From DVDs to AI: twenty years around Māori GIS, mapping and capability
One of the first things GIS teaches you is confidence. A line appears on the screen cleanly, precisely and apparently without argument. You can zoom into it, measure it, colour it, calculate its area and print it at A0. Once information has travelled through a computer, it acquires a peculiar authority. It looks as though somebody, somewhere, must know exactly where it is and exactly what it means. One of the first things Māori claimant mapping taught me was almost the opposite. Historical land blocks did not always line up neatly with the modern cadastre. Old plans could disagree with one another. Rivers moved. Survey descriptions depended on natural features that had changed or disappeared. Names varied between sources. Areas written on historical records did not always agree with areas calculated by modern GIS. Sometimes a plan was missing altogether, and sometimes two apparently authoritative documents contradicted one another.
In one of the Northland mapping methodologies we had a wonderfully unfashionable description for this: State of Unknown. Rather than forcing the available information to produce an answer simply because the software was capable of drawing one, we recorded that the answer remained unresolved. At the time it was a practical classification for awkward historical geography. Over the next twenty years it became something closer to a professional philosophy. The temptation in GIS is always to resolve, align, categorise and complete. State of Unknown was a reminder that sometimes the more professional thing is to preserve the uncertainty.
What changed more slowly for me was the way I thought about the thing underneath the map. In conventional GIS the land can easily become a surface on which information is located. Parcels sit on it, roads cross it, rivers run through it and points are placed where events occurred. Working alongside Māori claimants, practitioners, researchers, iwi and hapū repeatedly challenged that habit. Whenua was not simply the backdrop to the information. It sat within relationships, whakapapa, occupation, memory, responsibility, history and identity. The map could help represent parts of those relationships, but it could not contain them all.
I am not Māori, and I do not claim to explain te ao Māori or speak for the people whose knowledge shaped the projects I worked on. My part in this history has been much more practical. I have built maps, recovered old plans, configured GIS systems, organised workshops, trained people, supported conferences, found data, developed resources, fixed broken things, produced programmes, dealt with government agencies and occasionally discovered that an ambitious twenty-first-century mapping project was being held together by a cardboard box of DVDs and somebody remembering where the extension lead had gone. What changed my practice was being in rooms where somebody else understood the place, history or meaning far better than I did and learning that the technical role was most useful when it supported that authority rather than competing with it.
The convincing line
By the time I became involved in claimant mapping in the mid-2000s, the work already had a substantial history. Earlier Central North Island work had involved cartography and acetate overlays, with people such as Simon Patience contributing to that earlier phase. The work was increasingly moving into digital GIS, with historical blocks being reconstructed, place names corrected, Crown transactions mapped and claimant research brought together spatially. I did not arrive at the beginning. I arrived at a point when the technology, expectations and volume of material were changing quickly.
My role at the Crown Forestry Rental Trust became that of Mapping Facilitator, which sounds considerably tidier than much of the job actually was. It involved historical research support, finding source maps, scanning and georeferencing old plans, preparing data, working with claimants and researchers, coordinating contractors, producing map books, locating imagery, supporting hearings and gradually helping groups think about what their own mapping capability might look like. A lot of the work sat somewhere between cartography, historical research, project management, technical support and interpretation.
The physical details now make that period feel older than it really is. Historical maps were scanned and moved around until files became too large for ordinary email, at which point a CD was often the sensible answer. Large source plans arrived in pieces. Some projects involved paper maps, scanned TIFFs, spreadsheets, Word documents, JPEGs, shapefiles, handwritten annotations and boxes of research accumulated over years. If a file was too large to send electronically, there was no long discussion about distributed storage architecture. Somebody wrote it onto a disc and sent it.
The history of GIS is often told through software releases, satellites and technological milestones. The practical history is also made up of printer drivers, missing fonts, broken relative paths, unreadable discs, map tubes, badly folded plans and people asking whether the map can be ready before the hearing on Thursday. The software was important, but a surprising amount of the work depended on ordinary logistics and on people being willing to keep sorting through untidy evidence.
What changed most for me, however, was not the technology. It was my understanding of what the map represented. A claimant map might bring together Native Land Court blocks, Crown purchases, Public Works takings, reserves, historical roads, contemporary property boundaries, waterways, place names, sites of significance and information derived from oral or traditional history. GIS was extraordinarily good at putting those things together. Relationships that were difficult to follow across separate reports and historical plans could suddenly become visible. An old land block could be compared with a present-day road, a Public Works taking could be viewed against current property boundaries and a historic place could be related to several different sources at once.
But putting information into the same GIS did not give every source equal authority, and it certainly did not make the person operating the GIS the authority over all of it. If I put a place on a map because somebody with the knowledge and authority to identify it told me where it belonged, I had provided the GIS. I had not provided the knowledge. If a claimant researcher reconstructed the history of a block and I digitised the result, the polygon did not suddenly become my historical interpretation. If somebody corrected an ingoa wāhi and I entered the correct form into a database, the knowledge had not transferred to me merely because I knew how to configure the field.
The computer made all of these contributions look visually similar. The responsibility was to remember that they were not. This is one of the ways a Māori worldview began influencing my technical practice without turning me into a cultural authority. The GIS operator was not at the centre of the knowledge system. The operator was one participant in a relationship involving the people who held knowledge, the whenua to which that knowledge related, the evidence being used, the institutions around the work and those who would inherit the information later.
That problem became particularly obvious when old land blocks were compared with modern cadastral data. A nineteenth-century plan might depict a boundary following a stream while the modern cadastral parcel sat somewhere else. Another historical plan might use a different interpretation. An area recorded in acres, roods and perches might not agree with the area calculated from modern geometry. The natural feature used as the original boundary might have shifted over the previous century.
