Landonline

From register to transaction system

When Land Information New Zealand was created in 1996 it inherited two national record systems that had grown up for different purposes. The survey and cadastral side came largely from the Department of Survey and Land Information. The land-title side came from the Department of Justice Land Titles Office. Both were already partly computerised, but much of the legal and survey record still depended on paper, microfilm, office indexes and staff who knew how the pieces fitted together. Landonline brought survey, title and transaction records into one electronic operating environment.

The previous chapter ended with a national digital cadastre already in existence. DCDB held a maintained parcel framework and had become useful far beyond the survey office. What it did not do was turn the whole survey and title business into an electronic transaction system. Survey plans, title instruments and supporting records still moved through processes that had been designed around paper. Searches depended heavily on where records were held. The next step therefore involved records, law, professional practice and organisational change as much as spatial data.

Government approved development of Landonline in November 1997. Later Auditor-General analysis described the programme as initially divided into two phases. The first would create the electronic record base, automate internal processing and provide remote searching. The second would allow surveyors and conveyancing professionals to lodge work electronically, with the system applying business rules and validation before or during registration. A third phase was added later to finish the move away from paper and make electronic lodgement the normal route for almost all work.

The programme was large even by the standards of national information systems. The Auditor-General later described the first system phase as about $40 million alongside a historical back-capture programme of roughly $100 million. Contemporary ministers described millions of title documents and more than a million survey plans, together with geodetic records and cadastral data, being converted or linked into the new environment. These figures varied with what each report counted, but their scale is consistent. Landonline required changes to statutory processing and professional practice. The programme rebuilt a national statutory information service while it continued to operate.

Stage one

Stage One concentrated on making the existing record usable electronically. Titles were converted into computer registers. Survey plans were imaged and indexed so they could be found from the spatial and textual records around them. Cadastral information was progressively moved from DCDB into the new structure, while geodetic information and survey references were brought into the same operating environment. Internally, LINZ staff could process work against electronic records instead of relying on the old combination of paper registers, mainframe indexes and district collections.

Remote search was one of the earliest changes visible to customers. Otago and Southland were the first areas brought into the new service from April 2000, and Dunedin became the first customer region for the Stage One release later that year. The change went beyond viewing an image on a computer. A customer no longer had to depend on the physical location of a title or survey record in the same way. Search could become national, and the same electronic record could support both LINZ processing and professional access.

The rollout remained staged because the source material was staged. Land Districts had different collections, conversion status and workloads, and the underlying cadastral information had to be cut over in a controlled sequence. By early 2002 the remote service had spread northward, with Auckland among the last major regions. More than 500 remote access licences had been issued by March 2002, with many licences covering several staff. The customer base was still specialist, but Landonline had already changed the geography of access before electronic lodgement itself became routine.

On 31 May 2002, the Minister for Land Information signed what the government described as the last paper certificate of land title issued under the old system. The event coincided with completion of Stage One and the opening of the new LINZ Centre in Auckland. The symbolism was unusually apt. Paper had not vanished from land administration, but the authoritative register was now electronic and national. A title record no longer depended on issuing a fresh paper certificate as the principal operational expression of the register.

Stage One also changed LINZ's own working geography. Under the older arrangements, district offices and specialist registries held records whose location was part of the way the service operated. Converting those records into a common electronic environment allowed work to be handled with less dependence on the office where a particular register or plan happened to sit. That promised more consistent national processing, but it also meant that long-established local routines and knowledge had to be translated into common procedures. The database centralised information and altered how that information was lodged, checked and maintained. It standardised enough of the work around that information for processing to become less tied to place.

Title processing and survey processing required different approaches to automation. A title register recorded registered ownership and interests under land-transfer law. A cadastral survey record contained measurements and evidence used to define parcel boundaries. Both could be searched through Landonline, but the system could not safely treat them as two versions of the same record. One of the programme's harder design problems was therefore integration without collapse: users needed links between title, parcel, survey and geodetic information while the system preserved the different authority and purpose of each.

