In the live stream Mavor mentioned about picture in picture while talking about what they are going to add instead of a conventional mini map. That got me thinking. Could there be some sort of unit/ satellite that instead of just getting rid of the fog of war wherever you put it, give you a real time picture in picture view of what the enemy is up to or just a view of what is going on at a specific location? I think this would work with a satellite or a spy plane of some sort. Now discuss
Pretty much exactly what he said - If you want to have a unit give you a PIP view, just set up a PIP and set it to track it. Then you've got a PIP of your satellite. In theory, of course, we won't actually know how everything is going to work until we get the releases in our grubby hands. ---- What I'm interested in, though, is how Uber plans to deal with the graphical rendering cost problem (normally) associated with PIPs - I know that, for example, that while PIPs are often suggested as a solution to unrealistic scopes in FPSes, it's often discounted because it would essentially halve the frame-rate (due to having to render the scene twice). Now, of course, Uber are making an engine from the ground-up, so I wouldn't be surprised if they are being really clever about it (or, alternately, perhaps fine-grain graphical fidelity isn't going to be a problem? That also makes sense, after all, it's an RTS, not a FPS). Still, what they do aught be rather interesting.
I like the idea of having either a windows system in-game, or actually using extra windows in the OS. That would kick ***!
The graphical rendering cost could be a significant problem for people with lower spec machines if its not done right but I'm sure uber will find a way I would however like to know how the are going to get us battle information from a different planet ( your base is under attack, the enemy has just amased an army of bots, that sort if thing) without using a mini map
Well, fans have made a lot of suggestions that aren't explicitly minimaps, but fulfill the same role. This is a similar idea. Ideally, you'd be able to glance at whatever the quasi-minimap would be, check where your units are, and your enemies' are, see what units are damaged/under attack/not functioning, and see where conflict is occurring.
I'd be worried about the performance cost, not to mention it'd probably be a great deal more distracting (movement constantly catching your eyes) and require more screen "real-estate" than a simple mini-map (not a big deal for those with huge monitors I guess). Don't actually want this unless it's completely optional.
A minimap is, in essence, a PIP window that is zoomed out all the way to show the whole map. I don't understand what this whole argument is about...
Well the basic concept is user reliant, meaning if you don't activate it, then it doesn't happen. Simple as that. As an aside, I have three monitors to play with, so any screen space I can fill is what I want to use. Sixteen satellites in a surveillance grid would be glorious.
In many traditional RTS games, you can create a control group and selecting that group will automatically set the view to the group. This could get frustrating/poor-responsive if you have a group split literally on opposite sides of a planet, but this generally should not occur as control groups are used to send units together.
Supcom had cartographic mode, which removed the render of the game (made my pc go from 10fps to 20-30). Having a satellite then tracking it (default t in supcom) then setting the view to cartographic shouldn't chew to much resources (if the game supports cartographic as it did in supcom).
The PIP stuff is part of the chroma-cam I think and chroma-cam are basically smart rendering windows - no logic - just fast client side rendering panels for subsets of server-side data. All the real work - dmg calcs, win/loss, etc are all server-side. Meaning - opening a window to some coords will result in streaming a small set of data from the server to the client, rendering only happens when things enter frame. With smart ways of handling data now - basically observable data where a change in some aspect of a data attribute results in triggering a UI change - you can do a lot more with the cycles available and with the API's available for GPU's - much of that can be thrown at the GPU. The flip side is the network characteristics have a greater impact - latency, bandwidth and even minor things like DNS lookup times can cause headaches.
