wyoung76: the main problem I have with the authors point of view is that of a Modern World perspective. As evidence that this future is still many a generation away from becoming reality, we need only look at the Third World countries and witness the total lack of infrastructure in supporting such a society of high bandwidth and low local maintenance computing. The local computer is a fast, simple, and easy way of getting the required (or needed/desired) computing power to the people in poorer nations without worrying about the HUGE commitment in upgrading or installing the infrastructure that we modern nations are beginning to take for granted.
Doc Ruby: The real trend is mobile devices, DRM, and cheap bandwidth to home servers at local centers of always-on P2P networks. The huge mass market of less sophisticated/tolerant users, and the peripheral attention offered by personal mobile devices mean the devices will be multimedia terminals with wireless networking. The media industry orientation towards DRM means they’ll give away mobiles at a loss to sell their more scalable/profitable media products, while ensuring the terminals can’t copy the media objects. While the whole network will become much more complex under the hood, the market will demand that it all “just works”, like TV (IOW, when it doesn’t work, there’s nothing you can do about it but wait). That’s why Microsoft is evolving into a media company (games, interactive “TV”), enforcing the consumption of their lower quality products by perpetuating the applications that they prefer/require to “play”. So we’re going in the direction predicted by this story, but along the way the changes will be much different.
neurocutie: despite increasing bandwidth out to the Internet as a compelling force, equally powerful trends suggested the continued importance and popularity of the home PC. Most of these trends can be summed up as needing even higher bandwidth locally, as well as needing specific interfacing of other devices, both of which aren’t likely to be reasonably handled by some form of thin client. For example, all the reasons to burn personalized CDs or DVDs. It is not likely that burning CDs or DVDs would happen straight over the Internet without some kind of fast local store (i.e. hard disk). Another is interfacing digital and video cameras and editing those results. Again it doesn’t seem reasonable to build a thin client to interface these device just to ship the many gigs of data (particularly video) out over the Internet to a remote fileserver and, worse, to perform editing against the remote fileserver — these applications, popular on the home front, pretty much dictate a home PC-like architecture with fast, large local file store
Craig Maloney: I think the argument for a more service-based PC has some major issues to get around: First, there needs to be some receiver machine at the home end. A reasonable computer can be had for around $500 nowadays. Unless this subscriber machine can be had for less than $200, there is no incentive to move to this model. Second, nothing is free. This service will be a subscription-based service. I think it would have had some bearing had people not been burned by subscriptions from other companies. Witness the cable companies and TiVo and how they’ve handled their subscriptions. Witness the cellphone subscriptions. Paying outrageous rates for using a computer won’t succeed if there is no conomic reason to do so. People will sooner purchase Macintoshes. Thirdly, there is the issue of control. You’re dealing with people’s data, and their private information. I will never relinquish control of my checkbook, nor my family pictures, nor anything else like that. Some people may be amenable to this, but many will not. The computer is a multimedia device now, and people have scads of personal data on their computers. It’ll take a very convincing argument, and a company with a reputation for integrity to wrestle away that desire for control. The PC as we know it will change, but I see that change moving more to a home entertainment/personal network than a service based machine.
Dutky: The solution to the increasing administrative burden on computer users is not hire someone to do the administration: instead, we need computers that actually reduce amount of administration required or make the task of administration markedly easier. This is what personal computers did 40 years ago, and it can be done again.
In the next two columns, we will summarise the argument for and against centralised computing. The final two columns in this series will then discuss the notion of centralised computing in the context of emerging markets and what Microsoft should do.
Tomorrow: The Arguments For Centralised Computing
TECH TALK Microsoft, Bandwidth and Centralised Computing+T