Data Portability/Data Interop – How do we get this done? With Whom? What’s come before? What’s stopping this? Consortium?
Convener: Mary Hodder, Doc Searls
Notes-taker(s): Scott David
Tags for the session - technology discussed/ideas considered:
Discussion notes, key understandings, outstanding questions, observations, and, if
appropriate to this discussion: action items, next steps:
Mary’s outline of discussion:
¥ List existing pieces
◦ terms and policy
◦ data schemas
◦ data dictionary and taxonomy
¥ What can we learn from companies and projects that have failed
◦ list of failed efforts and learnings
¥ List of needed schemas, rules and policies, taxonomy
¥ List of entities that need to be in the room
¥ Find and list gaps in ecosystem
¥ Use cases or verticals:
◦ emergent use cases
¥ Open source effort
¥ UX for taxonomy
Where does VRM go next?
Missing piece of control to distribute may be data flowing in ways that keep user at the center.
Even though it is nuts and bolts, it is a point of power – gatekeeping function.
Working on schemas and standards from organizations have silos with agenda. Interest in keeping
things not portable.
Kantara CISWG as starting point – looking at 2 pieces of the many pieces that need to be established for
a user-driven (not necessarily user “centric”) system.
Clarifying question: Desire for control and moving their data. But which data between which entities?
What is the scope of that.
Back to general issue setting: Lots of information can move easily.
So, session called to talk about the idea of bringing people together to fill the gap – standards making
and schema making body – what in individual at the center. Also a UX element that is important. Make
If putting the user at the center and in control – need it to be accessible to people. UX component.
Who is here. Who need to pull in. What work need to be done. Use this as an organizational session.
New comment: Pieces of a platform may be across a range of things. Distributed infrastructure.
Tracking it across platforms, etc.
information and rights around it. Move the information.
New comment: Don’t look at the movement of the data, look at the movement of the data rights – as
information. That manifests in value o and harm which is the source of the challenges and business
plans. what if it is information not data that moving. Data rights management can yield. Rights different
plumbing than data itself. Secrecy is dead, long live privacy
New comment: Challenge of connecting the shared data to get it into some type of stored thing. Project
to map ID System can provide useful cartography for parties and relationships in the system.
Noted Fing in France – 300 people put data in central store, and made it available to companies. They
have reports from 2013 from the experiment. Testing. Shared data under French data laws. Data store.
Only individual had access. What would an individual do if they got this information available to them.
Question with systems of “if you build it they will come.” does that work?
New discussion of scope: How do we self define scope of the notion. What parts are we seeking to
derive. Might consider narrow scope to help us stay on task on user centric notions and goals. What
changes in older systems when we move to user centricity.
What are focused pieces that are needed? What is boundary set?
Example provided in XML context. First question there was what is needed. Needed reliable data
repository, registry, data dictionary, etc. Had components identified.
It is genus and species question. List species to derive genus.
What can we learn from the companies that have tried and failed?
Consideration of end to end platform. Talk about all in concert, but never enough because not
“hummm” as one machine. Want data to be more standardized in terms of output. Company
willingness to make data public, not wanting to retain it in silos. Looking from systems perspective and
de facto standardization.
Comment: GDPR will make it happen – Article 20 – right to data portability. Will be a driver.
UK MyDATA project – in markets prone to cartels and inertia – drive data back to data subjects. Industry
managed to slow it down with standards. E.g. E-on energy when switch as customer is required to
provide data to the data subject.
Comment: This data is a manifestation of the operation of these kinds of systems.
XDI – semantic elements of link contract designed for data sharing. We are reaching a problem space
for which XDI was designed.
Question of whether all the schemas have been defined. Schemas may already exist. Maybe need to
start with mapping to figure out where holes are and with a larger set of stakeholders.
Question of subverting process. How can the “user at the center” vision happen, when it messes up a
bunch of business models.
Comment: Pushing from the wrong direction. More effective to define the end to end system that
delivers customer value. If can deliver user value, then pulls the other companies into it.
Question of whether this is still centralized.
Answer: Still user permissioned.
Back and forth on terms – They want to define the system with what they can control. Question of
enforcement. Question of whether can make a statement of will and whether it is enforceable.
Market solution and user solution. Is there a negotiation? Is this a notice and consent?
Question of assertion and enforcement of rights.
This becomes easier in system of user expression, need to sign a receipt. Enforceability through
reputation network is what is going to enforce these. Question of whether enforcement is manifested
through reputation systems.
Description of NSTIC/IDESG provided as suggestion of “reputation” type systems to compete on user-
Fact of rating not affect how good going to be with a particular user.
What are the ratings, and do they talk about how made public.
EFF runs a system like that now, like a consumer reports. Done on expert score, not aggregated
Need mapping to help with reputation mapping.
What things could we track and map and decide where there are holders. Where can we combine
Kantara examples – of things proposing to create.
Not need to work within one organization
Electric generation facilities all produce energy into system. Standards organizations.
Syntheses of concepts.
Create artifacts of mutual desire for interaction normalization in the gaps. The ideas of what works in
the gaps will perpetuate themselves through the chosen artifacts, but each will be incomplete. Get “end
to end” harmonization through coordination of the gaps.
Healthcare is a use case. Optimum takes all claims data and harmonizes it.
If going to do that, need formalities of inter-organizational licensing, and liability constraining rules.
Needs to be an open source effort. If protected in some way,
Not every aspect needs to be open. Need some subset of core functionality to be open. Can let folks
sell lemonade at the public park.
Let’s use everything we can that is out there. Not remake things already made.
Build universe of use where user is at the center
Suggest using wikispace to collect the information. What kind of information needed. What levels of
Possibility of this effort collaboration with the creation of a map to leverage existing pieces. Lots of
disparate pieces – how can we bring them together? Not need to build everything.
Expose what is needed to create user-centric market.
Health care may be a challenging area because of high level regulation and interested parties.
Maybe not try to hard. Organic emergence of structures.
Also, what can we learn.