GIS offers a powerful temptation in these situations because there is almost always something you can do. You can stretch the historical plan, change the control points, georeference it differently, snap a reconstructed line to a modern parcel or adjust the interpretation until the old and new information line up neatly. At some point, however, the ability to make everything fit becomes the problem. Sometimes the discrepancy is the evidence. A historical boundary that refuses to align with the modern cadastre may tell you something about surveying, changing waterways, later subdivision, transcription errors or unresolved historical interpretation. Forcing the sources together can destroy the very thing you were trying to understand.
The correct answer was not always to fix the line. Sometimes it was to preserve the problem. That was the beginning of State of Unknown for me.
Maps with people
A great deal of claimant mapping was not a GIS analyst disappearing into a room and emerging later with a completed map. There were mapping sessions, reviews, meetings and field visits. Drafts went out and comments came back. Someone recognised an error. A researcher found another plan. Somebody looked at the screen and knew immediately that a name or location was wrong. A field visit changed the interpretation of something that had looked straightforward on paper. Sometimes the fastest way to resolve a mapping question was simply to put the map on the table and talk.
That changed my understanding of what technology projects often call requirements. In a conventional project environment, requirements are frequently imagined as something that can be gathered at the beginning, written down and progressively transformed into a final product. Claimant mapping made that model look rather optimistic. Research was still happening. New evidence appeared. People learned things from seeing the maps. A map could provoke another memory, expose a contradiction or make somebody realise that the geometry represented the wrong thing entirely. The requirement changed because understanding changed.
There was also something closer to ako in that process than the usual consultant-client model implied. I might arrive with the GIS skills and a better understanding of how to manipulate the datasets. The person beside me might understand the whenua, whakapapa, research history or local meaning. The useful learning moved in both directions. I learned that a technically correct layer could be conceptually wrong, while others could use the map to see relationships across sources that were difficult to hold together on paper. The GIS was strongest when it became a shared object around which people could question, correct and interpret, rather than a finished technical product presented for approval at the end.
That did not mean the work could be vague. It meant the record of why something changed became more important. If a boundary moved between versions, why had it moved? If a place was added, where had the information come from? If a label changed, who had supplied the correction? If something was approximate, did the map make that obvious or did the cartography accidentally make it look authoritative? One of the most useful sentences in mapping is therefore also one of the simplest: no, that is not right. A large amount of bad GIS survives because nobody with the relevant knowledge has ever been given a comfortable opportunity to say it.
The whakapapa of information
Modern GIS is built around astonishing precision. Coordinates can be recorded to fractions of a metre. Satellite imagery can resolve individual objects. Survey systems can be extremely accurate. LiDAR can describe the shape of the land in extraordinary detail. That precision is genuinely useful, but it can also create the illusion that precision and truth are the same thing. They are not. An uncertain historical boundary does not become more historically certain because its vertices are stored to eight decimal places. A place name can be perfectly positioned and wrongly spelled. A point can be exactly where somebody clicked and still represent information that should never have been placed on a public map. A beautifully designed web map can be technically flawless and culturally inappropriate.
Once I understood that, metadata stopped looking like paperwork GIS people had been forced to complete by somebody in information management. Provenance became part of the map itself. Where had this information come from? Who supplied it? Was it confirmed or indicative? Was there another interpretation? Was it intended for public use? Had it been supplied for one project only? Could it be reproduced elsewhere?
Over time I also found the idea of the whakapapa of information useful, provided it is handled carefully. I do not mean that whakapapa can be reduced to metadata fields or that a technical lineage is equivalent to whakapapa in its fuller Māori meaning. The useful influence is the insistence that information has origins and relationships. A layer came from somewhere. Someone created it for a purpose. It may have passed through several people or institutions. It may carry obligations, permissions or limitations. Detaching the geometry from that lineage can make the data easier to reuse technically while making it less trustworthy culturally and historically.
Government information usually arrives looking like data. It has fields, schemas, dates, coordinates and perhaps even metadata. Historical and cultural information is not always so cooperative. It may be embedded in written evidence, marked on a photocopy, described in an interview, remembered in relation to another place, recorded under an older spelling or represented differently on several generations of maps. Turning that into GIS is not merely data entry. It is translation, and translation can remove meaning if you are careless.
A database field called NAME does not tell you whether the name came from a nineteenth-century survey plan, an official government dataset, a Gazetteer, a claimant researcher, a local correction or somebody trying to decipher a barely legible label on a faded map. A field called TYPE cannot adequately explain the difference between a remembered place, an urupā, a former settlement, a historical event location, a site of significance and a point placed only to identify a general area during a field visit. GIS likes categories. History and relationships are usually less cooperative.
The solution was not to abandon structure. It was to structure the information without pretending that the structure was the knowledge itself. The better database preserved enough provenance for somebody else to understand the origin and limits of the feature. The better map made uncertainty visible. The better project kept the relationship between people and information intact rather than assuming that once something had become digital it belonged to the database.
Place names are not labels
Place names made all of this particularly obvious. In GIS, a place name looks like a simple object. A geographic feature has a name. The name is stored in an attribute. The GIS displays the label. Problem solved. Except ingoa wāhi can carry history, language, identity, memory, whakapapa and authority. They can also carry layers of colonial transcription, simplification, spelling error and administrative replacement.
By the mid-2000s claimant mapping was already producing substantial numbers of corrections and additions to place names. Later, names became a recurring strand through Māori GIS conferences, Treaty settlement work, the New Zealand Geographic Board, the Gazetteer, Kā Huru Manu and many iwi and hapū projects. People such as Huia Pacey, Maxine Bluett, Wendy Shaw, Takerei Norton and many local knowledge holders contributed expertise in different parts of that work.
My own role was much more limited and technical. I could help make a name searchable, configure the database properly, make sure a macron survived export, connect a name to historical mapping, coordinates or a Gazetteer and help build systems through which people could find and compare information. None of those things gave me authority to decide what a place should be called.
