[Tinyos-alliance] Notes from IP / Source telecon

[David E. Culler]

Participants:

David Culler
Adam Wolisz
Lama
Phillipe
Matt
Kristin
Ramesh

Background
IP
- IETF style: provide notification, mechanism to contribute, 
organization decides if it wants to steer clear
- WWW / Zigbee: provide IP pool. When you join to share with others 
inside, but not outside.

Licensing models

BSD – do what you want, except remove copyright or sue
Apache – BSD with accreditation
MPL – you must return changes, but you can build around it without 
obligation
LGPL – dynamic linking to GPL code OK
GPL – give back whatever you do
Dual-License – buy out of opensource obligations (typically with GPL)

Discussion
-----------------------------------------------------

Culler - layout basic alternatives above

Ramesh - IETF processes
- WGs tend steer clear of proprietary IP
- not allow people that are based on standards that are based on 
proprietary standards

Matt - clarify IP vs source license

Lama - Intel tends to have trouble with Zigbee
even declaring all the IP that's relevant is problematic

David
Producing a full list of potentially relevant IP is problematic for 
universities too.

Kristin
what is the motivation for including a pool? Besides promotion of the 
value of the organization itself?

David
Companies working together want to ensure that others won't block them. 
All give and all get something back.

Philippe
good to obtain some of the benefits seen in IETF
mistake to not mention it at all

Ramesh
RFC3979 - IETF wg prefer open
discretion to adopt tech with FAND

General consensus on IETF

-----------------------------------------------------
Source License terms

David - outline of relevant models

Rob - Most sources carry something closer to academic license
- does not include advertising clause
- copyright retention in binary distribution

Adam
- what is the problem with Apache license?

Rob
- gives contributor a little more visibility

David
- pro: individual credit
- con: chain can get very large, apache you reference the organization 
not the individual

Adam
- can you reference an old chain

Philippe
- simpler if there was only one type of license
- keep it close to what it is now

Matt
- if all act in good faith, would not be an issue

Rob
- past issues: change in license did not have any consultation
- TOS2 wouldn't want to have just one to reference

Matt
- what if all had same position

Rob
- could rely on nesc doc

Lama
- apache: individual vs group attribution

Ramesh
- third option would be individual working groups

Adam
- individual attribution is a motivation

Philippe
- important for students to get recognition within alliance
- may not matter on the outside

Rob
- disagree, audience to much bigger community
- might motivate people outside the alliance

Matt
- who would actually hold the copyright? The alliance or the contributor

Adam
- two issues can be separate

David
- issue is what is the set that we are willing to accept

Lama
- is documentation in the source code sufficient
- does it have to be in all other documents

Rob
- it would need to be in the manuals

General interest in allowing something like apache
Need to find mechanisms that make it effective.


Lama
- what about checking in object code


Matt
- should we push for a single homgenous license
- eventhough it might be impossible

Philippe
- what about individual working groups
- might find agreement there

Adam
- too much heterogeneity is a problem

Rob
- alliance compliant license
- in favor giving copyright to the alliance with individual attribution

David
- is it naive to expect all these organization to agree to common

General consensus on expanding the options and providing attribution, 
attempting consolidation


next => consoldate and talk voting / committee process