Converting the archive

The record conversion behind that moment was more complicated than scanning. Survey plans could be imaged so the original evidence remained visible, but Landonline also needed structured data that software could query and validate. Titles had to become electronic registers rather than pictures of certificates. DCDB geometry had to be transformed into the new cadastral structures. Survey conversion used bearings, distances and other observations from the cadastral record to improve the position of a large proportion of parcels beyond the quality inherited from record maps.

The programme therefore made deliberate choices about what should become data and what should remain an image linked to data. A survey plan contains much more than a set of parcel lines. It may include observations, marks, annotations, diagrams and evidence whose significance cannot be reduced safely to a single vector layer. Imaging preserved that documentary record while structured survey observations and parcel relationships could be captured separately where the new system needed them. This created a hybrid electronic archive in a useful sense: searchable structured records connected to source documents rather than an attempt to translate every historic mark into database fields.

Contemporary technical papers describe survey conversion as a major effort in its own right. By the early 2000s a large share of parcel survey information in intensively surveyed areas was being captured from survey records so that the digital cadastre could be strengthened. The work treated cadastral position as an adjustment problem rather than a drawing exercise. Bearings, distances and observations could be combined, tested and adjusted so that the new coordinates reflected survey evidence more closely than the old map-derived DCDB linework.

The historical labour behind this remains less visible than the programme name. LINZ survey and title specialists defined rules and checked converted records. Data-capture staff, contractors and technicians scanned, indexed, keyed and reconciled material. Database and systems staff kept conversions moving between environments. Surveyors and conveyancing practitioners tested how the records behaved in real work.

A contract goes wrong

The conversion programme also produced one of the least tidy episodes in New Zealand’s geospatial history. On 16 June 1999 Terralink signed a subcontract with EDS for Landonline work. Terralink had been created only three years earlier from the commercial side of the old DOSLI organisation, so it carried substantial national mapping and land-information capability into the contract. The work formed part of the core programme. It sat close to the technical lineage from which many of the records and production practices had come.

By 2000 the subcontract was in serious difficulty. The government later said the cost and time required had been underestimated, creating losses and risks that Terralink could not absorb. A KPMG review was commissioned, the Crown provided a secured loan late in 2000, and on 15 January 2001 Terralink entered receivership at the board’s request. Contemporary government material described about 200 full-time Terralink staff and roughly another 80 people contracted to the EDS project. The company continued to trade under receivers while the consequences were worked through.

National record conversion contains work that is difficult to see from a sample. Exceptions accumulate. Old records vary. Quality problems do not distribute evenly. Staff who understand obscure cases are scarce. A conversion rate that looks reasonable in a pilot can collapse once the difficult districts and unusual records arrive. Terralink’s failure does not prove that any one of these factors caused the losses, but the episode demonstrates the financial danger of treating national spatial conversion as routine production.

The disruption had operational consequences. In an October 2001 speech to surveyors, the Minister for Land Information noted that survey conversion had been affected by the receivership and that EDS was working through a revised delivery plan. Otago and Southland were close to completion and Christchurch was under way, while national completion was then expected later. The survey and title system still had to keep functioning during the commercial failure. That separation between statutory continuity and contractor difficulty was one reason the programme could not be treated as an ordinary outsourced software build.

Changing the law

Electronic search could be introduced by changing how records were stored and accessed. Electronic transactions required more. The law had been written for instruments, signatures, certificates and plans that people could hold, sign and lodge physically. Landonline Stage Two therefore needed legislation that recognised electronic records and electronic lodgement without weakening the principles that gave the survey and title systems their authority.

The Land Transfer (Computer Registers and Electronic Lodgement) Amendment Act 2002 and the Cadastral Survey Act 2002 were both enacted in May and commenced substantially from 1 June 2002. The land-transfer amendments allowed computer registers, electronic instruments, electronic workspaces and digital authentication to sit inside the existing registration system. The cadastral legislation explicitly allowed cadastral survey datasets to be lodged and processed electronically. The legislation modernised the transaction medium while retaining the fundamental land-transfer principles and the Crown-backed guarantee of title stressed in parliamentary material.

