Board/Minutes/2020-12-04

From OpenStreetMap Foundation

Participants

The meeting was set-up in short time. The Licensing Working Group (LWG) was invited but only Dermot McNally could participate in such short notice.

Board

  • Allan Mustard (chairing)
  • Joost Schouppe (joined 24' after start)
  • Rory McCann
  • Guillaume Rischard
  • Paul Norman (had to leave 56' after start)
  • Tobias Knerr
  • Mikel Maron

Biographies.

Guests

  • Dermot McNally (Licensing Working Group - LWG). The meeting was set-up in short time. The LWG was invited but only Dermot McNally could participate in such short notice.

Minutes and background by Dorothea.

Background

The Licensing Working Group had created a draft attribution guideline , which was subsequently modified by the board. After meetings of the board with Corporate Member representatives on attribution (Jun and Oct 2020) and with LWG members (Oct 2020), the board version was modified. There was subsequently a Board and LWG meeting in October 2020, leading to further modifications to the board's draft.

Attribution related meetings, excluding recurring board and LWG meetings

Attribution related board minutes:

Attribution guidance related LWG minutes:

OSMF members related discussions:

  • Attribution guideline status update: 2019 Aug, Sep, Oct, Nov, Dec
  • Attribution: board letter to Facebook: 2019 Oct, Nov

Minutes

Notes

  • The order of the points is not strictly chronological.

Board members with conflicts of interest

  • Mikel (self-declared)
  • Paul (mentioned by Mikel)
  • Rory (mentioned by Mikel)

Current board conflict of interest policy

Issue

Resolve the inherent conflict between:
a) attribution guidelines that constitute a reiteration of the Open Data Commons Open Database License (ODbL)* in layman's language and
b) an aspirational document of what we (i.e., OSM community) want data users to do, which may or may not be mandated in the ODbL, i.e., may not be enforceable.

* OpenStreetMap data is made available for reuse under the Open Database License (ODbL) share-alike license for data (ODbL page on OSM wiki).

Licensing Working Group (LWG) members want the above to be resolved by the board, before moving on and resolving any other conflict.

On the draft attribution guidelines, in general

Paul

  • Feedback from 2019 State of the Map US conference: Potential users of the data want us to be specific. It was criticized that there was too much language that was vague and that they wanted specifics.

Mikel

  • If we take the practice of previous guidelines, the intention is to provide guidance where there is clarity and avoidance of situations where it's more of a grey area.

On LWG members’ conflict of interest

Guillaume

  • The interpretation of the ODbL by some LWG members is pretty unique and limited to those actors who have a vested interest in minimising attribution.
  • LWG members have a conflict of interest that we are aware of (#Post-meeting note: Michael Cheng is an employee of Facebook, Kathleen Lu of Mapbox and Jim Vidano of Telenav).
  • Complaints on fundamentals by LWG members only seem to be when it doesn't go in the favour in the conflict of interest these people have.
  • We might damage the trust between the community and the Foundation if it looked like we allowed regulatory capture and allowed two persons with known conflict of interests write rules that favour their employers.
  • We should make progress with LWG members without a conflict of interest.

Mikel

  • To reduce positions to lobbying is disrespectful and it's not in the spirit of how we operate in OSMF.

On board-LWG relationship

Mikel

  • Relationship with LWG in bad shape and there has been a lot of miscommunication.
  • We turned the corner in the board-LWG meeting, started to go in a productive direction. Abandoning that posture now is damaging to the foundation of this process.
  • Concern about process and trust with each other and with LWG.

Joost

  • We have a trust issue.
  • Feeling that not all LWG members are doing the best for the Foundation but having a different goal in mind.

On community expectations and time

Mikel

  • Dominant and loud voices in the community about what's required by attribution.

Joost

  • The guidelines have been dragging too long, some people will see it as a strategy. Suggestion to set a deadline.

