Take the GIMP. It still isn't really compatible enough to be a drop-in replacement for Photoshop. It still has limitations on basic stuff like colour handling and typography. It still has a name that sounds like a sophomore's dirty joke. If all you need is a basic image retouching application for photos on your web site, it's fine. For full professional use, it's not even close.
The same arguments applies to almost all of the "big name" FOSS competitors to dominant proprietary applications from the likes of Microsoft and Adobe: data portability isn't quite good enough, features are missing or underpowered, and usability is often inferior. They're fine as free substitutes and good enough for basics (which for many people is sufficient) but the "high" prices of the professional applications are justified in a heartbeat if you have real work to do as a power user.
What I don't understand is why FOSS continues to be dominated by cheap knock-offs in these fields. One man wrote TeX, albeit one exceptionally talented and driven man, and to this day a whole industry uses it in preference to anything else because it does what it does differently and much better than the mainstream products. There's still plenty its users would like to improve, but then that's true of Microsoft Office and Adobe Creative Suite too, and they have barely changed in years despite the vast resources available to their development teams.
In short, if the FOSS world wants to compete in the market for mainstream creative applications, it needs to start innovating and change the rules so the presumption isn't that the FOSS has to (try to) match incumbent proprietary products on their own terms. The question shouldn't be whether you can cycle between .docx and .odt without messing up your formatting or whether you can translate an AI file to an Inkscape-friendly SVG and back without losing details. The question should be why you'd want to in the first place.