My own experience with the development process is that they ran roughshod through the feature list with a minimal implementation in order to make August 18. They are intimately familiar with what features are missing, and what needs additional work, whereas we are looking at a proprietary mystery box and noticing things appear to be broken without a real way to track how or why. The game stays in that kind of broken, half baked state until enough users make enough racket that the developers prioritize that issue. This isnât a matter of them being clueless to some quirk that is hard to reproduce, and all they need is better documentation of it. Itâs a matter of clamoring for attention until they finish the feature set. Itâs why the moderators on here tell everyone to report their problems to Zendesk, even though the same issue has already been reported 1000 times, and why the post above this one had broken weather on the last ten flights. They know itâs broken. But hopefully if these discussion are ongoing and the pressure is maintained, attention will be directed where itâs needed, or else theyâre going to call this incomplete or minimally complete implementation âgood enoughâ.
The trope on here is âI wouldnât expect study level out of the boxâ, and that goes for weather too I guess. I also wouldnât expect highly dynamic convective systems to be picked up before or as theyâre forming like that August 10 derecho. The global ensemble 12-hour forecast should probably work well enough for 95% of the users and cases on here Iâd imagine. Even with the 6 hour lag getting the 12z Meteoblue run into Flight Simulator, it should have been able to get that derecho as it came into Chicago (while missing its morning track over Iowa). But usersâ experiences arenât anything like this.
The beauty of their approach is that once theyâve implemented this gorgeous volumetric weather engine with layers and tiles that blend, then actually setting it to live data should have been relatively straightforward. They need only query Meteoblueâs output for a variable list at location x,y,z at forecast time f(t) and f(t+1) and then average these variables to get the âcurrent weatherâ.
The accuracy of the sim weather using this particular model would suffer a bit compared to something like the GFS, which updates twice as often and finishes its runs hours earlier. But I totally get why they did it. The convenience of having Meteoblue produce simulator specific variables for them allows them to focus on other aspects of the game.
Itâs anyoneâs speculation why the live weather appears a day old or completely mismatched. But my own guess is that itâs that the implementation of live weather and the weather engine itself are only half done, works in progress like much of the rest of the game.
Iâd hope Asobo spends its time working on that core weather engine so the sim can actually draw the different weather types users expect, and process generic weather data as it comes in. I wouldnât expect them to start blending in a complicated mix of METARs, higher resolution regional or convection allowing models, or radar data, unless Meteoblue is also in this for the long haul and is actively developing for the game still.
So my other hope is that once they finish implementing the weather engine, they expose it to the third party developers with a user configurable source. The sim shouldnât care where the data is actually coming from when it grabs variables at a lat, lon, and time, whether itâs the GFS when youâre over the ocean, or the HRRR when youâre over the continental US or METAR when youâre at an airport. Perhaps this will never happen, however, due to some agreement with Meteoblue, just like we wonât see Google Maps imagery in the scenery.
But my worry is that the foundation on which the weather is built is shaky. We donât have cirrus, real haze or fog, or thermals, ceilings and tops that are accurate or controllable last I checked. Even if the sim is exposed to third party injection, there might be nothing to plug those variables into. Itâs the same worry I have for the aircraft. Everyone is saying, âJust wait for the third party developers to come along and youâll get planes that fly properly.â But there appears to be fundamental issues with the aerodynamics engine itself. You donât need much of any rudder correction in flight, and you canât use the rudder to effectively side slip. Third party add-ons arenât going to fix that if the simulator isnât even modeling some of these aerodynamics. It might be the same way with the weather, and if these issues donât start getting prioritized on these forums and development snapshots, this going to be know as the simulator with half baked weather for a long time.