Using GET allows the browser to cache the server responses and therefore reduces the load on the server and improves the loading times (for every client after they have already loaded the page once).
Ideally, POST should only be used for big queries. How big is not entirely clear, as browsers and servers do not accept the same limits. After a quick search on the Web, it seems like 4000 bytes should almost always be OK (except with IE), but limitations might occur beyond that size.
Therefore, we need to use GET whenever possible, but still use POST whenever the payload exceeds a set point (for example 4000 bytes as mentioned above).
Good first task: it's a conditional to edit on github, then to git pull the code on the server. Result will be barely visible, yet positive.