If you pay any attention to the IT news, you'll realize there's one major theme for this week: desktop virtualization\u2014a theme fueled more by \n\ncoincidental self-interest than technology. Maybe you've seen some of the articles all around this week:\n\n\n VMware pitches \n\nvirtual office desktops 101\n KVM Sponsor Premiers Desktop Virtualization\n HP Client Virtualization Announcement\n Citrix Rides A Wave Of Virtualization Success\n HP \n\nconfesses love for Citrix with mobile thin client\nSure, there's plenty of news. But, as with much news in the computer business, it focuses on what vendors are \n\nannouncing&mdasusually technology they've been working on for months or years and that they've tried to hold back and \n\nrelease at a time that's most advantageous for them. So when you see the headline "VMware Makes Thin Client Moves," it's not because someone at VMware has just discovered the existence of \n\nthe thin client. That news came out of VMware this week because Citrix is having its own conference this week and is generating headlines by \n\nannouncing partnerships and technology that would be just as accurate and\/or useful to customers they had been announced last week, or next.\n\nVMware and the other companies that get in on the fun (See"Neocleus - A different take on desktop virtualization") by announcing their own stuff are just stealing a \n\nlittle of the spotlight. That doesn't mean desktop virtualization technology's no good; it's been around a lot longer than server or storage \n\nvirtualization, in fact, and has been successful to a limited degree in every generation of computing almost since Grace Hopper chased the moth out of the Mark II.It's gotten better with every generation, too, to the extent that people who've never heard of desktop virtualization, \n\nremote sessions, thin clients or shared-host, terminal-based computing are happily signing up for Software As a Service \n\n(SaaS) offerings like Google Apps and Salesforce.com, remote desktop-session players like GoToMyPC or online storage and email services. \n\nNone of those is a complete desktop-virtualization solution; they don't make it easier to install, secure or manage vast \n\narmies of personal computers or reduce the cost and hardware requirements of the PCs themselves. They only slice off small pieces of things the end users themselves would like to do, and provide it on a shared-host, \n\nremote-session basis. But they work; and they attract end users who, for once, are on the winning end of the something-for-nothing calculation \n\nthat usually benefits only IT managers and the vendors who sell them things. VMware and the rest of them are selling a much more complete picture of virtualization, cost-savings and simplified \n\nmanagement. (See "Desktop Virtualization: Inside VMware's Strategy and Newest Plans" for more detail.) They're excited because the number of desktops out there is so much \n\nhigher than the number of servers that they figure if they can virtualize even a fraction of those, it will make them rich. \n\nThey're right, too. Most of the PC universe is pure greenfield for virtualization vendors. It will stay that way because \n\nend users, for the most part, don't want to be virtualized. They don't mind if someone else manages their line-of-business \n\napplications and makes their storage available without too much hunting, and keeps the network attached and running. But they don't want to have to dial in and stay connected while they're on the road, or working from home, or even be \n\nunable to load MP3s on their hard drive and play music in the office while they work (on headphones, of course). PCs are like desks; everyone knows they belong to the company, but mess with your workers' desks and the personal stuff in \n\nthem and you're in for a world of trouble. That's why desktop virtualization will be successful on only a limited basis. It will remain attractive to the same kinds \n\nof companies for the same kinds of roles that use remote sessions and thin clients and virtual desktops and all the other \n\nways to cheaply support PCs used by more than one person for relatively generic functions. Call center operators, bank tellers, data-entry clerks, retail clerks and all the other non-knowledge workers who need to \n\neither plug information into a computer or get it back out, are good targets for virtual desktops. Anyone who has to spend hours and hours in front of a computer and is expected to come up with answers to problems using \n\ntheir own imagination and experience, or use their brain as anything but a conduit for information from customer to computer, \n\nwants some control over their work environment. True, VMware (and probably others) are now making arrangements to allow virtual desktop users to work offline. But that's \n\njust an extension of the partial-VM that I mentioned before. As they have with gMail and Salesforce and all the other \n\nselective-virtualization services, end users will accept the virtualization of some aspects of their working environment. And \n\nthat level of virtualization will save some money and effort for their employers. It will be much easier to update \n\napplications that users access only through remote session on a secure server, for example. Web front ends became popular interfaces for business applications for the same reason; users didn't mind going through a \n\nbrowser, and all the code sat on the server. No more wandering around userland with a DVD updating everyone's fat client \n\nevery three months. Virtual desktop infrastructures will be more flexible than previous iterations, and probably more popular. They'll allow \n\nIT to leave all the "important" stuff in the data center and leave end users with what amounts to a sandbox that they can \n\ncarry around and play games or MP3s on. That should satisfy both constituencies. IT gets more control; users keep their independence.