I will assume that the second sentence you call the following text:
But there is another doubt. It seems to me that there are some non-obvious smart solutions that could reduce the number of necessary compromises. These solutions are somewhere in the field of designing the player's interaction with the game. We just haven't invented them yet:)
Although I'm not sure about that.
I'll try to explain with an example:
Let's say that we are designing a gameplay component of a game in which the player flies from the Sun to some point near the Earth, then flies around the earth and moves towards the Moon.
The simplest thing we can do is to move the camera at a constant speed along a given trajectory.
Here we meet the problem of boring proportions of reality. The first 99% of the way from the Sun to the Earth, the Earth will be just a pixel on the screen. It turns out that the player is forced to observe for one minute an almost static picture of approaching an incomprehensible blue pixel for the sake of half a second of beautiful movement relative to the 3D model of the earth. A similar situation will occur when flying to the moon.
We can increase the fun by changing the proportions between distances and sizes. But we can not change the proportions, but make the speed of the player's camera variable. The speed will be huge at the beginning of the journey, then it will gradually decrease. For example, we can use the formula:
(current velocity) = (current dist to target) * (const coef)
You can also use smarter formulas that provide a beautiful smooth deceleration.
The essence of the example is that the problem is solved by a small modification of the gameplay. I agree that the solution proposed in my example is not good in itself. Better solutions certainly exist. The question is, are there good enough solutions? That is, are there such solutions when we do not compromise with realism, but still get a funny result.
I agree with you about the target audience. Scientific correctness is relevant for a small niche.