Thank you, I had this dynamic of cost-to-market in mind and you provided the perfect example.
I would see the propagation of SaaS model as advantageous to competition and not detrimental and thus beneficial to all who consume software, as it will reduce the cost to market and make profit models actually deterministic.
I work in a company whose products deal in the structural design for construction. The energies put into retail, licencing, fighting piracy etc. are very off-putting for anyone thinking of starting a software business but they seem to be essential part of selling expensive desktop software. Also, the de-facto standard licencing server solution (I'm looking at you, FlexLM) is ... not so good.
I actually only now start to believe there could emerge viable competitors to Adobe in publishing/graphics design, Microsoft in Office and Autocad in 3D design. The age of the dinosaurs is finally coming to an end. Yay!
To the average consumer it does not matter whether the software lives in the providers server or in the local substrate. It's the added value, usability and the trust in the software that matters.
Adobe's CreativeCloud pricing pisses you off? Fine, start up your own desktop publishing software company, implement, say 20% of the features you need and expand. Of course, race to the bottom is never good but a product that implements say 10% of the Creative Cloud feature set does not need to be as expensive - it's in a totally different segment - and once it has a user base and traction the company can grow, implement new features and products. When this virtualization-as-a service really kicks in it actually will be just as realistic option to start a desktop-equivalent product as making a web-app as the dynamics of propagation _will_be_the_same_ as for any web service. Good products will propagate and creativity will flourish.
Huge proportions of cost of desktop software development are about licencing, fighting bugs in different user desktop configurations, etc, non value-adding, labour intensive things that you need to provide a usable desktop product which are mostly about platform fragmentation. Suddenly you can develop software on top of _only_one_platform_ that will be more or less _stable_ in terms of your core functionality.
People, this is not a catastrophe and return to the age of the mainframe behemoths, this will be a new renaissance (... just as long as the pricing for the core technologies will be such that they do not lock out small time players... fingers crossed).
The only downside here are software patents which probably could and will be used by large companies to fight scary small competitors who will not sell but I'm optimistic their time will go away... at some point.