Posted by sfharris
on May 17, 2006 at 5:38 PM PDT
Can you imagine a world in which everyone is wearing medical sensors? Where everyone is downloading 10 megs of medical data per day to some remote database. Where doctors can perform physicals without needing their patients to visit the office and researchers can spot medical trends such as flu outbreaks in real-time? Sound like science-fiction???
Well, of course, it is at the moment, but providing the designs and architectures that might enable such systems in the future is exactly what Jim Waldo and his team at Sun labs are researching.
It's a very interesting problem that requires a distributed solution that is reliable, federated, and long-lived (in other words: it has to stay up as long as you do). It must be robust in the face of changing technologies (sensors, computer languages, computer systems) and network topology changes.
It seems unlikely that we will see a full blown system implemented in our life times but certainly it is not hard to imagine such systems being put in place on a smaller scale in towns and cities around the world in the not too distant future and Jim and his team are paving the way.
If you missed his excellent talk today you can read more about the Neuromancer Project here .
On a less serious note:
Today I opted for some entertainment. I attended the Continuing Adventures of Java Puzzlers: Tiger Traps, hosted by Click and Hack the Type-It Brothers.
I must admit it was a humiliating experience.
The way it works is like this. Josh and Neal show the audience some Java code. They show four possible outcomes of compiling and running the program and ask the audience to vote by show of hands which outcome they believe to be the correct one. I think I guessed correctly about once. Often only about 5% of the audience would get the right answer.
I believe, however, that turn-around is only fair play in this case. Josh and Neal, I think you guys should write a little Web app where your fans can drop off puzzlers for you to solve. Sort of an analog to the Tap-it Brothers "Stump the Chumps".
The rules are simple. People type in some Java code (30 lines or less) and they supply four possible outcomes and indicate which one is correct. You then study the code and select which outcome you believe to be the correct one. The Stump-O-Meter would show exactly how many you got right and how many you got wrong.
Also I was a bit disappointed at the end. I was expecting that Josh would point to Neal look at the audience and say "And remember don't program like my brother". After which Neal would point at Josh and do the same. Well maybe that's carrying the whole Tom and Ray metaphor thing a bit too far.