That distinction became more important as mapping moved online. In the desktop era a spelling mistake might remain inside one GIS project or appear on a printed map. In the web era a bad name could be propagated immediately. One incorrect form could be copied into services, applications and derivative datasets until repetition itself began to look like evidence. Digital systems accelerate correction, but they accelerate error as well.
There is a wider Māori worldview influence here that changed the way I think about naming. A label in a GIS can look like a descriptive tag attached to a location. An ingoa wāhi may instead be part of a relationship with that place, carrying memory or history that the field itself cannot contain. The technical system should therefore be humble about what it holds. It can preserve the approved name, source, language and location. It cannot assume that storing the text means it contains everything the name means.
Put the GIS in the marae office
By 2006 and 2007 another question was becoming difficult to ignore. Producing good maps for people was useful, but what happened when they wanted to change something themselves? Professional GIS was expensive. Software licences mattered. Hardware mattered. Data formats were obscure. Training took time. Even when a community received an excellent map book or sophisticated GIS resource, that did not necessarily leave them with the capability to continue the work.
The Ngāti Manawa GIS pilot was one of the projects that pushed directly at this problem. The ambition was not simply to deliver finished outputs. It considered how GIS capability might remain with the group, including equipment, training, data, metadata, provenance and the authority attached to the information. Michael Gilmore was among the people involved. The basic idea was simple: put the GIS where it would actually be used.
That required a different way of thinking. Instead of asking only what maps the project should produce, we began asking what had to remain afterwards. Who would use the computer? What software would they have? What happened when the licence expired? Could the data be updated? Would somebody know how the folders were organised? Could the next person make sense of the system?
I would now describe part of that shift as a question of rangatiratanga and mana motuhake, although I would be cautious about pretending we were using that exact language for every technical decision at the time. The practical lesson was clear enough: a group had more control when it could maintain, interpret and move its own information rather than depending permanently on an outside specialist. Capability was not just access to the final map. It was the ability to make informed choices about the system itself.
That also brought an intergenerational dimension into the technical work. A system should not make sense only to the person sitting at the keyboard today. The next person should be able to understand where the information came from, what can be changed, what should remain controlled and how the technology works. Succession is not merely a project-management task. It is part of ensuring that the mahi and the information are not stranded when one person leaves.
When Māori GIS came on ten DVDs
Around the same period we developed what became known as ArcFlax and other claimant resource packages. They now look like artefacts from another technological age. ArcReader allowed people without a full desktop GIS licence to open prepared mapping environments. Packages could contain map layers, imagery, PDFs, two-dimensional and three-dimensional views and supporting material. At one point the complete resource system occupied ten DVDs.
There was something gloriously contradictory about trying to democratise advanced digital mapping while the entire experience depended on feeding a numbered sequence of plastic discs into a computer. At the 2007 Esri conference the idea was described as “Mapping for the Masses”. It was a slightly grand title for a system whose success depended partly on the duplication company getting several thousand discs copied and packaged correctly, but the underlying problem was genuine. How could sophisticated GIS material be placed into people's hands without requiring everyone to become a professional GIS specialist or buy expensive software?
The important part was not the plastic. It was the attempt to move capability and information closer to the people who needed them. The DVDs were one technological answer available at the time. Today the same intention might use QGIS, web services, offline packages, local servers or a cloud environment. The underlying question remains whether the architecture increases practical control or simply shifts dependence from one provider to another.
This was a period when the exciting digital future still had to be physically duplicated, labelled, boxed and transported. It is easy to laugh at that technology now, and I do, but modern systems have merely developed more sophisticated ways to fail. Then the map did not work because Disc 7 was missing. Now it does not work because a cloud authentication token has expired. The dependency has become less visible, not necessarily smaller.
What the ArcFlax period taught me was that accessibility and sustainability are not the same thing. We could provide the viewer, the data, the map and the instructions. We could install software and leave a computer behind. Yet every system remained dependent on something: a version of Windows, a file path, a software component, a licence, a hard drive or a person who knew how it worked.
This led to a question that has followed me through almost every later GIS project: what happens after the mapper leaves? If a system can only be maintained while one contractor, staff member or volunteer remains involved, then some of the capability belongs to that person rather than the organisation. Technology projects have a long tradition of building something important, documenting approximately sixty per cent of it, putting a shortcut on the desktop and assuming that whoever built it will be around forever. Eventually that person leaves. That is usually when everyone discovers what capability actually means.
Maps in hearings
While we were trying to make GIS more accessible, the maps were also becoming more consequential. During 2007 and 2008 claimant mapping was increasingly being used in Waitangi Tribunal hearings, judicial conferences and formal evidence processes. Large-format maps were produced, map books were assembled and digital material had to function in hearing environments rather than only in a GIS office. Judges and Tribunal staff sometimes requested additional maps because the geography was necessary to understand complicated evidence.
That period reinforced something GIS people occasionally forget: cartography is not decoration. A map can materially affect whether an argument can be followed. If a report refers repeatedly to historical blocks, waterways, Crown purchases, roads, settlements and old place names, the reader may spend a great deal of time trying to reconstruct the geography mentally. Put the same information into a well-designed map and relationships can become immediately visible. The opposite is equally true. A poor map can make clear evidence confusing. Scale matters. Labelling matters. Colour matters. Generalisation matters. What is included matters, and what is omitted matters.
This is also where I became more conscious of maps as instruments of authority. The map shown in a hearing can shape what everybody in the room sees first. A clean boundary can appear settled. A place omitted from the map can disappear from the visual argument. An official cadastral background can make the modern property system look like the natural frame within which the historical story occurred, even when the historical relationships preceded it.
For Māori claimant work, that meant the cartographic choices carried responsibilities beyond visual quality. The map needed to help communicate evidence without silently converting uncertainty into certainty or allowing state geometry to dominate every other way of understanding the whenua. The map could support claimant evidence, but it should not pretend that the GIS itself was the evidence's source of authority.
