Major/broad topic - certainly high-level intentions for both visual (in the HUD system) and sonic notifications based on real-time data, fully customizable, perhaps even an alerting-specific product implementing a server-side component for traditional alerts not connected to running a Win32 desktop app.
Another major topic. From day 1 in September I conceived of the major potential for WSV3 next-gen to serve purposes of forensic mesoanalysis. I'd like to make an expensive industrial product aimed at insurance so I can keep consumer costs low. There could be cheap offerings too based on an on-demand system. I have a vision of making WSV3 able to retroactively analyze past storm events after a period of on-demand server-side processing. For instance, it would be incredible to re-live the October 2021 100mph+ hurricane-force Nor'easter years later in WSV3, with 3D satellite, model/mesoanalysis data, obs data, and radar. The forensic case starts to multiply the order of magnitude of data costs, so I need to build everything else first as efficiency as possible and then see what's possible. But I am not building the real-time mesoanaylsis data in a bubble, but rather flexibly, so that the time axis can be extended and transposed for forensic data.
This is basically the entire WSV3 Tactical Cartographer product. More details soon. Still doing a few months of high-level top-down work before going back into the nuts and bolts of performant, fluid, 3D hardware accelerated map rendering. I intend to have a base vector layer of seamless vector tiles automatically loading street level data across the US without any custom import necessary. I intend to go for road name labeling and fully customizable shapefile import.
This sounds like a concept I was trying out in V6P, in the font editor window of the test new font engine and the new visual color palette editor. I love the idea of having Photoshop-like "preview" boxes in visual customization dialogs. I am building a generic styling system based on SDFs where you can style the edges of contours. I am trying to abstract this to many applications: geographic lines, contour lines, convolved polygon edges, coastlines, literal font characters (the new font engine you see in V6P uses SDFs, and we are running with that), SDF shapes/edges such as rounded rectangles you see in the V6P hardcoded city labels, and more. I am designing this in a way where, hopefully, one common set of vector styling and edge colorization settings can be applied and layered in flexible ways to produce incredible vector graphics. Custom blend modes will be a big part of this. But yes, a visual 2D preview window rendered in D3D11 graphics for these styling customizations is the way to go and already familiar from programs like Photoshop.
This is very interesting and I had some ideas. One idea was native integration of my own private hobby mesonet, MAREPAM.org, a nascent high-performance multi-sensor anemometry mesonet for coastal New England with unique and novel full-resolution sampling. Perhaps this could be extended. I thought of a concept, "WSV3 Tactical Anemometrist", which could feature hardware anemometer integration and special forensic tools for observation hunting from all the many sources, QCing, and objective analysis/IDW interpolation between sparse points. This shouldn't be limited to wind speed though, and a lot of these features I have planned anyway for the observation system in WSV3 Tactical Mesoanalyst.
Personally, I really care about the point observations. Much more so usually than radar/satellite/model data. WSV3 first-gen doesn't even come close to a competent and fully-featured tool that I would use for my own storm chasing anemometric missions out on Cape Cod for big tropical/Nor'easter events. I am going to do some very cool things with the observation system, including creating the ability to ingest the entire MADIS mesonet dataset, which is something like 150,000+ obs per hour. It is extremely heavy and likely will require an add-on data sub, but nobody has put all that full-res obs data in a real-time-rendered 3D virtual globe map yet, as far as I'm aware.
This could be extended eventually by some sort of private network as you mention. By creating the best tool first, with the highest value at the lowest cost, the value for such a proposition is established.