Linux, politics, and other interesting things
Eric Topol gave an interesting TED talk about wireless medical monitoring devices . The potential for helping diabetics and other people who need ongoing treatment is obvious as is the potential for helping pregnant women and other people who might suddenly need medical treatment at short notice.
One significant positive potential for this is integration with multiple services. For example Eric’s talk showed a graph of sleep levels on a mobile phone, the deep sleep (which is apparently the most restorative) was shown in sections that were significantly less than one hour in duration. I often receive SMS messages about network problems during the night, the vast majority of them aren’t that important and can be delayed without any problem. If my phone could determine that I was in deep sleep and delay sending me a NAGIOS report for up to 30 minutes then it would help me sleep while not making any significant impact on my work in repairing server problems – it’s not as if I would be immediately productive if woken from deep sleep anyway.
Eric Horvitz, Carl Kadie, Tim Paek, and David Hovel of Microsoft Research wrote an interesting paper titled “Models of Attention in Computing and Communication: From Principles to Applications” . In that paper they describe various methods for tracking user attention and filtering messages so that the user won’t be needlessly distracted by unimportant messages when they are busy. The next logical step is to integrate that with a smart phone (maybe Android would be good for this) to screen calls, unknown callers could be automatically directed to voice-mail and known good callers could be given an automated prompt as to whether they think that their call is important enough to be worth the distraction.
It seems to me that combining health and status monitoring would be the sensible thing to do. If a bio-sensor array indicates that someone is more stressed than usual then decreasing their availability for phone calls would make sense. It would also be trivial to analyse calls and determine which callers are likely to cause stress and block their calls at inconvenient times.
Of course there are lots of security implications in this. Having multiple networked devices tracking and collating information on health and all activity (including video-monitoring in the Microsoft model) has a lot of potential for malicious use. This is one of many reasons that we need to generally improve computer security.
We also need to have free software implementations of such things. We don’t want Microsoft to get a monopoly on status monitoring. Also it seems that smart Nike shoes that can work with the iPhone  are the leading mass-market implementation of health monitoring, everything related to health-care should be as open as possible. I’m happy for Nike to make their own special products, but we need to have them work with open systems and make the data available as widely as possible. There is the real potential for people to needlessly die if health data is not available in open formats! According to the Wikipedia page Nike has been shipping products using a proprietary 2.4GHz wireless network since 2006 , it would be a really good hardware project to devise a way of making compatible sensors and readers.
We also need some free software for monitoring the user status to avoid interrupting them.
Finally we need some software to integrate all the data. Canonical’s Desktop Experience team are doing some interesting work on managing desktop notifications that will be in Ubuntu Lucid , they appear to have some basic support for masking notifications based on priority. Please keep up the great work Canonical people, and please consider working on advanced status monitoring next!