23 hours ago, Paraffin said:

The downside is disappointment in the user base if a feature ends up not working well enough. And even more disappointment if the feature gets pulled out of the final release. I assume that's one reason why there is a private beta and password-protected feature list in the first place.

I still remember hearing about a feature -- maybe a year ago? -- where they had worked out a way to show more appropriate water colors in shallow areas and tropical regions, instead of that slate blue water color everywhere. It was expected to be in a current beta or final, but it never showed up. If it wasn't going to make the cut, I'd prefer not knowing about it.

P.S. as a side note, I still hope they can figure out how to do that. We can install small custom airport and region files that use orthos or hand colored bitmaps for tropical water, but ideally this is something the sim should generate in the terrain engine, from whatever data sources are available.

I don't care. I've lived with so much flightsim based disappointment that it just can't get any worse.

Shallowness would need to be determined from elevation data, which would require an elevation model for water regions, which in turn means less performance and more data to store.
You can, of course, use a waterclass based approach, but this would require careful blending to avoid hard edges.

 

Btw: Flightgear's water shader already supports varying depth by depth map, but I have no idea where they got one and how exactly they are processing it.

Edited September 23, 20187 yr by Bjoern

14 minutes ago, Bjoern said:

I don't care. I've lived with so much flightsim based disappointment that it just can't get any worse.

Hah! I can relate. However, the problems arise when people less patient and understanding about software development start whining about the developer hyping their product, breaking promises, etc. It's just bad P.R. to promise and not deliver.

 

Quote

Shallowness would need to be determined from elevation data, which would require an elevation model for water regions, which in turn means less performance and more data to store.
You can, of course, use a waterclass based approach, but this would require careful blending to avoid hard edges.

More data to store (a little more), but maybe not a performance hit if it's burned into the mesh tiles. Anyway, I suppose we'll get this eventually, one way or another. 

23 hours ago, Paraffin said:

Hah! I can relate. However, the problems arise when people less patient and understanding about software development start whining about the developer hyping their product, breaking promises, etc. It's just bad P.R. to promise and not deliver.

Ah, the whining will be there no matter the outcome.

Am I gonna go blind because I clicked on the link?