I made a conscious decision some time back to pursue a career in academia and I got back into graduate school this year, working in an 'applied computing' area--educational technology. While I thoroughly enjoy what I am currently working on, in the near future, I intend to conduct research in more mainstream computer science areas. My question is, how long would it theoretically take for one to radically switch research areas? I know there is probably NO fixed time frame for this, but I am especially interested in hearing from individuals that did this or attempted to do this early on in their careers. I would also appreciate additional advice on how I could start working towards this at an early stage. Incidentally, I came across an interesting CACM article  that ascribes a 10-year 'once-in-a-career' shift, mostly as a result of evolving technology.
I've done it a couple of times in my career so far.
I think the 10,000-hour rule isn't far off the mark.
If you have no other commitments, and have the motivation, you can just about do that in three years, if you manage to avoid burning out. Most people will burn out attempting it: it's three years of work-eat-sleep.
If you think you'll only be working 35-hour weeks, and taking holidays, then reckon on six years.
If you're doing it part-time on top of a full-time job, then 12 years or more.
I'm not quite so sure the 10,000 hour rule applies very well here. There are a couple things that need to be considered -- i.e., substantive expertise and methodological expertise. Although I am not in the area of computer science, the transition from educational technology to mainstream computer science doesn't seem terribly far apart. It seems like you have the aptitude for computer science, so you should not have really significant problems acquiring the substantive expertise. I would be concerned if you were saying that the switch was from educational technology to psychiatry or aerospace.
The other issue is the extent to which your methodological expertise generalizes to this new area. Have you acquired research skills that are relevant to your proposed switch? Do you have the core research skills for your new area? How familiar are you with the research in this new area. And what is the litmus test that you have achieved "expertise" in the new area?
I guess the issue really comes down to how much your knowledge and skills in your current area of work generalizes to your new area. So, I'm not going to put a number on this issue -- I just don't think it is possible without more information.
Hmm, it sounds like you are going to a relatively similar field... In any case, you can probably become proficient enough to be a grad student (or maybe post-doc) in your new field relatively quickly.
I know someone who made the switch from theoretical physics to biology mid-grad-school. It added a year or so. I also am in the process of switching from chemistry (my undergrad degree) to biophysics (in which I'll be getting a PhD starting next week). I'm hoping to be caught up by the fall. Loads of people switch fields at the undergrad-grad, grad-post doc transitions. After that I think it can be harder, at least until you get tenure.
Research (at least in the short term) involves being intimately familiar with a very narrow field of information. On the other hand, qualifying exams (in grad school) would require you to have broad based knowledge of the entire field. So it depends exactly what you will need to do and when.