Tobias

  • The drafting has been going on for 2 years and it’s hard to defend the board's inaction.
  • Problematic as the community watches us deepen cooperation with some companies that provide poor attribution. It undermines the community's trust in us.

On LWG’s initial draft of the attribution guidelines

Guillaume

  • The LWG members are pointing out niches and exceptions in the current draft, but the previous versions had lots of these as well.
  • The original LWG draft would be impossible to pass - it would face overwhelming opposition by the community, as discussion on Migurski's thread shows.
  • It might not be legal for us to grant an exception to attribution that the ODbL doesn't allow.
  • The points of unacceptability and legal impossibility were raised with LWG in March - impasse passed to the board to resolve.
  • LWG refused to discuss possible alternatives (disappearing attribution, which was discussed during the board screen-to-screen meeting).

Mikel

  • Guidelines prior to this were developed with community consultation and board input.
  • Simon Poole was part of the current guidelines since their initiation, who can't be painted as a friend of Silicon Valley companies.

On board’s draft of the attribution guidelines

The board’s draft has the suggestion that attribution can be hidden after 5'. The LWG commented that the draft should not be overly prescriptive and absolutist that we would lose a court case. (comment relayed over email and mentioned during discussion).

Guillaume

  • We could send the text to say "we find 5 is reasonable time on the screen for someone to read the text."

Dermot

  • Didn't find troubling some of the wording that has been challenged as problematic.

On what are considered to be problematic parts of the two versions

Dermot

  • Board draft: Lots of comments are only on a few parts of the document.

Guillaume

  • LWG draft: Redline (provided by board members) was a big amount of work and highlighted the problematic parts.
  • Board draft: Comments on the document (provided by LWG members) show left-over disagreements. Agreed on the November LWG meeting that the new comments would be the review/list of disagreements.

On “loopholes” to make exceptions for attribution

  • Small screens, device-independent pixels.
  • If there are many sources.

On suggestion of attribution via splash screen

Rory

  • Idea seems reasonable.

Dermot

  • For cases that we would like to be credited but people say otherwise, the splash screen approach (or a medium appropriate) allows to do that. Precedents in auto-motive navigation systems and handheld Garmin devices.

On framing the document

Dermot

  • The purpose of this document is to anticipate common use cases and offer a set of formulae/patterns that can be followed, as people would like to know what we consider as adequate attribution.
  • We absolutely have to outline our understanding of ODbL.
  • The document is not binding as ODbL is. Act in accordance with our community.
  • Suggestion to ask for our preferred outcome.
  • Certain cases where we have to accept no attribution whatsoever. If that's in agreement with ODbL (e.g. negligible part), happy with it.
  • Where text became specific, this happened to not be very general.

Guillaume

  • The first duty of OSMF and the board is to the map and the community and the guidelines must reflect that.
  • We're bound by the contributor terms that say we have to licence data under OdbL. There is an obligation to make people aware.
  • The guidelines are one step lower in the legal system than ODbL, and we can't give away something that the ODbL says we can't.
  • The guidelines are our interpretation. If someone comes with a better way of doing attribution, we won't sue them.
  • Guidance is not intended to be minimum but best practice.
  • We would be better off asking for the most favourable interpretation.

On wording

ODbL

"4.3 Notice for using output (Contents). Creating and Using a Produced Work does not require the notice in Section 4.2. However, if you Publicly Use a Produced Work, You must include a notice associated with the Produced Work reasonably calculated to make any Person that uses, views, accesses, interacts with, or is otherwise exposed to the Produced Work aware that Content was obtained from the Database, Derivative Database, or the Database as part of a Collective Database, and that it is available under this License."

Allan

  • Concerned about the use of “should” on the LWG draft, compared to “must”.

Dermot

  • Mindful about the different use of "must" and "should". RFC 2119 defined must and should (incl “may”)

Rory

  • Previous author of attribution guidelines (Richard Fairhurst), said that using that terminology was a mistake.