The Central Claims Charitable Trust mapping work was a good example of how dynamic this could become. Significant places were corrected and added as the work progressed. Claimant feedback altered later versions. Field visits changed interpretations. Mapping did not simply illustrate research after the research was completed. It became part of the research process itself.
That created another problem: the authoritative-looking map was not one static truth. A version printed in February might not be the same as one produced in April. A name might have changed. A location might have been corrected. A boundary might have been refined. A site might have been removed because it was not appropriate to reproduce. The map was therefore part of an ongoing evidential process, not the final word on it.
Possession is not permission
There were also situations in which information supplied within one claimant mapping process could not simply be reproduced in another. The details of individual cases are less important than the principle. GIS makes copying extremely easy, and that technical convenience can encourage the assumption that if a person possesses a digital layer they also possess the right to use it. They do not.
The practical questions were simple but important. Who supplied the information? Why was it supplied? Was it intended for this project only? Who could see it? Could it be reproduced elsewhere? If somebody lent us a historical map, did digitising it give us any wider rights over the information? If culturally sensitive locations had been mapped for a particular purpose, did that mean they belonged in the next map book?
The easiest part was often obtaining the file. The difficult part was understanding the relationships around it. This changed the way I thought about datasets. A shapefile is not simply a shapefile. It has a history. Somebody collected it, somebody paid for it, somebody may have shared it under conditions and somebody may be represented in it without having agreed to wider publication.
This is an area where later Māori data-sovereignty thinking gave stronger language to something we had encountered practically for years. Rangatiratanga over data is not just a question of who has a copy. It is about authority over collection, access, use, interpretation, sharing and future reuse. Kaitiakitanga is not a convenient translation for data stewardship. It carries a wider sense of responsibility that makes the question “Can we publish this?” less important than “What responsibility comes with holding this, and who should make the decision?”
The existence of a technical capability to publish something does not create an obligation to publish it. Sometimes good GIS practice means leaving the layer off the map.
A community begins to organise
By 2008 the work had expanded far beyond individual claimant projects. There was already a much wider Māori GIS practitioner community. People were working in Treaty settlements, iwi organisations, government agencies, local authorities, environmental projects, universities, consultancies, cultural mapping initiatives and Māori land contexts. What started to emerge more deliberately was infrastructure through which those people could meet one another and maintain relationships between events.
The 2009 Māori GIS Conference in Christchurch was an important point in that development. Huia Pacey was a central organiser and proponent, with support from Eagle Technology and a wider group of people who saw the value in creating a national forum. More than a hundred people attended, bringing together Māori practitioners, iwi representatives, researchers, government agencies, consultants and others working around mapping, place and data. I helped promote the conference nationally and was involved in the wider network around it, but the conference itself was collective work, with Huia playing a major organising role.
The more interesting story is how the community developed. The conference created momentum for people to remain connected afterwards. A Māori GIS Google Group, established by Moka Apiti, provided one simple way for questions, opportunities and useful information to keep moving after everybody went home. A second Māori GIS conference followed in 2010 at GNS Avalon in Lower Hutt, with people including Huia Pacey, Roland Pomana, Gina Mohi, Tony Elson and others contributing across the broader organising effort. I helped with programme production, presentation coordination, practical delivery and a session on managing Māori GIS projects.
Looking back, whanaungatanga is a better way to understand what made these networks useful than the usual language of professional networking. People were not simply collecting contacts. Relationships were forming across iwi organisations, agencies, universities and the private sector. Someone could ask a question because they knew another practitioner would answer it. Knowledge could move when people changed jobs. A government specialist could help somebody in an iwi project understand a dataset, while the iwi practitioner could change how that specialist understood the real-world implications of the data.
That movement of knowledge was also a form of ako. The community was strongest when it did not assume expertise flowed in one direction from a technical presenter to Māori participants. Different people brought different forms of knowledge. A software specialist might know the tool. Someone else knew the whenua. Another person knew the policy environment, the tikanga, the research history, the te reo or the community relationships. The useful thing was the combination.
The work around the conferences also fed into Te Kāhui Manu Hōkai, the Māori GIS association and practitioner network. Like most volunteer organisations, its development was more gradual than a neat founding story suggests. Meetings happened, a name developed, governance evolved, websites and systems appeared, and different people contributed leadership, practical work, networks, language, sponsorship, cultural knowledge and time. My role changed through those years. At first I was often the technical helper, building things and solving practical problems. Later I became more directly involved in komiti work.
The purpose was not to create another institution for the pleasure of having one. It was to give the practitioner community some continuity.
What happens on Monday?
Conferences have a structural weakness: they end. A good event can make GIS seem completely achievable for two or three days. The demonstrations work, people are enthusiastic and everyone leaves with ideas. Then Monday morning arrives and a participant sits at their own computer and discovers that they do not have the same software, the same data, the same licence, the same menu or anybody sitting beside them who remembers what the trainer did.
That gap between event enthusiasm and practical capability became increasingly important to me. In 2011 Ocean Mercier invited me to contribute GIS and mapping practice to MAOR210 Cultural Mapping at Victoria University. In 2012 and 2013, after I moved into DOC as Geospatial Information National Lead, practical Māori GIS and geospatial outreach work increasingly focused on what happened after the initial exposure to the technology.
Managing Māori GIS Projects workshops developed during this period and were delivered in places including Rotorua and Whangārei. We talked about data, software and mapping, but also about the questions technology projects often postpone until after something has already been built. Who owns the project? Who maintains the data? Who pays for the software? Where are the backups? What happens if the person doing the work leaves? Which information can be shared and which should remain controlled? What will still be understandable in ten years?
These were not glamorous questions, but they were the difference between a demonstration and capability. They also brought kaupapa and succession closer to the centre of the technical discussion. A project did not succeed simply because a map had been delivered. If the people who needed the GIS remained unable to update it, interpret it or control the information, then the project had delivered a product without necessarily delivering much rangatiratanga.
