[Tinyos-alliance] March 14 notes

[David E. Culler]

Phil
Lama
Rob
Jack
Ramesh
Deborah
=====================================

Phil
  IETF open process striving for interoperability and technical excellence
  not against standards that have embedded IP, but want it to be known
  requirements:
    - assurance of RAND
  obligations on individuals (good faith)
    - disclose in timely manner any IPR that you are aware of
       - tell what you know, not that you must do a patent search
    - if cannot disclose, should not participate in the process
   documents
    - if you have IPR that makes its way into RFC, then must grant IETF 
rights to dissemnate RFC
    - grant rights to IP if only way to implement the RFC
    - if there are multiple interoperating implementations without IP 
complaines,
         IESG takes this as indication that there is not an underlying 
IP issue
    building of systems is outside the IETF process
    - bake offs

    by time has been moved forward to a standard, must have 
interoperable implementation
    two interoperable implementations in order for become a proposed 
standard
    
Lama
   -
Jack
   - Want it simple
   - references, vs standards, vs documents
   - reference would be nice to have free of IP
      - but permit proprietary
   - need to get it written down.  Nuances are important.
   - WG's should ...

Rob
   - like IETF model
   - goal is to create a system that would welcome contribution
     - don't want companies with IP portfolio to hesitate contributing
     - want to avoid compulsory licensing of IP

Lama
   - main issue for Intel was not diluting the value of IP
   
David
   - are we worrying about IP entrapment

Rob
   - don't want to put actual contributors at a disadvantage

Phil
   - IETF tries to encourage as much knowledge as possible.


---------------------------------------------------------------------------------------


Reference implementation
   - is the source licence term enough?
   - does there need to be a 'multiple implementations' element?

Deborah
   - multiple implemetations as a matter of custom
   - still much smaller than the IETF situation

Ramesh
   - reason for interoperable because of different vendors
   
David
   - what about NET working group protocols and standard APIs

Phil
   - another example is timer interface

Jack
   - encourage of multiple impls, but doing it through working groups

David
   - presence and history of open implementations takes pressure off 
need for multiple impl
   - encourage existence of multiple interoperable impl, even if internal IP

Rob
   - allow working groups to utilize either open reference or multiple 
implementations
     as a way of establishing "safe"-ness of the proposed standard
   - RAND terms

==> action: draft TEP

---------------------------------------------------------------------------------------