And that's exactly my point. How is EESchema (or any other schematic-design program) supposed to know that wires that cross are supposed to be connected or not? An advocate of "no junction dots" might say, "well, the user can just draw the wire, click on the orthogonal wire and then continue drawing the wire to the next endpoint." That doesn't help when dragging wires, that doesn't help when deleting wires, that doesn't help when the connection is made when one isn't expected.
From a simple usability perspective, the junction dot is an unambiguous indicator that "these wires are connected." The netlister doesn't have to guess whether the wires are connected and the person reading the schematic doesn't have to guess. The fewer assumptions that a program needs to make, the better.
Just because @Sprig has had a bit of experience with schematics drawn one way, that doesn't mean that his way is now the new standard. As I noted earlier, every schematic package I've worked with in my career has used junction dots to indicate wire connections.