A geospatial outreach workshop in Wellington in 2013 attracted around forty registrations and a waiting list. The level of interest reinforced what had already become obvious: people did not simply want to hear what GIS could do. They wanted to do it, and they wanted somewhere to return when the first real problem appeared.
My own measure of success changed accordingly. Earlier in my career, success often meant producing the finished thing. The map was complete, the data had been delivered and the project was finished. Increasingly, success meant somebody else knowing how to continue without me. That is a much less impressive result in a project closure report, but it is also probably the more important one.
Indigenous Mapping Wānanga
The next phase developed through the meeting of local Māori GIS experience and international Indigenous mapping networks. There was no single originator. The Firelight Group in Canada had developed Indigenous Mapping Workshop approaches, with Steven DeRoy among the important people in that lineage. Google Earth Outreach had international trainers and tools. In Aotearoa, Māori organisations, practitioners, Digital Navigators, government agencies and others helped adapt and host those approaches locally.
In November 2014 an Indigenous Mapping Workshop was held at Te Whare Wānanga o Awanuiārangi in Whakatāne and was oversubscribed. That mattered because it changed the question. Earlier discussions about GIS access sometimes sounded as though the challenge was persuading communities that mapping technology might be useful. By 2014 the problem was more immediate: more people wanted to attend than there were places available.
The format mattered. Participants could bring actual projects. Mapping tools were taught in relation to practical needs rather than as isolated software functions. International Indigenous mapping experience, local Māori organisations, trainers and technical people all contributed. The workshop was not significant because one particular software company was involved. It was significant because Indigenous mapping capability itself was being treated as something worth investing in.
The wānanga model also changed the atmosphere of learning. A conventional training course tends to place the trainer at the front as the expert and the participant as the recipient. Wānanga created more room for exchange. The technical trainer could demonstrate a method, but the usefulness of that method was tested against the participants' own kaupapa. People brought questions from their whenua, taiao work, heritage projects and organisations. The learning could move between participants as well as from the front of the room.
That reciprocal quality matters to me. I have facilitated many sessions where the most useful contribution was not my presentation but a participant explaining how they had solved a related problem somewhere else. Ako is a better description of that than delivery. The facilitator helps create the conditions for learning; they do not own all of it.
Cloud mapping without the cloud
By 2017 the Indigenous Mapping Wānanga at Claudelands in Hamilton had become a large collaborative event involving Digital Navigators, Waikato-Tainui, Google Earth Outreach, Te Kereru Associates, LINZ, Firelight Group and a broader group of organisers and trainers. Through LINZ, I was heavily involved in technical leadership, programme delivery and troubleshooting.
The event also delivered one of the better reminders that technology is only as good as the infrastructure underneath it. Much of the training relied on online and cloud-based tools, but the venue internet did not share our confidence. A large room full of people opening web mapping applications at once is an extremely efficient way of testing whether a venue's Wi-Fi is genuinely fit for purpose. The answer in this case was not especially encouraging.
Connections struggled, trainers improvised and participants worked around failures. The problem became significant enough that the venue hire was eventually refunded. It was frustrating at the time, but it became one of my favourite illustrations of how technology projects actually work. We had an Indigenous Mapping Wānanga partly built around the exciting possibilities of cloud computing and, for periods of time, insufficient access to the cloud.
The lesson was not really about Wi-Fi. It was that access is infrastructural. A global mapping platform can exist and still be inaccessible to the person sitting in front of you. Free software is not genuinely accessible if the computer cannot run it. Public data is not useful if nobody can find or interpret it. A cloud platform does not create capability simply because the account can be created without charge.
There is also a manaakitanga dimension to training that technology planning can overlook. If people have travelled to a wānanga, given their time and arrived ready to learn, the environment should support them. Good facilitation is not only having the correct slides. It includes making sure people can connect, participate, ask questions, work at different speeds and leave with something useful. The technical infrastructure is part of how people are looked after in the learning environment.
The 2018 wānanga expanded the practical toolset further. Google products sat alongside QGIS, Esri Survey123, Mapbox and other approaches. Another Rangatahi Day was included. I was heavily involved as a trainer, and LINZ remained a major supporter. The programme was increasingly software-neutral because the real question was not which company had the best platform. It was what people needed to get useful work done.
The support after the event was often where that capability became real. Participants asked about government datasets, controlled information, data access, field tools and how to move from workshop exercises into their own organisations. That was Monday morning again, and Monday morning continued to be the part that mattered.
From Facebook to regular online wānanga
By the late 2010s there were several overlapping communities around Māori GIS. Te Kāhui Manu Hōkai had its association and conference history. Indigenous Mapping Wānanga had a related but distinct lineage and stronger international connections. Government agencies, universities, consultants, iwi organisations and practitioners moved between initiatives. At the same time another community developed in a much less formal environment through Facebook.
A Māori GIS Facebook Group became an active practitioner space in which people shared questions, events, datasets, maps and technical help. Facebook worked largely because people were already there. It required almost no behavioural change. Someone could ask about a dataset in the morning and receive useful suggestions from practitioners elsewhere in the country without waiting for the next formal event.
Facebook Live added another dimension. By 2018 I was running live Māori GIS sessions online around three-dimensional mapping, storytelling, virtual reality and related tools. The informality was an advantage. There was no need for a dedicated learning platform or a national communications campaign. A session could simply be announced, people could join, questions could be answered and the recording could remain available afterwards.
The technology mattered, but the relationship mattered more. The online group worked because people were prepared to share knowledge. That is another expression of whanaungatanga in the practitioner community. The platform was temporary. The habit of helping one another was more durable.
Around 2020 I deliberately reduced my involvement in the Facebook group, partly because a community that depends permanently on one person is not much of a community. My attention shifted toward a more structured model of practical online learning. The Facebook group later disappeared after conflict within the group. By then I was no longer involved in its administration. The more important story for me was that the practical online learning model had already moved somewhere else.
