That's a very good point - about buying on the strength of magine reports.
How the magazines 'test' new hardware/software is very important. I'm guessing that most of the 'journo's' don't have the time to put e.g. a new piece of software through its paces in a 'real' world situation.
They probably just hook it up to their own little 'test bed' (read that as their own set of sub test tools) and check that the unit performs "as good as" or "better than" the last piece of kit thy tested...
How many magazine reporters actually bother to load up e.g. some cubase variation (doesn't have to be Cubase) and check a VSTi as just that, a VSTi running along other VSTi's in a song they've composed, that say 'drives' the hardware in a more 'real world' scenario.
A good example would be, Not the fastest dual core PC out there, running sequencer software, loaded with audio, midi tracks and 3 x VSTi's being driven by 3 x midi tracks.
Stand alone is fine, but most of us will want to use the harmonic content of software synths in our own creations (VSTi/RTAS etc) and it is here that we will want to know how they perform.
I love the sound of the Jupiter 8 V - in standalone (that's the only way I can run it :-(
If the magazines had bothered to correctly report this behaviour, I'm betting that the Arturia guys would've plugged this issue nice and early (or maybe even re-developed - if the architecture is to blame).
I guess ultimately, we should all buy on the strength of the demo alone.
Or maybe ask that Arturia allow some of us to Beta test 'all' their software - hmmmm, that would be nice