I am curious, why not use XPMode for your 12-year-old LOB application?

Nobody could possibly want one of these. Especially not a college student...

The company would have to take a fairly substantial risk in not releasing it on Windows, though, for that to happen. Otherwise people will just get the game on Windows. To bootstrap a platform in the face of entrenched competition, quality exclusives are necessary.

2TB NAS? Preferably one that is cloud-storage aware (or has its own clients for tablets/phones) to be broadly accessible even when not on LAN? (I guess I could settle for SkyDrive's Fetch feature, but would rather not have to go through that route).

Even a general-purpose computing device can be sold in a configuration that by default "just plugs in and works."

We (the techies) are not incompetent to be unable to switch to a more productive environment - just that environment does not need to be in everyone's face all the time. Rather than think of this as the PC being "dumbed down," think of it as a call to action to make all the other "appliance" computers be more full-features, but without the historical complexity of PCs, before we knew how to make them easy to use.

Er... not quite. Yes, clustered distributed systems exist, and are growing much faster than mainframes, but there is still a lot of stuff that is done on mainframes, if only because it is a lot more convenient to have a really beefy box for that large, mission-critical multi-user database/application, without having to figure out how to partition it and without being subject to CAP.

Distributed clusters are generally better for non-real-time processing. Real-time HPC stuff is still the realm of supercomputers (not to be confused with clusters, even though both generally tend to run Linux or some other embedded POSIX OS)

And they will still pay Microsoft royalties for Android. From the B&N press release (link from TFA):

Above emphasis mine.

Disagree. It is possible to construct a CryptoFunction C(t, k) such that exists k_0 != k where D(t, k_0) = t_0 != t, with t_0 "plaintext" that appears correct but is, in fact, wrong, by inserting specifically tailored "trash" data.

I would actually advise against installing those program to [C:\]Users\ instead create a parallel [C:\]Program Files[tag]\ folder with proper permissions set. Since those programs will still want to screw around with the registry as they get installed, (and all the uninstall information is in the global registry), it makes sense to let them be wiped and require a reinstall while preserving user data. With a manually created parallel Program Files folder (I usually use "(ul)" to tag it) you get all the advantages (and disadvantages for that matter) of not having to deal with UAC for these applications, but do not pollute the [C:\]Users\ folder.

As far as I can tell, all that needs to be done is to implement a system-wide Windows Media Foundation component to render Theora, and IE will pick it up if it sees a Theora video.The MSDN docs there are really bad, though; they will tell you how to build out the component itself, but nothing about how to get the system to pick it up. However, what I guess you can do is run regmon while installing the WebM MF components (which is all the installer does) and see how it publishes its MediaSource implementation out. Given that there are macros to create media type guids for arbitrary FOURCCs, it should not be impossible to register arbitrary filters to be used by this.

If you can stick a meta-tag into the HTML generated by the application there is a one-line fix that will trigger IE7 rendering/DOM mode in later versions of IE for backwards compatibility. Here is a good overview of how it works, and how to manually trigger it, overriding doctype sniffing.

Actually, Google provides a WebM fiter (tech preview) that IE can use to display WebM video. The Theora folks could do Theora as well that way.

Actually, if you configured it to work like it did in 95, then you should have no problems with it taking over your screen