Then it'd have to be to-scale. That's not good if you only care about the order.

It would technically be wrong because the dates would have no appropriate scale. The difference between each node could be wildly different and also irrelevant to the solution space this virtual representation is trying to convey. Egg before chicken, check

If you actually wanted a phylogenetic tree to scale you would end up with a huuuge tree that has many more branches because it obviously is not as simple as depicted above. Take birds for example: There are all the dinosaurs in between that weren't birds but have their own branches. It has actually been a tough question where to draw the line between dinosaurs and birds (there is a whole article on wiki). And if you have any paraphyletic groups in your tree it gets even messier! If you are already displaying other groups at the subfamily level, you should then display all groups at this level.

All this is to say that the level of detail contained in a phylogenetic tree (or any graph for that matter) is highly dependent on the information you want to convey. Ideally you should draw it as detailed as necessary and as simplified as possible. In this case, we get all the information that is necessary but are not overwhelmed with facts that if you draw Squamata (lizards and snakes), you would also have to draw Rhynchocephalia (monotypic order) in.