<troll_mode> What's the big deal? It's still a fly! Evilution won't be proved until they turn a fly into an aardvark or something! </troll_mode>
By the end of 1984, 62 genera and 2,822 described species of this family were known worldwide, and at least 4,000 species were predicted to exist (Wheeler 1986). This catalog records over 1,000 species in 65 genera and subgenera from the Australasian/Oceanian Regions.All of these forms are still in the one big family. When something evolves all the way out of that family, we get to wake up fzx12345.
Except the way cladistics works, we probably never get to wake up fzx12345. That's because when you can't describe Drosophilidae with just species, genera, and maybe the occasional recourse to sub-genera, you don't create a new family alongside the existing Drosophilidae. IOW, if something is known to have originated within Drosophilidae, we try to preserve that relationship forever.
Thus what you do is you bump Drosophilidae up a notch and now it's an order rather than a family. Nothing ever wanders "out" of it.
Of course, the order that used to contain Drosophilidae bumps up a notch, too. Maybe you just call it a super-order to contain the ripple effect, or maybe you call it a class. If you do the latter, now Insecta becomes a super-class or something. For sure, you don't create a new phylum unless there's a whole new body plan, so you have to get inventive somewhere along the way up the hierarchy.