Again, that entire explanation is not wrong or disagreeable, but it doesn't speak to the definition of persistent that is important here.
A column can be either "Calculated" or "Persistent" - that's the intended use, but since this causes headaches, I've devised a new plan.
What I'll do here to remove the ambiguity from the argument, is retract the use of the words "persistent" and "fleeting" in all the above posts and replacing it with the word "Scoriginy" and "schponkish". These words do not exist yet in any language, I checked, so nobody can argue with me about their meaning.
Here is the excerpts form the great Dictionary of Ryan - aka The "Ryanomicon":
(I say "the" excerpts rather than "some" excerpts because so far they are the only two)
```
Definition:
Schponkish - Adj.
def 1. - To exist at the mercy of its peers. To their whims, even
though said existence may seem very real and persistent,
it can be altered in an instance by peers, through no fault
and/or doing of the object itself or its creator or other
authority - though those may change it too.
Scoriginous - Adj.
def 1. - To exist firmly and by design, how it was made to be. Can
only be altered or removed by its creator/originator or
proper authority. Its peers holds no sway over its
existence.
```
So now I can say with confidence: A column can either be
- a normal Scoriginous data column, or
- a calculated Schponkish column.
While the normal scoriginous columns are ALWAYS persistent, the Calculated columns may be persisted through special addition of the "STORED" directive, but they will still always be very schponkish and change/update the moment any peer within their expressions change, regardless of their "STORED" status.
I'm hoping you will find this a more agreeable statement than the previous. :)