Joost

  • There's a clause which seems to imply that the document is just community opinion: "based on the wishes of the community"

On legal standing of attribution guidelines

Mikel

  • We're not entering a contract by providing data.
  • Saying we will abide by the terms of ODbL, which does not provide specifics, is a pretty distinct situation from a provider having requirements in a contract. We have to be cautious.
  • If lawyers disagree, that's stronger than amateur lawyers disagreeing.

Allan

  • A licence is an implied contract, that an individual will abide by the requirements of the license and can be enforced in much the same way that a contract can be if we were ever to litigate.
  • Nothing that demonstrates that the understanding of ODbL in the community is false or that one interpretation is more likely than the other.
  • Whatever we come up with, you must provide attribution of some short. Not a huge ask.

Dermot

  • Assumption that if the document isn't licence, the licence still applies.
  • If we under-reach in our recommendations and a local court of law looks at the licence and determines that our recommendation was not enough, that would be awkward.
  • If what we demand is considered reasonable in a court of law we will have a much higher probability of prevailing than if we are absolutist in our approach.

Paul

  • If we produce a document that falls short of ODbL requirements we get into a very bad situation where someone has followed our advice and they've got in legal trouble with it because they didn't follow the ODbL. We aren't the only people who have standing ,we're using third-party data sources through imports. People have licenced data to us under ODbL.
  • What we produce has to be ok - people have licenced data to us under ODbL.

Guillaume

  • If we under-reach, we run the risk the court deciding is not important at all.
  • If we say in the guideline that it's fine to not have attribution in certain cases, then we can't go back on our word and adjust it later, as made clear by LWG members on the New Zealand case.
  • Once the guidelines get published they will set clear expectations and will become common understanding of what is meant by the ODbL.

Suggested next steps and points to consider

Allan

  • Paul, Guillaume and Dermot to sit with the rest of the members of LWG and identify areas in both documents (LWG draft and board draft) that are problematic. Identify and highlight points of contention.

Mikel

  • Do not recommend any decisions made without some legal analysis with direct conversation with us.
  • We have not fully clarified the opinion of the pro-bono lawyer and it is not his area of expertise.
  • Get outside counsel, but not shop around till you get the opinion you want. There were people involved with LWG in the past. None of us are experts.
  • Recommendation to not exclude LWG from that process: it is valid for them to give input and be aware of it.
  • Supporting Dermot's suggestion of brief summary that extracts central points in a separate document, to brief someone else.

Joost

  • Suggestion to ask someone who is neutral and not involved with OSM to have a really good look at the guidelines.

Dermot

  • Create a sparse document with the unresolved issues that can be arbitrated, point-by-point.

Guillaume

  • Suggestion to publish the current draft working document (board draft including comments, making clear what the disagreements are).

Mikel

  • This should not be a rushed job but thoroughly and carefully done.
  • No-one suggested that no attribution is ok – there’s disagreement about what is compromised attribution.

Decisions and action items

Decisions

  • First step: ask the LWG to list their disagreements. Dermot will circle back to LWG.
  • Publish current board draft with comments, after asking the LWG if this is ok.

Action items

  • Dermot to circle back to the rest of the LWG members.
  • The board to ask the LWG if they can proceed and publish the current board draft with comments.
  • Guillaume to present a report at next board meeting.

Suggestions:

  • Set a deadline for the board. Wrap up by March/April.
  • Ask LWG to provide the list of disagreements no later than January.

On finding legal experts
Suggestions:

  • People with FOSDEM talks/presentations/on panels about open source licences. [Rory]
  • Allan to ask Allen Gunn (Aspiration Tech - facilitator of past board face-to-face meetings.

Future step

  • Find all points that we agree with LWG and present them to outside counsel and tell us if it would stand up in court and does not inadvertently undermine the ODbL.

Meeting adjourned after ~ 1hr 20'