[Tinyos-2.0wg] Notes 7-12

[David E. Culler]

Attendees:

Phil Levis
David Gay
David Culler
Jonathan Hui
Gil Tolle
Vlado Handziski
Martin Turon
Jan-Hinrich Hauer

Status

Vlado will commit TTDA5250 Radio with new USART interface on his special 
sub-branch
  - will require changes to other code that uses SPI and USART
  - need check before rolling into main branch

Phil
  - update from Alliance working groups
  - TEP draft -> final approval will involve review process under 
steering committee
   => using WG as surrogate and appoint review coordinator
  - use the Wiki to solicit public review (after Phil is ready)

Phil
  - merged release into devel branch and moved onto the trunk.

David
  - Ready to try to construct a CSMA tep based existing code
  - groups suggest that it should be 105 from scratch
  - DGAY to forward discussion n changes to the control interface
  - Vlado: against using three layer decomposition when speaking about 
protocols
     - don't confuse protocols with software decomposition

TEP 101 ADC
  Vlado raised concerns about naming issues within the TEP.  
  Lots of discussion around desire to have hardware independent sensor 
drivers.
  Discussion about whether TEP prevents the construction of an HIL above 
it. Conclusion NO.
  Should HAL2 actually be a HIL?  
  Code above would need to utilize platform dependent wiring.

  Phil: ReadNOW. "it MUST release access via the Resource interface when 
it is finished"
  needs to be clarified.  Refer to Resource TEP for actual constraints.

  Phil: HAL1 guideline "its design SHOULD follow the guidelines" => should

  Phil: "HAL1 SHOULD instantiate the standard round robin arbiteall ADCs 
should be arbitrated
  in round robin fashion, i.e. the HAL1 SHOULD instantiate the standard 
round robin arbiter.".  
  Ought to indicate the desired semantics, but should not perscribe the 
implementation.  
  What about multiple ADC ports?  
  Jan: Desire to have uniform semantics.
  Actual goal is to avoid starvation.
  Consensus to adjust statement to a more general one.

  Ben: provide detail on read, readnow interfaces in the TEP.  Concern 
over length.

  Add remark about value and shifting.  Shifting should be specified in 
sensor board tep.  

Next week revisit TEP 101 with timeout on HIL question
Focus on 102.