I often do not explain things this well. Each paragraph is clear and helpful. I was glad to see this when I got up and poured coffee this morning.
TheTechNoir said:
The rest of the universe was also compact. I think what he is saying is that it is believed that our observable universe used to be that small (watermellon, atom, etc. something small by a human scale) but that doesn't assign anything special to our oberservable universe (the horizon of which is constantly seeing less and less). Using the omnicient thought, if you could see another section of the universe that were spatially far removed from our OU but the same size, it too would have been the watermellon/atom size at the time of the BB. And so too would every other equivalantly sized section of the universe.
There are some lesser accepted theories that may indicate however that, that is not correct such as chaotic inflation. I think that would make what I said incorrect anyway, but I don't know much of chaotic inflation to say for sure, if you would like to read up on it go ahead. And that isn't to say what marcus said was in anyway wrong, it was not wrong at all.
EDIT: Also, while not popularily believed to be the shape of our universe, it is not impossible for the U to be finite and flat (and no it will not have a boundary - think of a pacman screen what happens when you keep going in one direction - or for a much more satisfying explanation search for toridial universe)
I think from your first paragraph the reader can see how the universe could have been infinite already at the start of expansion (an infinite number of finite watermellons?).
The fact that our OU began as a small finite vol does not imply that the entire U did.
We simply do not know at this stage which model is closer to the truth, spatial infinite or spatial finite ("closed", "compact") like for example hypersphere.
I think your second paragraph is a good point to make! Working cosmologists use a simple model U (called LCDM based on Friedmann equation) that does not involve a multiverse of different random inflations with possibly different laws of physics. The mainstream picture is just one U, governed by Einstein GR eqn., assumed approximately uniform, with no boundary. (and there are two main cases, spatial finite and spatial infinite, because our data is not yet good enough to determine the curvature accurately enough to distinguish.)
And like you said there ARE these other models like "chaotic inflation". They are not supported by data, they are speculations, they are not models that mainstream cosmologists use, but they interest people. As multiverse scenarios they appeal to some string theorists, for example. Your second paragraph strikes the right tone. It acknowledges these interesting scenarios which people think about but do not use for fitting observational data to.
Your third paragraph, the EDIT, also strikes just the right tone. There are these pacman geometries which the U could have. They look locally flat. As far as we can tell they would look like the flat infinite case. But if you could travel very very far (outracing expansion) you would eventually come round, like a pacman. Maybe life is too short to constantly be acknowledging the different possible geometries the U could have.
Pragmatically, you need a mathematical model to fit data to. Two or three cases of LCDM seem adequate based on measurements of largescale spatial curvature (pos, zero, neg).
So far signs of pacman (torus = donut type) have not appeared so it gets ignored or downplayed.
All I think I'm doing is amplifying what you said. It is either that or remain silent.
Thanks.