Ngā Poutama Matawhenua
Ngā Poutama Matawhenua was operating as a practical online Māori GIS mapping wānanga by October 2020. It grew from the same needs that had driven much of the previous decade: practical teaching, accessible tools, ongoing contact and support that continued after one event finished. Roland Pomana and others were part of the early connections, with Te Kāhui Manu Hōkai and LINZ contributing to the wider collaborative environment.
What mattered was what the programme became. Ngā Poutama was intentionally practical. Someone could attend because they wanted to understand StoryMaps. Another person might be trying to work out how to access public data. Somebody else might be building a hapū GIS. Another participant might be opening QGIS for the first time. The programme did not assume that everyone needed to become a professional GIS specialist. It assumed they needed enough capability to do something useful.
The wānanga format also gave the idea of ako more room to operate than a conventional software course. People could return with questions. Guest presenters brought specific expertise. Participants brought their own experience. The facilitator did not need to pretend to be the cultural, environmental, technical and historical authority in every session. Sometimes my role was to explain the GIS. Sometimes it was to find the right person to explain the subject and then get out of the way.
That is a significant change from the traditional expert model. The useful facilitator is not the person who keeps proving how much they know. It is the person who helps connect the knowledge already present, brings in what is missing and leaves people more capable afterwards.
One principle became particularly important: helping people did not require taking control of their information. The phrase eventually used was simple: we will never ask you for your data.
That sentence connected the modern online programme directly to lessons from claimant mapping years earlier. A trainer does not need somebody's culturally sensitive information in order to explain how to organise a database. A facilitator does not need a copy of a wāhi tapu layer to demonstrate symbology. A commercial cloud platform does not need to become the permanent custodian of local information simply because it is convenient for a training exercise.
Methods can be taught with sample data while the real information remains under the control of the people to whom it belongs. That is a practical expression of rangatiratanga. It changes the relationship between trainer and participant. The participant does not have to surrender information in order to receive technical help.
As Ngā Poutama developed, the topics expanded across StoryMaps, public data, QGIS, Survey123, LiDAR, three-dimensional mapping, Māori land information, project planning, data security, environmental monitoring and many other areas. The real measure of success was not whether a person remembered every button demonstrated in a webinar. It was whether they came back later asking better questions.
A beginner might ask where the tool is. Someone developing capability starts asking whether the data can be connected as a service instead of downloaded, then who controls it, where it should be stored, whether it can be shared, what happens if the person who built the system leaves and whether the map should be public at all. Those later questions are much more interesting because they show that the person is no longer simply following instructions. They are making decisions about the mahi.
Local control
The contrast with the mid-2000s is extraordinary. A person starting a GIS project then might have required specialist desktop software, an expensive licence, powerful hardware, imagery obtained through purchase or personal contacts and physical media large enough to carry the data. Today someone can download QGIS at no cost, connect directly to national web services, view modern basemaps and historical imagery, use Census information, create mobile field forms, work with LiDAR, generate three-dimensional terrain and publish interactive map stories.
The technical barrier has fallen dramatically, but that has not made GIS simple. It has moved the difficulty. There is now so much data available that choosing and interpreting the correct dataset can be harder than obtaining one. A search may produce several plausible layers. The problem is understanding which one actually represents the thing you need. The same is true of publishing. A person can create a technically impressive web map in very little time. The harder question is whether it should exist.
The idea of locally controlled GIS has become much more achievable than it was when we were trying to place dedicated mapping computers into community offices. A hapū GIS no longer has to mean an expensive enterprise system. It might combine QGIS, local files, public web services, controlled cloud components and simple publishing tools.
The harder part is governance and succession. What information should the system hold? Who maintains it? Who can edit it? What is public and what remains private? What happens when committee members change? Does the organisation know where the passwords are? Are there backups? Can the data be exported if the platform changes? Does an external provider retain any rights over what has been uploaded?
These questions are not exciting, which is exactly why they matter. Technology projects rarely fail because nobody can draw the map. They fail because nobody decided who would still be responsible for it three years later.
A Māori worldview adds another dimension to that continuity. The people maintaining the GIS today are not necessarily the final users of the information. Some projects are concerned with histories and places that extend back generations, while the information being created now may need to remain understandable to people decades from now. Thinking intergenerationally changes mundane technical choices. Open formats matter. Provenance matters. Succession matters. Password ownership matters. The explanation of why information was withheld may matter as much as the explanation of why another layer was published.
A good system should not merely work today. It should remain understandable tomorrow.
Maps as stories
Another significant change has been the movement of GIS from specialist analysis into storytelling. StoryMaps and similar tools allow maps to be combined with photographs, text, historical imagery, video and other media. That is especially powerful for histories that are already spatial.
A written description can say that a landscape changed. Historical aerial photography can show it. A map can place the change geographically. Modern imagery can show what is there now. A narrative can connect the pieces. Three-dimensional mapping adds another dimension where the shape of the land matters. A ridgeline, catchment, pā site or route may be difficult to understand from a flat map if the reader does not already know the place.
Used well, these tools can help a reader understand relationships between places rather than merely the locations themselves. An awa can provide the organising structure of a story. A sequence can follow movement through the landscape. Historical imagery can sit beside present-day restoration mahi. A map can help a reader see why two places belong in the same narrative even when they are far apart on a conventional map.
The caution is that digital storytelling can flatten relationships just as easily as databases can. A StoryMap template likes a sequence of stops. A cultural narrative may not be a sequence of independent locations. Whakapapa, relationships and meaning cannot necessarily be represented by moving the camera smoothly from point A to point B. The story should determine the technology, not the template.
The same applies to publication. The public StoryMap can be a carefully approved view rather than the system of record. The detailed source information can remain local and controlled. Rangatiratanga includes authority not only over what is shown, but what is withheld and how the story is framed.
