On the question as to whether RT will 'take off' or not is dependent on many - potentially infinite - variables. But the thing that drew me to the RT platform was what I considered - and still do - its radical forward-thinking ness. Now, with the caveat that all this could just be a reflection of my own biases (and preferences), I will say this - while it may not be the case that the specific RT platform that we know today will survive, something like it is most likely to be the dominant OS (or, perhaps more appropriately, application underwater) within the next 5 years or so. Why? Because, as Jeff pointed out in a post above, the concept of the wholistic OS (like Windows, MacOS, Linux etc.) is dying a slow death and the reason for that is because the macro socio-technical culture is in a midst of a radical transformation.
What precisely does this mean? Mainly, this: Information Technology has now become an infrastructure - much more so than in the 1980s when the current dominant model of local installations of operating systems was conceived (at least in the consumer space). If you guys recall, there was a point in time when Thin Clients were a rage (this would have been around the late 1990s and early 2000s) and there was talk of screen scraping to reducing processing times and lightening the end-users headaches involved in maintaining operating systems that were and remain complex entities. Of course, that project died early probably because the technology was not mature. But today, it is. And there is emergence of cloud computing (arguably an old model, but now re-wrapped in a new garb and underwritten by an incrementally-becoming-ubiquitous wireless network architecture). Of course it could be argued that this is a First World perspective and yes it is. But a recent trip (within the last week) to Asia convinces me that what we in the West think in terms of WIFI networks is being leapfrogged by mobile data networks on which much of Asia works. Naturally, given the still-woeful nature of the infrastructure there, speeds and consequent capabilities are sub-optimal, but the numbers and momentum cannot be denied. And, what is the basic device that people there (as is the case in large parts of Africa) using? Mobile phones - the majority still use enhanced feature phones, but low-grade (by which I mean utilitarian and without the bells and whistles that we are normally used to in the West) smart phones are growing in numbers (unfortunately, Android has a very big slice of this market). Nevertheless, the point that I trying to make is that for these people (who constitute the largest and growing number of network exploiters - and I use this term advisedly), the fundamental computational device is the mobile phone. This does have an installed OS, but it is one that rarely, if ever, requires end-user intervention. So, what can be a viable upgrade model for these people?
My answer to this is Thin Clients where the OS (or at least the functionality of the OS) is provided as a service. And who are at the forefront of this emerging trend? Google (with their Chrome OS, which is made attractive by the likes of Acer who release USD 200-300 devices) and Microsoft (with the RT platform - notice how the desktop which is where the complexity of the Windows OS is most evident is at a discount in the RT scheme of things - and, of all things, Office 365).
So, at least in future-oriented terms, I think RT does have a future. But much depends on (1) how MS is able to play this - the preparatory phase will involved releasing Office Gemini and getting Windows Phones to use RT as an moderating system; (2) If the future is all about an OS being a service, then MS should seriously think about slashing the costs of the OS. They can do this in a number of ways. Arguably, the best would be make the OS and the application layer free (or available at a very minimal cost) while focusing their revenue streams on applications. For this they need to create a cult of MS developers who, if necessary, need to be bankrolled by MS. If MS could write off USD 900 Million on the first edition of the Surface, they can probably spend two to three times that amount to capture and retain developer loyalty; (3) MS needs to actively reimagine hope they engage with the enterprise sector. At the front-end, MS needs to hand-hold enterprises to develop LOB applications and to provide a cascading cadence of support to them. In other words, MS cannot involved another XP-like situation, which is arguably developing now Win 7. They need to avoid this like the plague.
What then does this mean for RT-centric application developers? Simply this: Engaging with the RT platform will mean betting on the future - not simply on the platform, but more importantly on the global socio-technical future. In the immediate term, I think developers will wait till Win RT and Win Phone 8 OS merge with Win Phones operating on the RT platform. Personally speaking based simply on anecdotal evidence (which really means nothing more than opinions), the inflection point will when that merger takes place. From that point on there will be only two dominant players - Google and Microsoft (aside from any new comers who may show up - particularly from the Eastern part of the world). The most frenetic activity will them shift to application developers who will fight to reimagine the informational needs of the future and the ways and means to service them.