In reference to this recent nature article: https://www.nature.com/articles/s41567-018-0241-6
Specifically, does this warrant a new type of gate?
This does not warrant a new type of gate. When we write down quantum circuits, each 'wire' corresponds to a single qubit. However, we do not (usually) specify what technology any of these qubits is made out of. You might typically assume that they're all the same technology (e.g. solid state, photonic,...) but there is no need to do so. There are very good reasons for wanting to interface different types of qubit, particularly static and flying qubits, to take advantage of the benefits of both.
So, on your quantum circuit, you specify that you want a swap gate. This is an abstraction, and does not tell you how you're physically going to realise it. The cited paper is showing one way of implementing it when the two qubits are two specific (different) types. But it's not a new gate.
As for the significance, that comes back to why you want to have both static and flying qubits in your experiment. Flying qubits are great if you need to connect two distant components because they can travel long distances with relatively little decoherence (usually more relevant to different quantum information protocols, rather than computation specifically, but some current quantum computer designs require distributed blocks). Static qubits are great if you actually want to manipulate the quantum state, interacting with other qubits etc. For example, we often talk in quantum information about processes such as "Alice sends a qubit to Bob" and then we talk about Bob holding the particular state he's received. So Alice probably sent it using a photon because they were far apart. But if Bob wants to hold onto the state, planning on doing something with it later, he needs to transfer it onto a static qubit. Perhaps, for example, Alice and Bob want to share a Bell pair, but their quantum communication channel is noisy, so Alice will send many Bell pair halves to Bob, and they will later perform a distillation protocol. The distillation process will be highly non-linear, and probably makes more sense on static qubits.