AI arrives
Artificial intelligence is the latest addition to this toolkit, and the change is unusually large. For much of my career, technical knowledge created substantial barriers. Writing code, transforming data, configuring systems and troubleshooting GIS required accumulated experience. Now an AI system can explain a script, generate a workflow, produce a draft web application, translate technical documentation or help somebody work through a software problem in seconds.
That could be enormously useful for community GIS capability. It can help someone understand code rather than depend on a developer, turn a technical error into plain language, draft metadata or explore a public dataset. Used well, it can reduce barriers and support mana motuhake by helping people do more of their own work.
It also creates a new category of risk. GIS has always been capable of being wrong. AI is capable of being wrong fluently. A GIS error often looks wrong. The layer is in the wrong place, the projection is clearly incorrect, a field is missing or something fails to load. An AI error may arrive as an elegant paragraph explaining something that never happened.
For Māori GIS the implications are obvious. An AI system can invent a plausible explanation of an ingoa wāhi, confuse iwi, generate an unreliable boundary, treat repeated internet material as authoritative or combine fragments from several contexts into a story that sounds coherent but is historically false. It can also encourage people to upload information to external systems without thinking carefully about where that information goes.
AI therefore brings us back to the same old questions with greater urgency. Where did this information come from? Can it be verified? Who has authority to answer? Is the system describing public information, or is it pretending to possess knowledge it does not have? Should this information have been put into the tool at all?
A Māori worldview makes the limit especially clear. AI can assist with mapping; it does not acquire cultural authority because it can speak confidently. It can help organise public information; it does not become the holder of whakapapa. It can suggest a technical workflow; it does not decide whether culturally sensitive information should enter that workflow. Those decisions remain with people.
The technology is new. The discipline is not.
From mapper to facilitator
When I look across the last twenty years, the biggest change in my own role has not been from one software product to another. It has been from producing maps toward enabling capability.
In the early period I was much closer to the actual production work: find the plan, scan it, georeference it, build the layer, fix the geometry, create the map, print it and get it where it needed to go. Later the work moved outward. Find the people who need the capability. Find the person who has the knowledge. Find the trainer. Find the data. Find the funding. Find the venue. Work out why nobody can connect to the internet. Find out who owns the account. Help the participants continue after everyone goes home.
The practical GIS skills remained useful, but the role became increasingly connective. That is not always how technology careers are described. Professional expertise is often presented as accumulating greater control. You become more senior, learn more and therefore become the person who makes more of the decisions.
My experience around Māori GIS pushed me in the opposite direction. The more useful I became, the more conscious I became of the decisions that were not mine to make. A technical person can explain storage options, but the community decides what should be stored. A GIS practitioner can explain the implications of putting a layer online, but the people with authority over the information decide whether it should be published. A facilitator can make a wānanga happen, but the participants decide whether it was useful.
The work also changed my understanding of expertise. Expertise does not have to mean standing in the centre. Sometimes it means knowing enough to support the person who should be in the centre. Sometimes it means recognising that the most responsible technical recommendation is to keep the information local. Sometimes it means saying that the map cannot answer the question being asked. Sometimes it means being corrected and being pleased that somebody else now understands the system well enough to correct you.
Being corrected is an underrated measure of capability. It means knowledge is no longer travelling in only one direction.
Maps and authority
Maps have authority because people believe maps. Put a statement into a report and someone may question it. Put the same statement on a map and it suddenly looks official. Digital basemaps amplify the effect. Roads, coastlines and imagery may be extremely accurate, so whatever boundary is drawn over the top inherits some of that visual authority.
Except it may not deserve it. A reconstructed historical boundary can be approximate. An iwi boundary shown for one administrative purpose may not represent another kind of interest. A place-name point may represent a broad geographic feature rather than one coordinate. A culturally sensitive site may be deliberately generalised. A historical area may have several competing interpretations.
Good GIS therefore has to do more than display information clearly. It has to display uncertainty honestly and authority proportionately. That is difficult because cartography prefers clean answers. Users like clean answers. Senior managers particularly enjoy clean answers. The world is under no obligation to supply them.
This is why State of Unknown has stayed with me. At first it meant that a historical boundary could not be resolved. Later it could mean that an ingoa wāhi required confirmation, a permission was unclear, an event was proposed but could not be confirmed as delivered, or I remembered being involved in something without being certain enough to claim a particular role. Now it can also mean that an AI system has produced a convincing answer without enough evidence to support it.
There is constant pressure in professional work to remove uncertainty. Sometimes the more professional action is to preserve it.
What Māori GIS became
Over those two decades, Māori GIS became much easier to access and much harder to define narrowly. It includes Treaty research and claimant mapping, but it is not only Treaty mapping. It includes cultural mapping, but it is not one method. It includes environmental monitoring, Māori land, planning, place names, heritage, infrastructure, story mapping, field-data collection, imagery, climate information, cadastral data, three-dimensional visualisation and community-controlled information systems.
Some work uses Esri. Some uses QGIS. Some uses Google tools. Some uses custom systems. Some should probably remain in a spreadsheet. Some should not be digitised at all. That variety is healthy. A mature capability is not defined by one platform. It is defined by people being able to choose tools appropriate to their kaupapa.
The software industry naturally encourages the opposite idea. Every vendor would prefer its platform to become the answer to everything. Increasingly, platforms want to become the place where data is stored, analysed, published and governed. Sometimes that is useful. Sometimes the correct architecture is much simpler: a local QGIS project, a controlled folder, a backup, a public service for information that is genuinely public and no culturally sensitive information anywhere near an external platform that does not need to hold it.
Architecture should follow purpose and authority.
That is one of the places where Māori GIS has most changed my general technology practice. The normal IT sequence often begins with platform, architecture and data model. The stronger sequence is often kaupapa, people, authority, information, then technology. What are we actually trying to do? Who should make the decisions? What relationships sit around the information? What must remain controlled? What capability already exists? Only then should the software become the interesting question.
