Improve docs and add custom execution plan example

[gabotechs]

This is a docs-specific PR that:

• Improves the docs by revisiting the wording and the terminology used
• Adds an example for distributing queries containing custom execution plans

[JSOD11]

Looks good to me, just made a few wording suggestions in the docs.

I like the code example you came up with, I think it highlights the behavior of the TaskEstimator well. I ran your examples and they work for me.

[JSOD11]

Is NetworkCoalesceExec at the top not considered a stage? Do stage indices start at 0 or 1? It could possibly be helpful to make that explicit by writing e.g. Stage i where i is the stage index rather than just stage.

[gabotechs]

Is NetworkCoalesceExec at the top not considered a stage?

I don't think so. That part of the plan gets executed locally in a non-distributed context, so we usually talk about it as the "head" of the plan. It might be nice to choose a specific term for that though.

[gabotechs]

For stage enumeration, I'd not over index to much about it in this drawing, as I wanted to land it as clean as possible. There's not a strong enumeration order requirement, fwiw stage numbers could be just random numbers and it would work the same, so I think it's fine to not represent it in this drawing

[NGA-TRAN]

@JSOD11 has great comments. I only wonder whether we can also display the plans with the examples