Awesome work :) Thank you!!
As for charging for it - I think you guys absolutely deserve to get paid. There's no forcing in providing a good quality service with awesome features and qualities such as smoothness and relative simplicity, listening to the folks who use your stuff, and doing it all sustainably w/o going out of business. Keep at it!
Boundary layer would be an interesting part to see visualized.
You could make it premium, label it experimental and stash it in some weird container where it can live even if it was buggy or otherwise non perfect, and subject to change or removal later on.
Then it'd be a matter of defining a scale and creating & serving tiles for a specific parameter of a model right? You could also add more parameters for experimentation without perhaps committing to them fully. Might even generate the tiles runtime, even if that was slower, just to make it easier to experiment with parameters and see what would finally make it's way to the actual list of layers
I know the official reason for not being able to choose a default layer is cache fragmentation. And the secondary reason is that "this is windy and not radary etc" which likely is more of a joke, but anyway 😛
I'd suggest adding a setting "remember last open layer"
Think of it this way, if a lot of people from a certain geographical area are looking at radar (or any other layer) instead of wind, the caches would actually adapt to that behavior and (could) make the experience better for all.
Label the setting as "experimental" and kill it if it causes trouble. Or make it available to premium users only.
Also, I would not mind if the app still preloaded wind layer data on background, or if the caches were configured or set up in a way that gave preference to the wind layer if you think this is needed to preserve decent initial load speeds for most users.
Two things to balance and I'm sure you've thought of these: initial load speed vs getting to what you want to see speed
@korina I see something like this was added. Thank you :)
Would appreciate it if the 1hr setting in forecast got persisted, either locally in a cookie or account wide, won't matter much which way
3D globe is unusably slow on my Chrome if particle animation is enabled. Disabling particle animation brings up fps from 1-2 to maybe 20-30.
On Safari it worked better (even) with particle animation enabled; fps is at around 30. Obviously the chrome-issue could be related to browser settings or something, as on my other machine I do get 20-30 fps on Chrome with particle animation enabled.
Not sure what all is going on but clearly if there's a setting that causes the 1-2 fps lagginess then that could be documented somwhere or worked around. To me it seems like there is an issue somewhere.
The max fps of 30 could be improved to 60 but for some reason the composite step only happens on every other frame... anyway, would be way smoother and nicer at 60 🤓
Can't think of anything to check in Chrome settings. Hardware acceleration is enabled and everything else functions properly, including some webgl games.
Anyone else have similar issues? Suggestions?
@korina compatibility mode switch has no noticeable effect (even after restart), but disabling particle animation did make it more usable... what do you think should be checked in browser settings?
@Korina if I disable particle animation through the lower right corner toggle then it becomes somewhat more usable. Not quite as smooth as it could be, but usable. No dropped frames, but with a gputask visible on only every other frame. Frames are green (16ms) but feels more like 20-30 fps probably because of the gputask getting triggered only on every other frame.
@Korina the 3d globe iirc was smooth when it was released, but now (for some time already, can't remember when it started) it is pretty much unusable. Chrome shows most frames being dropped with one green frame every one second or so, and with GPUTasks taking 700-800ms each between frames. If it helps I could send you a chrome performance profiling json if someone wants to take a look.
@idefix37 Ah so it seems! Totally forgot that there's the isolines box in the drawer :) Thank you
Awesome work :) Thank you!!
As for charging for it - I think you guys absolutely deserve to get paid. There's no forcing in providing a good quality service with awesome features and qualities such as smoothness and relative simplicity, listening to the folks who use your stuff, and doing it all sustainably w/o going out of business. Keep at it!
Also, obviously, would be superb if the pressure toggle got persisted locally in a cookie or in account details similar to how the 1hr forecast toggle could be persisted
Can you add the pressure lines toggle to weather radar? Would make it way easier and faster to infer in which direction the rainy areas are likely moving without having to either check the wind screen or scan back and forth with the time slider and then try to infer and extrapolate from that :-)
Would appreciate it if the 1hr setting in forecast got persisted, either locally in a cookie or account wide, won't matter much which way
Boundary layer would be an interesting part to see visualized.
You could make it premium, label it experimental and stash it in some weird container where it can live even if it was buggy or otherwise non perfect, and subject to change or removal later on.
Then it'd be a matter of defining a scale and creating & serving tiles for a specific parameter of a model right? You could also add more parameters for experimentation without perhaps committing to them fully. Might even generate the tiles runtime, even if that was slower, just to make it easier to experiment with parameters and see what would finally make it's way to the actual list of layers