Present: Arsalan, Phil, Rodrigo
This was a quick meeting, summarized below...
Phil: In 7 days we'll have a codefreeze for 2.0.1, and it would be good to clean up the CTP code, verify TEP 119's requirements, and verify that the code is compliant with the TEP.
Phil: I will ping Gill about dissemination also. David Gay found some significant issues regarding initialization. Dissemination is initializing the radio and itself automatically. Collection does not do that, it has to be explicitly initialized, and that is the right thing to do.
P: After we do that we should then clean up the implementation section of the TEP to make sure it is coherent with the interfaces in the code.
P: Arsalan, can you talk about the Link Layer Abstraction?
A: This is something that I will do in the near future. Joe has a T2 implementation of SP. One of the driving forces of my implementing it will be DSN (Declarative Sensor Networks, a project in Berkeley). We are going to release code for T1, but will need to have code for T2 soon. This means either a port or a reimplementation/redesign. I prefer the latter.
I sent a draft TEP to Net2 before, didn't get much feedback. What I will do now is to go back and revisit that draft, see what's changed. Comments from the group would be great.
P: what were the important features Joe mentioned in his talk (a talk Joe gave at Berkeley the other day relating his experiences in industry with Moteiv, and contrasting with academia).?
A: feedback mechanisms were not very useful. Message Futures were useful, and the abstraction of different MACs were useful.
A: but there are some things, like rate limiting (think Flush, IFRC), that SP will not support.
P: SP ends up being used mostly on top of TDMA. For TDMA, rate limiting makes no sense. Likewise, Message Futures is not so useful for CSMA. The question is whether we should try to support everything or not.
R: I will take a look at the TEP 119 and the CTP code and come back with comments and changes.