Regarding engines, if you're buying in something 3rd party you'll have to accept limitations of some sort. Be it in asset management, visual fidelity, scale, etc. there'll be *something*. CIG chose CryEngine because it is visually stunning, and they've shown with what they've delivered so far what a space game on modern hardware can achieve visually. Unfortunately, the compromises they had to make were with broken netcode and limited map sizes. Upon learning of these limitations, they would have had two choices: 1) pick a new engine and hope it's better, and sink the costs of converting their assets and processes to the new engine; 2) modify the existing engine to do what they want. They went for the second option, and to be frank I don't think it was the wrong choice, especially as they were able to hire in specialist expertise to do the work for them. I may not like how their ships handle, but that isn't really an engine limitation as much as a design choice.
Frontier were in the perfect position to develop ED, in that they already had an engine that had been built with an eye to a potential Elite successor. While I'm sure they've had to do a *lot* of work under the hood, particularly around the audio and networking side of things, they already had resource onboard to do that work. I would say that COBRA's visuals aren't on par with CryEngine yet, and that's really where Frontier have had to compromise.
From the FPS vs. space flight perspective, CIG effectively get FPS for free (though I'm sure the zero-g stuff was an interesting challenge even then) and had to put a lot of work into space flight. For Frontier it's likely to be the other way around; while we know that COBRA has at some point had FPS features implemented (The Outsider) it's still likely to take a lot of effort to implement. It's certainly going to be interesting to see how the two games develop over time and particularly whether Frontier and CIG can keep their respective engines contemporary with improvements in PC hardware.