The long arc
The technologies make an easy timeline: acetate, desktop GIS, scanned plans, CDs, DVDs, ArcReader, Google Earth, Google Groups, Facebook, Facebook Live, cloud GIS, Zoom, QGIS, StoryMaps, LiDAR and AI. That looks like the story because technology changes are easy to see.
The deeper history is much more consistent. At the beginning we were asking whether a historical boundary could be trusted. We are still asking whether a boundary can be trusted. We were asking who supplied information and whether it could be copied. We are still asking who supplied information and whether it can be uploaded, shared or published. We were asking what happened when the mapper left. We are still asking what happens when the person who built the system leaves.
We were trying to make GIS available to people without specialist software. Today the software barrier is much lower, but people still need support, governance, confidence and access to the right information. We were trying to make mapping useful rather than merely impressive. That remains the objective.
The technology matters because every generation of tools changes what is possible. CDs and DVDs allowed larger amounts of data to move. Google Earth made three-dimensional geography accessible to people who had never opened professional GIS. Web mapping removed installation barriers. Facebook and Zoom reduced geographic barriers to training. QGIS reduced software-cost barriers. Government web services made public information easier to obtain. AI is beginning to reduce specialist technical barriers.
Every time one barrier falls, the next becomes more obvious. Eventually the questions are no longer technological. What should we map? Who controls it? What should remain private? What does the map mean? Who has authority to say? What responsibility do we have to the people who will inherit this information? Those questions were there all along.
Leaving capability behind
If there is one idea connecting most of the work I have done around Māori GIS, it is the desire to leave something useful behind. At first that meant maps, then data and resource kits, then computers and training, then networks, conferences and online groups. Eventually it became something less tangible: confidence, relationships and independence.
A person who knows enough to ask better questions. A hapū that can maintain its own mapping. A participant who knows where public data can be found. A group that understands that it does not have to surrender its information to an external provider in order to receive technical assistance. A young person who realises GIS is something they can use themselves rather than something done elsewhere by specialists.
That kind of capability is harder to measure than map production. There is no neat dashboard for it. You notice it later when somebody who once came to a workshop is doing the work themselves, or teaching somebody else, or correcting you. You notice it when a group decides not to publish a layer because it understands the consequences. You notice it when the software changes and the people can move their data rather than starting again.
There is an intergenerational idea in this that I did not fully appreciate when I began. The most sustainable GIS is not just one that survives a software upgrade. It is one that leaves enough knowledge, context and authority with the people and organisation that somebody in the future can understand what was done and why. The information should not become an orphaned technical artefact separated from the relationships that gave it meaning.
The map is not the whenua
The old expression that the map is not the territory has been repeated so often it has become almost decorative. Māori GIS made the idea literal for me. The map is not the whenua. The polygon is not the relationship. The point is not the history. The label is not the full meaning of the name. The database is not the knowledge.
GIS is a representation. An extraordinarily useful one, but still a representation. The danger begins when the representation becomes easier to access than the knowledge it represents. A public map can become the source people rely on simply because it is available. A simplified boundary can become definitive because the methodology behind it has disappeared. A shortened name can outlive the context in which it was created.
The answer is not to make fewer maps. It is to make better relationships between maps, people and their sources. That means provenance. It means allowing uncertainty to remain visible. It means retaining authority with the people who hold it. It means remembering that a technically perfect GIS can still be wrong in every way that matters.
The more Māori worldview influenced my practice, the less comfortable I became with the idea that the map was the centre of the work. The map was one tool within a much larger set of relationships. Whenua was not a blank surface waiting for the GIS to describe it. People were not simply data providers. Knowledge did not become the property of the technical system when it entered a database. Capability did not mean making everyone dependent on the expert. Good work required listening, reciprocity, responsibility and an understanding of where my own authority ended.
From DVDs to AI
I still think occasionally about the ten-DVD GIS package because it captures its period almost perfectly. At the time, it felt like a way of breaking open access to mapping. Large amounts of information, imagery and GIS functionality could be put directly into people's hands without requiring the full professional environment that had previously surrounded it.
It was ambitious, and it was also ten DVDs. Today that amount of information could move across a good internet connection without the user thinking about the physical files at all. The map might open on a phone. The data could stream from a national service. An AI assistant could help write the expression required to classify it.
It is easy to look back at the old systems and see how primitive they were. I prefer to see the continuity. The DVDs were trying to solve a capability problem with the technology available at the time, and so are the modern systems. The real question has never been whether the software is sophisticated enough. The question is whether the technology leaves people with greater practical authority over their own mahi and information.
That is a much tougher test. A technically advanced platform that creates permanent dependence on an external provider may produce less genuine capability than a modest local system people understand and control themselves. A simple map built by the people who need it can be more valuable than an extraordinary national platform built for them. The best system is not necessarily the one with the most features. It is the one that still serves the kaupapa when the consultant, grant, software version or fashionable platform has gone.
That is why I remain fond of State of Unknown. It sounds like failure, but usually it is the opposite. It means we have reached the edge of what the evidence, authority or relationship can support and decided not to pretend otherwise.
Looking back, I began this work believing the difficult part of mapping was getting information into the computer. I no longer think that. The difficult part is understanding what the information is, where it came from, whose authority it carries, what responsibilities come with holding it, what should happen to it and when the computer should stay out of the way.
GIS can draw almost anything now, and AI can describe almost anything. Neither capability gives us the right to pretend we know more than we do. Sometimes the most accurate thing we can put on the map is an unresolved boundary. Sometimes the most respectful thing we can do with information is not publish it. Sometimes the best technical contribution is to help somebody else build the system. Sometimes the strongest professional judgement is the admission that the evidence stops here.
That is not where the mapping failed. It is where the map became honest.