There's nothing wrong with that until the first fault takes place. Then, you've greatly raised the probability of the system trying to kill you.
With a grounded system, accidental ground connections do one of two things: (1) not make a lot of difference, because the fault is in the grounding part or (2) trip off the system until the problem is fixed.
With an isolated system, it's much less likely that a fault of e.g. the hot wire to actual-water-pipe-ground will trip the system off. That seems quite convenient: "We're still running!" But, in many cases, that fault will now have put what should be the ground and neutral connectors at service voltage. Again, not yet a problem. All's fine until you get an insulation fault in a kitchen appliance or a hair dryer, and now their cases (which in a grounded system would have been at ground) are hot.
TLDR: The first fault might be invisible and feel "free", but the second one can set the system up to be very dangerous.
You might be under the misconception that grounded equipment is primarily meant to protect people. Good grounds do help with that, but the real purpose is to force the system to trip off when it faults; that's what protects people and property. If you mess with that by not having a proper grounding system, you're not getting that "There's a fault here!" feedback.
Floating/isolated systems are sometimes needed, but done properly they have both equipment and people routinely (i.e. often) checking for faults and taking the system out of service when one is found.
Finally, a comment on the difference you try to draw between "legal/regulatory/standard requirements" and "technical/safety reason(s) or benefit(s)": there really is no such distinction, at least for national standards. The US National Electrical Code (and, as far as I know, similar national and international codes elsewhere) has been professionally developed by electrical engineers, practicing electricians, fire protection experts and other technologists. You shouldn't underestimate the amount of technical and safety knowledge that has gone into those "legal/regulatory/standard requirements". It's an amateur's mistake to believe that one can anticipate everything that can go wrong, every situation that will arise. Unanticipated things still do happen, and as a profession we bring those good and bad experiences to future practice through the codes.
There's a nice discussion of the solar PV codes here. It's in the context of the 2012 NEC, but it's a good starting point. From there, this article will help you understand the changes in the 2017 NEC.