No software is free of bugs and there are always things that annoy users or drive them crazy, there is no work around. Even a gazillion bugfixes can´t satisfy the whole human beings.
Have a look at really big companies like Microsoft. They have an exact time schedule how long they will support their OSs. There is even an exact date when Windows8 won´t be supported anymore even there is no new system out yet.
It´s exactly like you say, there are things that can´t be done easily or at all with the current version of Construct, so why should I invest 100 hours to make my 50PS car a 60PS car instead of taking 200 and build a 150PS one? I can definitely understand your point, but to me it´s more than time, to get away from this "workaround-needing-prototype" (sry 
Ashley ) and develop something really serious, that can at least try to compete with some major players on the market.
People are damned prejudiced when thinking about software. In C2s case it´s perhaps something like "HTML5+Javascript is slow/not as good as native code" or something like this. Even if you do everything to improve things like this, there will always be the same rumors on general gamedesign/programing boards.
I of course don´t know, how far C3 came already or how much of the source code of c2 can be used ( I guess some cause of the import-functionality of c2 projects in R3) but marketing wise it´s not a bad decision to announce the new software. It´s also best practise in the marketing industry. Apple for example announces their product month before there is a even a picture of the product, just to keep rumors and disscussion alive and ongoing. Just have a look here for example. People seem to really freak out!
Well, just my two cents
PS: I´m absolutely with 
alvarop regarding the ability of making games with C2. People with skill can actually make the games they desire with C2, without lowering their sights. Even if it´s a visual thing game programing and design needs a lot of knowledge. No engine can do that job for you.