First update 2007..
Tuesday, February 13, 2007 - 12:12
Hmm, I should really write in here more often. The Jennic project is still waiting for me to get documentation for the Jennic integrated radio. I have no idea what the status of that is right now. So I’ve started rewriting the bits and pieces I have for the Jennic chip to be TinyOS 2.x compatible.. Yay 
With regards to my 802.15.4 stack I’m in the process of writing a test harness for the stack to be able to do testing on the non-hardware dependent parts. Hopefully this will be useful for revealing some errors and to do regression testing. I’m also in the process of considering a move of the 802.15.4 to TinyOS 2.x. Unfortunately I found out last week that the timer design in TinyOS 2.x seems to be a bit of a problem. It seems that the guys writing the timer interface found it reasonable to define milliseconds as being 1/1024 second and microseconds as being 1/1048576 second. This means that a symbol period timer ticking every 16 microseconds must either be created by:
- Rounding the 16 microseconds off to be represented in the TinyOS 2.x “binary” time representation
- Writing a separate timer module for 802.15.4.
The first option is not feasable since the maximum time deviation is specified as being +-40 ppm in the 802.15.4 standard. A round-off will yield a deviation that is 3 orders of magnitude higher than the 40 ppm allowed. The second option will require each chip implementation that could possibly make use of an 802.15.4 radio to reserve a timer for this purpose. If this is not done, the TinyOS 2.x could claim all timers for its timer virtualisation leaving none for this special purpose. I’ve commented on this problem on the TEP review wiki found on tinyos.net. I haven’t seen any responses yet.