The more I think about neat apps I could make from the home office, the more and more I notice that, no matter how unweildy the concept is, distrubuted, subscription-based services are where computer OSs are headed. I don't know that Microsoft, usually pretty good about seeing good ideas overall, started in the right place. Offering Microsoft Word as a subscription doesn't make much sense at all when there are so many alternatives (previous versions of Office being the most widespread, of course) out there.
But Apple's .Mac services are closer to being on the money. Does everyone need 100 megs of virus scanned storage on the net and another email address? Well, yes and no.
What's the most aggravating part of reinstalling your operating system? That's easy -- cleaning up your hard drive and making sure you've got every file that you're going to need later backed up on CD-R or the like. Now think if your email, essential files, registration receipts, taxes, etc, existed in some limbo while you're reformatting your hard drive and could be immediately replaced after you're done. Not bad, but nothing special.
For people like myself that interact with more than one computer a day, often before I even start work or leave the house, imagine if each computer had access to these same files, regardless of platform. This, again, is nothing new, but is obviously the way things are headed.
Now let's think about this from a programmer's point of view. If I want to offer this "mobile user" services, an experience that they can pick up from whichever computer they're using, their desktop, laptop, a friend's computer, a university workstation, their work-workstation, or even a "internet bar" machine, I need someplace where I can reach these files and this information. I don't want a user to have to drag a CD around with them everywhere they go, and I really don't even want them to have to bother hooking up something like an iPod or PDA to their current machine to grab these services. I need my apps centralized but secure resources.
Anyhow, back to the point. These three needs could all be fixed by a distributed computing model (wow, I'm mangling some words there). The ability to create distributed computing resources should be part of an OS. And slowly slowly, they're finding their way there (glad my fast typing managed to get all the "there's" there correctly that time) with apps like Apples iCal. There are parts of iCal that you just can't use unless you're a .Mac user.
But of the millions of people who have OS X now, only 100 grand are using .Mac -- and I don't blame them. There's nothing I absolutely need there at the moment. So what if you can't see my iCal entries?
Apple needs to do two things. First, it needs to have more of these sort of distributed, internet-requiring features in OS X, to the point that people "learn to not be able to live without them. Then Apple needs to pull a Microsoft and integrate .Mac subscriptions into the OS. Buy OS X 10.2 and have a year of .Mac. Buy OS X 10.3 next year and get to keep playing. Otherwise, there are enough iFeatures in the iApps that require the iNet that you're outta luck -- and feel the pain.
Which brings me to the bit that's got me convinced this subscription services are the way to go and viable -- Madden 2003 (and FIFA 2002, for all that matters). There was a day when you could buy a game that you could play online and you were able to play online forever. Find another player and *poof*, you could play with the 1999 rosters until 2020 if you wanted. Now these games go to EA's master servers and make sure that game can still be played online. If your one year of online play is over, the servers say, "Forget it; buy Madden 2004". To continue to play with officially updated rosters, I seem to be willing to pay $50 a year (heck, that's even without online play, but you get the point -- another set of users are going to be upgrading to continue to use the online resource, including me, more than likely, if I ever get an ethernet port for the PS2). And that's "just a game".
If there's something compelling about operating systems that require the inet, I'm sure you'll be able to get bigger bucks outta users. Lindows is trying it for quite a bit more -- $129 yearly for click-n-install or whatever? Heck, you pay $40 a month for cell phones, why can't OS companies get a little more dough outta the yearly renewal as well? All they need is a featureset to match. Apple's on the way there, I believe.
(yes I realize many business OSs have done this with maintenance fees for a while -- I'm talking on the client side now)