Games are consumed commodities, just like shoes and baked beans. Steam has become the Tesco of gaming, and while there are dangers with that kind of thing (eg the decay of game availability on the high street, just like butchers and fishmongers) there are big advantages.
In truth there are simply too many games now. We cannot hope to play them all, so we look for value. And where customers are looking for value there is room for clever marketing and pricing. I would never have bought games like Duke Nukem or Bulletstorm (which are disposable experiences), or even Civ5 (because I don't need another time sink!) unless they were reduced to silly money.
I may well never play them even though I have access to them now. But the games industry gains a little money from me, supporting developers, and I gain a big pile of games to suit every mood and inclination.
I think it's inevitable that the Steam approach will be rolled out -- reluctantly -- to all digital media everywhere. It's either that or go under, because the internet has democratised quality control. Poor products cannot survive for long, especially at full price, in an age where word of mouth can kill something dead in a matter of hours.
PS The Defence Grid example is -- for me -- an example of a project which was never worth much money. It was good fun, but they seem to have invested a lot of money tarting it up which ignored the fact that the gameplay was very, very basic but satisfying.
The core game was only ever worth a relatively small amount of money. The bling on top was a largely unnecessary distraction. Nobody played it for the plot, and the only levels I remember are the tiny number which were tough -- and therefore satisfying -- to crack. So I don't think it's a good example of anything other than poor project management decisions.