Good point zridling. The model does sound logical from a software marketing perspective, but equally illogical from an end-user's point of view (unless one gets sucked in by all the marketing hype).
Perhaps the hope for the future of desktop software lies in the hands of freeware and shareware developers who are actually out there to provide the best product they can. I can see a possible split between two types of shareware apps: time-based upgrade policies and free-for-life upgrades.
The way I see it, the latter will be less likely to approve of a service-based model for the same reasons they chose to give their users free updates for life; because they believe in it themselves. They usually don't have a marketing department or someone whose sole purpose is to make money telling them what to do.
It wouldn't surprise me if a few years down the road most commercial software used a service-based web model to deliver the entire app. I wonder, though, what will happen to retail software in the longrun. Will companies still package the software and distribute it commercially (and therefore probably charge the same price for the package) as well as charge a service fee?
Interestingly, this all relates quite closely to the direction PC games have been headed in.
I don't know how many of the subscribers here are gamers, but the new model sounds a lot like the one used for most MMORPGs out these days. The game itself is sold online or in stores for the same price as other non-MMORPGs, but in order to play the game one must have a monthly subscription, one's character(s) and game information is stored on the company's servers, and the game obviously cannot be played offline. A year down the road, if someone's no longer forking over the monthly payment, they still have the game content but it is completely useless without the subscription service. So what do MMORPG gamers really own?
On a related topic, games in general might be headed in a similar, though much less intimidating, model which is episode-based. Several companies (eg. Valve and Ritual to name a couple), have lately been releasing, or working on, episodic games in which the user pays a smaller (but not necessarily proportionately so) amount for each episode, which usually includes substantially less content than a full release. The selling point to gamers is that episodes are released much more often than, for instance, full sequels or new games, and that new technologies can be incorporated into the game as they become available. Gamers pay more overall to have all of the episodes, but it is partially justified in that development takes less time. While there are some large variations from a service-based model to this one, it almost seems to be the logical alternative that allows developers to get more money, and more of a constant stream of it.
My related point is that if you add a web-based distribution system to the episodes and throw in some automated billing, you've basically got a "web 2.0" game which requires one to pay regularly for the new content. The difference being that so far the episodes are entire games themselves that one can keep whether or not they are actively buying following episodes, but I wouldn't doubt that's going to change eventually, since it's ultimately the publishers who make most decisions.
My last point relating this to games are the recent "free" (or not) online games you can download that sell things from within them that players buy with real money. In some games this is in the form of "modules" that add new features for a one-time or subscription charge, and in others it's a game with a virtual economy where players pay small amounts for things like in-game clothes, houses, vehicles, weapons, etc. Basically, if you want the full experience, then the game isn't free, and isn't usually much cheaper than standard retail games, if at all.
Anyway, I'm just trying to relate the recent changes that have been going on for a while now in the gaming world to the desktop/web software world. It seems even things which wouldn't do well in a fully web 2.0 environment are capitalizing on the same underlying concept.