I take the middle road. It's always good to have perspective on things, but it's also good to have quality and ideals to strive for.
There's a difference between subtle issues that are difficult and costly to revise, and straight-out "why didn't you at least change elevator_trim_effectiveness from default 1.0 to 6.0 in the aircraft.cfg before release?!"-type stuff.
I think it's the latter those of us whose lives aren't depending upon complete accuracy get frustrated by; just a little more attention to detail, or even an attempt to trim the Staggerwing, might have caught that. Or, alternately, one sentence in the docs saying (if true) that the Staggerwing has a completely ineffective trim for elevators.
That's the kind of stuff often missed that _someone_, _somewhere_ in the process could be responsible for reviewing before release? Or is it more complicated than that?
Alternate perspective: release it with such issues so the FS community will discover and comment and fix themselves, to their own satisfaction, make that part of "the game."
Then there are the maddening little things, case in point: the (few?) of us who purchased the Nemeth Designs Expansion Pack for the AS350 (beatifully modelled visually, btw! and lots of fun to fly!) have -- for those who purchased it upon release -- been waiting over a YEAR for a stupid VC Rain effect that is apparently built into the .mdl file.....??? and that produces epileptic fits since one frame goes black as the animation cycles, for any kind of response from ND. What's not clear is how complicated it is to resolve such an issue. Is it a simple edit of an animation loop and recompile? Did the .mdl file go missing in Hungary? Are there laws in Hungary preventing AS350s from flying in the rain, so that this was an intentional reminder? Is the denial of the issue in initial response the result of a true, "works fine on MY system" situation, or simply lack of attention? The animation only starts epileptic-seizure-flickering when the helicopter starts moving forward in the rain -- how difficult would that have been to catch before release?
I just deal with the above scenario myself by saying, oops, AS350 Expansion choppers, they melt in the rain, for VFR on Sunny Days Only. Just like any FSX technique.
But it's really a philosophical issue: we can't know "outside" what level of complexity or what nature of situation allows resolution or fails to lead to resolution of problems of various levels. I have, in my relatively naive exploration of payware, learned pretty quickly that often what look like "simple" issues really aren't, and alternately, am shocked when some super-simple issues get passed through.
Not just payware for FSX, either; this is an endemic issue in software in general. Programmers and graphics artists can be a moody lot and just take the piss on their audience, too, even.
There's the other "audience bug" of what we used to call "grognardism" in the turn-based historical wargaming genre many years ago; the more "expert" players of the genre became, the more increasingly insane the noodling on details became, making life hell for anyone who wanted to attempt to make a wargame. At some point, opening up the ability to tweak variables to the user's content provided some solution, but then the element of surprise was often lost. So, this is a persistent problem that, hopefully, Twitter will solve (increasing Airheadism worldwide) some day. It is truly a challenge for software devs, as the more dedicated the audience, the more anxious they become about mis-placed nose-hairs on the pilot modelled in the cockpit.
So it's best to remain calm, but continue to push patiently for quality and resolution, and then to just happily scratch your head when it doesn't happen. And over time, as these forums prove, the developers' quality or lack thereof will gain or lose audience.