The legal design also recognised that electronic processing required control over who could do what. Conveyancers and licensed surveyors would work in electronic workspaces and authenticate submissions using digital certificates. Routine transactions could be checked against business rules before they reached the register. Survey data could be tested electronically for compliance with survey requirements. A system that accepted structured data could perform checks that were impossible when its only input was a paper document arriving over a counter.

At first, electronic lodgement was an option rather than a universal requirement. The legislation allowed the government to make electronic lodgement compulsory later by regulation or order once the system and its users were ready. That sequencing reduced the risk of changing the law faster than the service could operate. It also created a long transition in which paper and electronic processes coexisted, sometimes awkwardly, while professions adapted.

The benefits case depended on more than faster searching. If professionals still prepared paper, LINZ staff still re-keyed it, and the electronic register merely stored the result, much of the old handling cost remained. Stage Two was intended to move structured information closer to the person who originated it. That changed where data quality was checked. A professional could discover a routine error during preparation, not after a couriered or counter-lodged instrument had entered a formal examination queue.

Automation also forced business rules out of manuals and tacit practice and into forms that software could apply. That was straightforward for some tests and difficult for others. A missing field, invalid reference or mathematical misclosure could be checked mechanically. Questions involving evidence, legal interpretation or professional judgement still needed a person. Landonline therefore did not automate the professions out of the process. It shifted the boundary between mechanical checking and expert decision.

The survey becomes data

For surveyors, Stage Two changed the relationship between a plan and the national cadastre. A cadastral survey dataset could contain structured observations, coordinates, parcel relationships and supporting information rather than arriving only as a finished drawing. Contemporary Landonline papers by Tony Bevin, Anselm Haanen and Neil Sutherland describe a system designed around digital submission, automated validation and closer integration of geodetic, cadastral and title information. The surveyor was no longer only submitting a document for LINZ staff to interpret. The surveyor was becoming a direct producer of structured national data.

Pre-validation was central to the idea. Software could test whether required fields were present, whether observations closed, whether parcel relationships were coherent and whether specified business rules had been met before final submission. Some checks remained professional or legal judgements and could not sensibly be automated. Others were mechanical enough to move into software. The attraction was obvious: find routine errors while the surveyor still had the job open rather than after formal lodgement had started an examination cycle.

The architecture also encouraged standardised exchange. Contemporary papers discuss LandXML as a developing mechanism for moving survey information between professional software and Landonline. That work connected New Zealand cadastral automation with international efforts to describe survey observations and land-parcel information in structured formats. The product name is less important than the direction. A survey dataset could move between systems without being flattened first into a printed plan.

Nationwide e-survey functionality became available in November 2003. Survey conversion was completed in December that year, with LINZ reporting about 1.35 million converted parcels loaded. The system could generate plans from submitted digital data and support online certification by territorial authorities. Drafting, checking, certification and database updates moved into a shared digital process, changing who performed parts of the work as well as making some stages faster.

The dealing becomes data

The title side went through a comparable change. Lawyers and conveyancers had long prepared instruments whose legal effect was registered by the land-registration system. E-dealing moved routine work into an electronic workspace where transaction details could be assembled, validated, digitally certified and submitted. For straightforward dealings, software could compare the submission with the current register and apply business rules before registration.

The statutory system retained authority over the register. Digital certificates identified authorised practitioners and supported the trust model needed for electronic signatures and certifications. Responsibility did not disappear simply because the transaction became easier to transmit. In some respects it became more explicit: the practitioner supplying structured information was responsible for getting that information right before submission, while LINZ could automate checks that previously required staff to re-key and compare paper forms.

The efficiency gains became substantial once use increased. Later Auditor-General work found that electronic processing reduced double handling, paper storage and internal re-entry, and shifted more data capture toward the professionals who originated the transaction. Processing that had taken hours, days or sometimes longer could occur in minutes for routine work. The gain was partly purchased by moving preparation and validation effort outward. A solicitor or surveyor could do more from the office, but also had to maintain compatible systems, authentication, staff capability and familiarity with Landonline’s rules.

The change also made service availability a new operational dependency. In the paper system an office could continue some work during a local computer problem. In the electronic system, outages could stop searching, validation or lodgement for many users at once. Ministers were already acknowledging such disruption during the early rollout. Landonline traded the limits of paper distribution for the strengths and weaknesses of a national online service.

