Comment by boomlinde

Well, ideally you'd have both. When I "GET" I assume that it's GET as opposed to any of the other standard HTTP methods, not "GET" as in Indiana Jones getting the golden idol from the pedestal.

This is not only a problem with browsers preemptively requesting URLs, but also when it comes to caching. What happens when the URL I'm GETing is cached? Absolutely nothing, as far as the original server is concerned.

The current situation with browsers doing smart things to make slow websites appear fast is a bit like compiler writers doing smart things with UB in C, though. Speed a lot of things up by utilizing every undefined nook and cranny of the spec, breaking tons of legacy software that make pretty sound assumptions about how things actually work. I use a bunch of poorly designed legacy systems where GET often has intentional side effects. They break because the browser starts issuing HTTP requests long before I have finished typing an address. Let slow sites be slow and leave the speed problem to the people that should be dealing with it instead.