You are reading content from Scuttlebutt
@mix.exe %EfEwGxRUSw5+DB1E0hcOLt9hlfK/iSwUC3l8K+uFIO4=.sha256

Indigenous Data Sovereignty (Āhau: Māori collective identity through p2p data structures)


Āhau is a FOSS project from Aotearoa New Zealand, built to meet the data needs of Māori. This talk will share insights from thinking in a less Eurocentric way about data, and how the need for data sovereignty has lead us to p2p data models, which fit really well with human-focused systems. Over the last 3 years, we've built a desktop application on secure scuttlebutt that is currently in use by hapū and iwi (tribal groups) to record family histories and create tribal registries.

Full Abstract

Ideas: data sovereignty, self-hosting, katiakitanga (guardianship) of data, history as a path to connection and future growth
Technologies: p2p databases (scuttlebutt, hyper/DAT), CRDTs, subjective histories, encrypted groups

this is what I submitted as a proposed talk for fosdem. I'm posting it here because I forgot what I wrote and I might keep this along with other notes / thoughts as I prepare a talk this coming week. Feel free to weave yourself in by asking questions or lending observations :seedling:

@mix.exe %HKUc8OL12yVKfNo+ftPKy7auC27V/wfvgMcMDeroTkQ=.sha256

our traditional knowledge teaches us that all natural objects have a wairua and a mauri. If our mauri is healthy then our wairua is intact and will make the natural object healthy. Moreover, if we touch something, share our thoughts or are involved with an event, then our mauri is left with that object or event.

The same is also applicable to Māori data. Māori data contains mauri of the individual, whānau, hapū and Iwi. Therefore, any system that has Māori data then has peoples Mauri. If our data and mauri is being used for things that we are not aware of, stored overseas, or in culturally unsafe environments, then this will impact our wairua which will in turn make the individual, whānau, hapū and Iwi unhealthy.


@mix.exe %FnrKq63GB7kRLEFtoj2bNBIlEwOIDwdnx28fXLA2lBQ=.sha256

Thinking in a less euro-centric way about data

NOTE the following aren't absolutes - every whanau is different, there is no one Maori perspective - so these are example of some differences I've seen.

  • where the data lives

    • European: data is just ones and zeros, ... maybe it's private, but if it's encrypted it's basically just ones and zeros that make no sense to outsiders
    • Maori: it's not really right if I'm holding data that's no my families, and the idea of people I don't know holding my data is uncomfortable
  • family trees

    • European: a family tree goes down and down and down
    • Maori: what about whangai - my father was raised by his grandparents, but we need to allow people to decide what the most important parental relationship is when drawing this graph. Oh also, non-blood relationships may or may not be relevant to a graph, it depends on the family
  • free access to info

    • European - information is power and should be free. maybe not free to everyone but if you're in the family you should be able to see everything
    • Maori - family members shouldn't be able to see everything, we have a practice of having kaitiaki who look after some sensitive knowledge (about culture, families), and they decide who learns somethings. (this is because information is power, which needs to be handled by people who have wisdom)
  • hierarchies

    • European - we have super admins who have access to everything
    • Maori - each level (whanau, hapu, papa-hapu, iwi) is self governing, and we coordinate across those levels. The iwi want the hapu to be in charge of defining who it's members are - they know best - and we also don't want the iwi to have access to all the data a hapu does - we the iwi shouldn't know everything.

Ways Ahau has build differently because of Maori perspectives about data

  • Ahau the company holds no whanau records
    • we run data relays/ patakas (storehouses) which are optional and allow no access to content
  • private groups are non-negotiable
  • application processes to join groups are needed, and allowing groups to define the questions they ask is important
  • allowing anyone to publish records, but having an approval / attestation process is needed
  • subgroups are implemented just as another group
    • they're not nested, rather just linked to on another
    • groups can become subgroups after they are created
    • there can be subgroups which not all members of the group know about - e.g. for all the wahine, as some cultural knowledge is specific to them
@mix.exe %GosyUosC/T7g/x+0L+FkYo+8tTiNXIzCHLxNOcNYrTo=.sha256

Why is p2p a good match for data sov?

If you want to be self-sovereign, then you need to run your own infrastructure. We've assessed other current systems (e.g. murkutu), but saw some challenges with these approaches were:

  • requires you to run a server - probably specialised physical hardware
  • you need a tech expert to maintain infrastructure
    • physical hardware
    • software
  • single point of failure
    • if there's a natural disaster/ hardware failure, then you lose all your records
    • some systems run an offsite backup, but that's only a bus-factor of 2, and requires more specialist maintenance

We looked at p2p databases because although more experimental, it hard special affordances.
What we've learnt is that building on p2p is much slower. But it's also enabled different guarantees, and the subjective, communal nature of the data in such a space seems to work really well with the context.

@mix.exe %cUf0fMFkzA8/wPyusKx+l51t+iXQVfZo2WNML4hv7x4=.sha256

Event sourcing means we have a history of edits, and can comment on and attest particular edits/ states.

User has not chosen to be hosted publicly
@Rabble %Rqp+Fnkq17CQROyZquoOdfzZoDz06B78QkBZdHTirgo=.sha256
Voted # Indigenous Data Sovereignty (Āhau: Māori collective identity through p2p
User has not chosen to be hosted publicly
User has not chosen to be hosted publicly
@mix.exe %bJ9Vrg1tKcDukJoRi915yuzN1nqHFfDmBxQOD7ngOtw=.sha256

At one point, a Pakeha, Trader Jack, interrupted. 'Begging your pardon, Sir,' he said to Hobson, 'but it's that Mr Williams. He's not translating a good half of what the Māori say. He's not translating half of what you say either.' Another Pakeha, Johnson, who understood Māori language said that the Māori were saying a lot about the missionaries taking their land and that Williams was not translating it. The rangatira who spoke at the end of the day were in favour of signing the Treaty. They persuaded most of the other rangatira to sign.

User has not chosen to be hosted publicly
@mix.exe %LXaoZGwQwD9sS0A5Gh7s4RALnmhoPP2uIun4pvcp4uQ=.sha256

phew, writing and recording that was stressful.
Have some talk slides #spoilers


@mix.exe %mA+kn7eX+H+sM8cB1hetrnkHJHiom6vcpyXekURztAs=.sha256
Voted I really appreciate reading these ideas :)
@mix.exe %DM6QDF6g1Igd6GdwDgQjipgbpWsyCheXutbexJ52Rrw=.sha256

Talk and Q+A went well!

You can watch the talk here if you're interested :

User has not chosen to be hosted publicly
User has not chosen to be hosted publicly
@Jacob %vlK6yOBxfgxtzasRJA2oUqP3j73Kvm2uyGA7jRKOB00=.sha256

@mix.desktop @Mix Android cool talk and so impressed with how much you've pushed ssb.

what's the ux/ui around merges?

@mix.exe %Sxsa6ery79nHDDPJA8WNiPrElrhRyhDl8x0TvsglpTI=.sha256

@Powersource (pc) alas there's no UX/UI around merges (YET). We have auto-merging and conflict detection built in ssb-crut, but ssb-crut-authors needs a little more work, then we will be wiring it all into ahau. At the moment the current state is "most recently edited branch wins" which is far from perfect, but has been working well enough while frequency of edits is relatively low for particular records (relative to replication).

Got any cool ideas for UX/UI?

User has not chosen to be hosted publicly
Join Scuttlebutt now