#comment-154362

We're working on "Plan B" (which actually is more like "Plan C" because B suggested that we stop working on 4.0 until about October and then mount another effort for a Christmas release). We managed to wrestle a few more weeks of bug fixing and testing time from the work schedule, and I intend to use the full amount to deliver the best possible result under the prevailing circumstances.
 
What that means is that we probably must cut back some ambition in order to bring at least most of the advertised features in reasonably good condition out in July, and then provide a major update half a year later, approximately. What exactly that means is currently being evaluated and tested by the Beta team. The final decision has not yet been cast, but I'm cautiously optimistic that we can wing it.
 
The likely outcome is that the high resolution terrain will have to wait until about Christmas. It's not that it doesn't work, it's just that we decided together with our army customers that we need to find a better balance between good performance and retaining user flexibility; previously we had fucused very much on the performance side of things (both as far as frame rate and loading times for maps and scenario files were concerned) at the expense of some loss in flexibility.
For example, with the current terrain engine we modify the height profile at runtime depending on whether a given point in the terrain is covered by water. If that is the case, the surface point is depressed in height by a certain depth. This works well because the number of terrain grid points is reasonably small so it can be done at virtually no waiting time while loading a scenario. It does however NOT work well at all when you do this with a terrain resolution that is more than 500 times as high.
 
What's more, you also need to offload some computational load to data "preprocessing" to speed up scene rendering and, in particular, line of sight calculations. Preprocessing costs even more time, at least about 20 minutes per map, and up to several hours in extreme cases. Nobody wants to wait that long when you fire up a mission, and you wouldn't want the current 3.0 frame rates cut down by a factor 500 either. But just as well "brute force" is not a viable strategy either.
At the same time however retaining the ability to modify a terrain for a scenario in some places here and there is a legitimate demand, and the original concept would essentially have required to spend all the preprocessing time whenever you wanted to modify a map. In practice this would have meant that maps were to be treated largely as a static thing, and that users would have to give up the ability in Steel Beasts to do quick & dirty map touch-ups.
 
So we decided that we need to go back to the drawing board as far as the balance between performance and user flexibility is concerned. But it was nothing that we could do on short notice without creating a very grave risk of ruining everything for everybody. Also, we didn't want to publish a new terrain engine where we'd know well in advance that we'd change the file handling half a year laters. So our idea of a workaround is to keep pretty much all the features that we advertised in the 4.0 videos, except that we would provide the the high resolution terrain as a free update later on.
 
 
That's where we are and what's happening behind the scenes right now. We don't know for sure yet whether this will work. Right now it looks as if things might work out, so there's reason for cautious optimism. But I wouldn't break out the champagne just yet.