My colleague, Marc Wagner, and I have written a couple of articles lately on thin clients and we have fairly different perspectives on them (The truth about thin clients, and Why you should actually pay for thin clients). Marc's been in this game a little longer than I have and has seen thin clients come and go in various incarnations as people have repeatedly turned back to full-fledged desktops when they hit the limitations of server-centric computing. However, what this really points to now are the fundamental differences between K-12 and university-level Ed Tech. While we in K-12 are charged with bringing as much technology to the masses as we can on a shoestring (in most cases, at least), higher ed needs to provide serious computing tools to meet dynamic and growing needs of students and faculty in a highly heterogeneous environment.
These differences, in my experience, are two-fold. The first relates to money, while the second relates to student requirements. I spent several years studying and working at Johns Hopkins before heading into private industry. In particular, most of my work focused on IT in support of research endeavors and statistical programming. We never lacked for cash. Every new grant saw newer, faster PCs capable of dealing with our increasing needs for data analysis and manipulation. As Marc has pointed out, at this level, high-end PCs are put to good use and few client-server models (with the exception of mainframes and supercomputers) could satisfy our need for speed. Even at less well-endowed public universities, there is usually enough money moving around from internal as well as external (government- and privately-sponsored) research to keep things moving quite nicely.
Even undergraduate computer labs need to be able to support such a wide variety of computing requirements (whether undergrads are typing a collection of poetry or working on advanced mathematical modeling with a professor), that it pays to ensure that these facilities are as robust as possible.
This isn't just a question of whether and where to use thin clients. Rather, I'm suggesting that we keep in mind the very real differences between the two environments. I wouldn't buy SAS or create a Linux supercomputing cluster for my middle school in the same way that I'm sure my college counterparts would never consider buying refurbs for a student lab. This, as in all things IT, comes back to requirements. The requirements of students and staff in K-12 tend to be pretty mundane. Can you access the Internet? Is malware under control? Can you type a paper or create a presentation? Great, 95% of your computing requirements have been satisfied. In primary education (and even at the middle school level), this is even more true. In K-12, our focus needs to be teaching smart and responsible use of computers and the Internet so that these skills are second nature when students hit higher education or the workforce.
Do thin clients have a place in education? You bet. So do teraflops of processing power for analyzing radio signals from outer space. Just not in the same setting, and probably not administered by the same folks. In many ways, higher ed and K-12 are two different worlds. We can learn a lot from each other, but ultimately need to rely on our users and thoughtful consideration of real requirements to find our way.