Of course all of the above is good advice.
I think the issue of perspective is also part of what Karsaa might find troubling about the recent picture of the church. Small changes in angle of the camera make big differences in perspective with wide angle lenses.
I would like to pose a further question on the issue of managing vertical perspective. For purposes of this discussion assume I do not own a TS lens, since I don’t!
To take a picture of a vertical object ( like the light houses already mentioned or the tower attached) one could get closer, angle the camera upward to fill more of the frame and then correct perspective in software. There is some loss of detail in the areas that the software expands. That is what I did in the example attached.
tower by Danny Watson, on Flickr Tower 2 by Danny Watson, on Flickr
Alternatively one could move farther back to get the camera horizontal, fill less of the frame with the object and then crop/ enlarge. That loses some detail, too.
Either one carried to extreme does not seem optimal.
How do you experienced guys decide how close / how much vertical angle to have in the original image?