Wave function collapse is a real measurable phenomena, what you describe is not measurable so just a belief, don't mix in your beliefs with actual results.
There is a different view that the wave function is just an representation of the knowledge someone has about the state of the universe or some part of it, in which case it can collapse for some and not others. On the other hand there is not really any collapse at all in this case, it is just a fancy term for updating once believe about the state of the universe or some system.
It is also important to note that a collapse is not a total collapse into a point. Measuring different things will collapse the wave function along different dimensions, only measuring everything in the universe would collapse the wave function along all dimensions into a point, or as far as the uncertainty principle permits.
There are 5 possible scenarios:
1) The wave function collapsed when the photon colides with the screen. You can calculate the time using distance/c. All the other steps are classic. You just open the box and see where the spot was.
2) When the photon colides with the screen there is no colapse. You get a superposition of two screens. A few milliseconds later when the gas inside the box hits the screen there is a collapse and ow you have a classic spot in a classic screen. Later you just open the box.
3) When the photon colides with the screen there is no colapse. You get a superposition of two screens. A few milliseconds later when the gas inside the box hit the screen, the gas gets entangled and later the walls of the box get entangled. There is no collapse until you open the box. Anyway, when you open the box everything get's classic and you see the usual result.
4) When the photon colides with the screen there is no colapse. You get a superposition of two screens. A few milliseconds later when the gas inside the box hit the screen, the gas gets entangled and later the walls of the box get entangled. There is no collapse. When you open the box your eyes and brain get entangled. There is no collapse! Anyway, each of the versions of you see the expected result because each one is entangled with a screen that has a spot as expected.
5) Shut up and calculate!
I'm a strong proponent of 5 because people pay me to calculate the distribution of the probability of the spot and not to tell weird stories.
The easy way to get the result of 5 is to use the scenario in 1, and that is what is teach, and that is essentially the Copenhagen interpretation and anyone that is not insane will use that scenario to make the calculations because it's "obvious" that all the other add negligidle corrections and are infinitely more difficult.
Here the problem is to define what is a measurement. It's "obvious" that the screen is good enough, but there is no good technical definition of what "obvious" means.
The problem is that the only non magical scenario is 4, where there is no collapse and you get entangled too. There are still many technical details to solve, but if it's real you would see the same experimental results than in 1, 2, 3 and 4, but the advantage is that there is no magic rule of collapse. In this scenario, measurement does not collapse wave functions, it just an illusion when some of the parts is big enough to get decoherence and make all the versions so different that you get (almost) no interference effect.
Anyway we are still not sure, so you can believe any of them until someone solves the problem, probably in 50 or 100 years.
Of course, a second measurement device that's completely isolated from the first measurement device can eventually interact with the first measurement device and the two will converge and share the same wave function. However, until the second measurement device interacts with the first one or with the original quantum system being observed, then from the point of view of the second measurement device, the first measurement device is in a superposition of all the possible states that it could have observed.
This can continue on and on, with a third measurement device which is entirely isolated from the second measurement device having a wavefunction that's a superposition of all the possible observations that the second measurement device will observer...