Customer adoption was also a technology problem in the ordinary sense. Firms needed computers, network connections, supported software and staff who could use the new workflows. A large surveying or legal practice could spread those costs across many transactions. Smaller practices had fewer people to absorb training and system administration. LINZ used training, fee settings and staged compulsion to push adoption while giving the market time to adjust. The change altered government processing and the equipment and skills required in private professional offices.

The professional organisations were deeply involved because Landonline redistributed responsibility. A system that pre-validated a survey dataset or dealing could reduce LINZ handling, but the original professional still certified the work. Digital certificates tied submissions to named authorised users. This was a technical security mechanism and a professional accountability mechanism at the same time. The system became faster partly because it trusted structured information supplied by people whose authority and liability were already established outside the software.

From optional to normal

Electronic lodgement took several years to become the default. LINZ trained users, changed fees, released additional functionality and worked with the New Zealand Law Society, the New Zealand Institute of Surveyors and other professional groups. Consultation helped identify professional requirements and implementation problems. Surveyors and conveyancers were being asked to alter workflows that affected statutory rights and professional liability. Poor adoption would have left LINZ maintaining two complete channels indefinitely and would have undermined much of the expected return from the investment.

By 2006 the government had decided that the remaining paper route would be phased out. Mandatory electronic lodgement was introduced in stages rather than on one date. Discharges of mortgages became compulsory electronically in May 2007, followed by major categories of routine title dealing. All cadastral survey datasets had to be lodged electronically from September 2007. The last categories of title transaction handled by lawyers and conveyancers became compulsory through Landonline in February 2009.

Compulsion changed the organisation behind the service. Once most work arrived electronically, LINZ could consolidate processing and reduce public counters built around paper. Public counters in Auckland, Wellington and Dunedin closed in early 2009, while processing was concentrated into fewer offices. Staff roles changed as routine handling and re-keying declined and more work centred on exception processing, quality, support and system administration. The same technology that let customers work nationally also allowed LINZ to reorganise where its own work was done.

Professionals experienced the transition differently. Earlier branch closures had already raised concerns about losing local access to historical paper records, including concerns from Māori communities where land history depended on material not fully represented in the new electronic system. LINZ considered copying or preserving records that fell outside the normal conversion scope. The conversion covered operational records and left parts of the wider archive undigitised. A system can become more accessible for routine national business while particular local histories become harder to encounter.

The long coexistence of paper and electronic channels helps explain why the programme's benefits were slow to become fully visible. Running two channels meant maintaining two sets of procedures, staff knowledge and exception handling. Users who stayed with paper could avoid some immediate change costs, but LINZ could not remove the infrastructure that supported them. The expected financial benefits depended increasingly on professionals using electronic transactions. Mandatory lodgement was the final organisational step needed to turn an optional digital service into the default national operating model.

A national utility

By the end of the 2000s Landonline had become something quite different from the project launched a decade earlier. Survey and title professionals searched nationally, lodged structured survey datasets and submitted most title dealings electronically. Business rules checked routine work. A large share of straightforward title transactions could update the register with little or no manual re-entry. The system had moved from digitising records to mediating the transactions that changed those records.

Landonline was a statutory production and transaction environment with controlled access, professional responsibilities and legal and personal records. It made the authoritative survey and title processes digital before many of their non-personal spatial outputs became easier to download and reuse through later LINZ services.

The system continued to require maintenance and development. Software had to be maintained, standards updated, security managed, users trained and business rules revised as law and practice changed. By 2010 the programme had passed through a third phase intended to remove most remaining paper handling, but the service would continue to be rebuilt and modernised in later decades. LINZ established a continuing electronic service. It was an operating model in which the national cadastre and title register could be changed through electronic professional transactions.

Chapter 13 ended when the parcel became maintained national data. Landonline went further. It turned the process of creating surveys, registering dealings, searching records and updating the national land system into digital work. A different national mapping problem was undergoing its own conversion. Topographic mapping also became a maintained database, but without the legal transaction machinery that made Landonline such a distinctive New Zealand information system.