> Of course a city isn't a tree. A city isn't a series of roads either. And yet maps still somehow manage to be useful.
If the thesis of the article had been simply ”A city is not a tree”, I would have agreed with this criticism. It’s easy to find cases where a model is not a 100% fit.
But the purpose of the article isn’t to demonstrate how a tree is a worthless model, rather, it tries to argue how a semilattice is a better model. More than anything, it’s an argument for the importance of picking the best data structure for your problem, because if you make a poor choice, it’s going to constrain your program further down the line.
If the thesis of the article had been simply ”A city is not a tree”, I would have agreed with this criticism. It’s easy to find cases where a model is not a 100% fit.
But the purpose of the article isn’t to demonstrate how a tree is a worthless model, rather, it tries to argue how a semilattice is a better model. More than anything, it’s an argument for the importance of picking the best data structure for your problem, because if you make a poor choice, it’s going to constrain your program further down the line.