Re: [Flightgear-devel] Volcano - scenery model or AI scenario?

[Thorsten R. <tho...@sc...>]

> Add a checkbox to the rendering/scenery UI "Display volcanos as active",  
> set a
> property somewhere and make the embedded XML react to that?


> couldn't that be where a slider comes into play?

Okay, please bear with me and visualize this.

Fast-forward into the future, someone has added plausible visuals for all  
volcanoes in Iceland. You want to fly to Iceland to see an active volcano.  
How to set up this scenario, and what happens then?

***

If there's a GUI checkbox 'Display volcanoes as active', you'll have all  
30 volcanoes in Iceland set to active at once. They're not fuming a bit,  
they're spitting ash and brimstone into the air, showering hundreds of  
square kilometers of scenery with debris, creating a carpet of ash clouds  
across all of iceland. A volcano eruption is usually not a very localized  
event, you can see (and hear) it from up to hundreds of kilometers  
distance - Hawaii is a bit unusual here.

But... all volcanoes of Iceland going off at once is completely unflyable.


***

If there's a slider, what does it set? Probability of volcanoes being  
active? Visual quality of active volcanoes (I actually don't plan to code  
LOD levels of the visuals, so unless there's someone else who wants to  
code them, we're not even going to have the function)?

If it's probability, you do your flight to Iceland and... probably see  
nothing at all. Because the chance to see an eruption on a particular day  
for Icelandic volcanoes is pretty small.

So to see the volcano going off, you need to check the next 300 flights in  
Iceland whether this time it is finally active.

Not too satisfactory for the user who wants to try something out.

***

So, my design requirements are:


* I want to be able to set something like an activity level _separately_  
for every volcano
* I don't want any 'random' activity model, because that in many cases  
defeats the purpose to experience a particular scenario
* ideally I want to be able to do this from a significant (but not  
infinite) distance (as the ash plume is very visible from afar)


The FG design requirements are:

* we can't have a dialog listing 1200 volcanoes, so to be future proof  
(see carrier discussion) there needs to be distance culling
* we need an 'opt-in' model so that users on weaker hardware aren't

Our way of handling AI solves all of that without any additional  
ingredient needed, our way of handling scene models does not.

***

It actually was James pointing my thoughts to Iceland which made me  
realize that this should be an AI scenario. I was thinking Hawaii before,  
and there the case is less obvious as it's a more or less continuous lava  
boiling in a localized spot. For an Iceland eruption, I'd never have  
thought of scene model in the first place.

So, unless someone can outline a compelling design within the  
requirements, I'll go with the AI scenario approach.


* Thorsten