Hi Pmont,
A lot of comment on there!
As far as building an app that you expect to live for 5, 10, 20 years... depends on what you mean by "live". Seems like many people discussing this in the context of modern UI technologies are treating it as if a technology is "dead" as soon as the last planned version of it ships. Silverlight is not dead, there are just no planned future versions that anyone knows about at this point. It is still just as viable to build a new Silverlight app today as it was before Silverlight 5 shipped. Just as it is still viable to build a new Windows Forms app today as long as you understand you are not going to get any new features. I think it is reasonable to expect a 5-10 year lifetime out of any of the current UI technology choices. I think it may be naive to expect much beyond that without some degree of rewrite - if not for technology sake, for your user's sake. Can anyone really say that their current business model and use cases will be mostly the same more than a decade from now? I doubt it.
But it is the world we are living in now - we want rapid innovation from the software platforms, but we somehow want it to be stable and unchanging for the long term. That is just an unsolvable conundrum in my book.
I agree with your perspective on what a LOB app is - to me it is any app built by a business to support what they do as a business, regardless of whether it is consumer facing or for internal employees. But the latter case aligns a little better with the traditional concept of LOB.
Your point on the growing expectations of users is I think the most important one of why we think LOB guidance for WinRT is needed now. People will want the same fast and fluid experience for their business apps, especially if given touch devices or screens, and they are only going to get that from a Windows Store app. But the simple fact is WinRT at this point in time has not matured to address the full range of business needs that go outside of the sandbox. Microsoft has said very clearly that Windows 8 is a consumer focused release. I think they will get to those things in future releases/OS's, but there is nothing p&p can do to show how to make the platform do something it was specifically designed not to do.
So the Kona project that is focused on LOB guidance for Windows Store apps is really focused on trying to show how you can eek out as good of a business app as possible based on what the platform currently supports, particularly how to structure it for longer term maintainability and testability, and not so much on how to go outside the bounds of what WinRT says you can do.
p&p's job (my words, not theirs) is to provide guidance on how to best use the current platforms, not how to hack them into doing things they are not designed to do. If you don't like the design, you need to try to get that feedback heard by the product teams that produce that platform (Windows Division in this case), for which the Connect site is unfortunately one of the only paths to doing so at this point.
I too sincerely hope Windows Division will open up a bit and operate more like DevDiv and the Windows Azure and ASP.NET teams in particular currently do. But neither I nor p&p have much if any sway in making that happen.