Feedback

[thkruz]

Way better than anything I made when I was 17!

Please don't take the critiques too harshly - I just wanted to focus on giving you a list of ideas on how to make it better.

A few positives up front. The UI looked nice, it didn't crash, and it felt professional. I really liked the hint that popped up to let me know I could snap windows when I was moving them.

1. Adjust the image of the earth that is used as the background for the installer. It makes it hard to read the black text and this is the first impression users will have of your product.
2. Offer the option to launch the application after installing. I am not overly familiar with .msi creation so this might not be an option.
3. After the installation I couldn't find it on my desktop or start menu. The default location of Program Files usually implies it will do that automatically. When I use applications that want to install to somewhere like C:\ it is more assumed that I will have to make the shortcuts myself.
4. Welcome screen is great as-is. Got my attention and was able to jump right into a sample file.
5. Auto-close the loading screen/menu after Scene initialization complete.
6. I am on a big 4K tv/monitor and Telemetry Dashboard takes up about 75% of the width and the Viewport takes up 25%. Feels backwards compared to something like STK.
7. Rename Viewport to something more user friendly like "Globe" "3D Globe" "3D Map"
8. 1st Person Camera was very unexpected as the default. I see now that it says free-fly, but every Astro tool I have ever used defaults to an orbital camera with the earth in the center.
9. Shift for -Y movement was not obvious and I still haven't figured out how to go +Y. I'd recommend using Q and E for Up and Down.
10. Give that you are very early on in development it is expected that stuff will break - if you plan on disabling multiple things like camera views then add a toggle in the settings to let me enable them at my own risk. Anyone willing to enable things that will likely break is going to offer up feedback that might help you get those things fixed faster.
11. I would move the Scene Resources to the far left and have its menus tab onto the Telemetry Dashboard window instead of the viewport (this one is purely my preference because it mimics other tools I've used)
12. Chandra X-ray Observatory seemed to be closer to the earth than expected, but that might just be the fact you aren't using SGP4.
13. When I Alt+Tab away (to type these notes) it doesn't release me from the Viewport so when I Alt+Tab back it is very jarring as the camera shakes.
14. The Play/Pause button is not obvious. I really don't have a great suggestion on this one - perhaps Green for Play and Red for Pause to get the users attention. Now that I know where everything is it is fine, but nothing drew my eyes to it.
15. The dark side of the earth is completely black. Not sure if that is a bug or intended but it looks broken. In the screenshots you have on github it looks like there is a visible dark side of the earth with a small transition (make the transition bigger/across a larger part of the earth)

It functions and looks like an engineering tool. I think my biggest critique is that the description doesn't match its current capabilities at all. I would make less promises of what it is going to be able to do, display something about it being an "Alpha Build".

I walked away thinking "what am I supposed to do with this in its current state. Maybe I missed something." I'd recommend you adjust the welcome message to thank users for testing and participating in its development and add a very short roadmap to set expectations. I can kinda guess that you are working on the camera controls and a new propagator, but the next time I load it up I want to see the last 2-5 things you added and the next 2-3 things you are working on so that I can immediately tell what features I should be trying and what I should expect to be broken.

Again - very nice start. I look forward to seeing where it goes from here.

[thkruz]

Rereading your release notes you clearly explain this is an alpha build so ignore those comments. I was multi-tasking and didn't read.

[ButteredFire]

The issues with the installer not creating shortcuts, the black background, and the camera being limited to free-fly mode will be fixed in the next release (v0.2.0-alpha), along with many of the UI/UX improvement suggestions you've outlined. Specifically, issues 1, 2, 3, 5, 8, 9, 11, and 15 will be addressed. Other issues have been noted, but I'm not sure I could resolve them in v0.2.0-alpha. I've already implemented SPICE and am trying to understand Vallado's SGP4 implementation, so hopefully Astrocelerate's scientific integrity will also be strengthened in the next few releases.

Thank you so much for your feedback! I'm deeply grateful for any insights I could get.

[thkruz]

Take a look at SGP4Funcs inside the Cpp/SGP4 folder. There is a SGP4Init function that conversation two line element sets into satrec structures. That satrec structure is then the input into the sgp4 function to propagate the satellite to a specific time.

Everything else is either helping that all happen or advanced stuff you don't need at this stage.

There are a few people on the Celestrak organization that is hosting the code that I know would be happy to clarify anything that is confusing. Good luck.

[thkruz]

Continuing feedback on v0.2.0:

1. Installing v0.2.0 went to the v0.1.0 folder automatically which didn't match the instructions. Minor but it did make me notice that the welcome screen doesn't say the version number (I found it in the help) and that might be worth adding.
2. I loaded the SGP4 test and noticed the time didn't match real-time based on the lighting but it wasn't intuitive where the time settings are located. Edit: After review there is not ability to set the time and that should be a high priority for future versions
3. Time Scale can't be changed while paused.
4. The start/stop button is under the viewport settings but the time scale settings are under simulation settings. It would be easier to find the controls if all time controls are centrally located.
5. I tried increasing time 400x to confirm the earth was rotating. You have a clear warning that will break things. The earth ended up tilted 90 degrees off of its axis, but I think your warning makes it clear that is my fault.
6. I did it again at 1000x but centered the camera on the earth and didn't interact with the camera - no issues. So the problem seems to be how rotation is being applied while the camera is being rotated. Not a priority right now, but the earth rotation due to time should be applied before any user rotation in your update loop. I suspect its reversed right now so that if the user moves the camera the spin of the earth due to time of day gets applied along the wrong axis.
7. Selecting what the camera is centered on works well and feels intuitive. I would default to the earth in your examples instead of first person view.
8. The earth clips completely when you rotate inside of it - I might end up looking into that approach myself.
9. When you first select the earth or moon the camera is zoomed in completely. If you don't guess that you need to zoom out you might think it is broke. I would default to a more zoomed out view.
10. Clicking into the viewport to interact with it and then hitting esc to exit out of interacting is very unintuitive feeling. For now it is workable. Zooming should be enable any time the mouse is over the viewport and the scroll wheel is changed. Moving the camera should start when left mouse button goes down and end when it goes up.
    10b. None of that applies to FPS view. Your implementation makes sense for FPS view (but again I wouldn't use that as the default).
11. When the camera is a large distance from the satellite, replace the satellite with a dot. Your current process works with one satellite model but won't scale very well. 1000 Chandra models on a low end device will be slow.
12. The satellite is obviously not to scale. This feels especially weird since the moon and sun appear to be.

Overall you are definitely moving in the right direction. I'd focus on time controls since it will also help you to be able to validate if what you are displaying on the screen is correct/matches other tools.

[ButteredFire]

I very deeply appreciate your feedback!

I'm still trying to resolve the issues surrounding the installation process, so hopefully that would be fixed by v0.3.0.

Your feedback on UX was particularly quite detailed and helpful. Thank you!

One major hurdle I have yet to overcome is the physics validation. I'm currently learning GMAT to recreate the scenarios from the test files and try to compare its output against my own, but nothing is working out just yet. It's a monumental challenge to understand and properly validate astrodynamics/orbital mechanics simulations, but I'm willing to learn. My being inundated with schoolwork doesn't help, though.

Again, thank you so much for taking your time to test v0.2.0 and craft such detailed feedback!