This is a common challenge for many Cadastral or Cartography projects, which at the time 2000-2010 integrated Microstation Geographics as a spatial data engine, considering reasons such as the following:
- The arch-node management was and continues to be extremely practical, for cadastral projects.
- The DGN is an attractive alternative, considering its versioning in the same file, which has not changed in 15 years, contrary to other formats in which we have seen many versions incompatible every three years.
- At 2002 free software was a dream away from what we have today.
- OGC standards did not weigh even on proprietary software.
- The shp files were limited for high draft projects and the spatial bases were still very closed to non-standardized schemes that compromised the performance of servers ... and silver.
- Remote connectivity was incipient compared to what we now have.
Thus, implementing a GIS based on a "linked CAD" scheme was a viable solution, although usability was sacrificed for attractive presentation purposes. The VBA API was abundant for programming transactional management routines connected to ProjectWise for the control of physical files and the possibility to use GeoWeb Publisher for spatial analysis from the server, although the publication would be limited to ActiveX in Internet Explorer (which in that year was the Single browser).
The problem is not to have evolved gradually and instead of moving to Geospatial Server or more robust versions of ProjectWise, wanting to make a GIS survive from physical files, having all Oracle Spatial licensed potential and the ability to develop. So that was our challenge.
1. The database: Postgres, SQL Server or Oracle?
In particular, I would have preferred the former. But when you're in front of a non-service-oriented transactional system but working well, where part of the logic and integrity is like PL in the database, switching to an OpenSoure database is not an emergency. No, unless you aim to develop a new version of the system that is not in the immediate term.
Nor is it a Taliban action to disparage everything that smells private. So staying with Oracle is a wise decision, if it is working well, if the size and requirement is broad, if it is well designed, protected and if support is being leveraged. Theme for another occasion.
So what remained was to develop functionalities for the data to be migrated to this database, publishing services and transactional management tools for vector data.
To control the roles and users that were previously managed from ProjectWise, a modular tool was created that allowed:
- Manage users and roles from the BentleyMap VBA.
- Assign from the user with administrative rights, right to departments and municipalities.
- Assign right to cadastral file by project.
- Right to tools available in the modules of Construction, edition, publication, consultation and administration. In this way, only new applications are created and users appear according to their specific role or assignment.
- This login panel also simplifies the common complexity of the BentleyMap projects, such that with just entering you will see the tree of categories and attributes defined in the Geospatial Administrator.
A panel of this solves problems of little understanding and risks of new users to functionalities like Data Interoperability. That is another roll, because Bentley edits natively in Oracle Spatial, which is wonderful but also risky if you do not have transactional control.
Thus, for example, the Construction module had the following tools:
- Assign Features
- Geographic Linkage Wizard
- Batch Space Migration
- Delete objects
- Edit polygons
- Export Shp / CAD
- Import Shp / CAD
- Geological Migration
- Migration Geopunto
- Georegion migration
- Register map
- Link Geo-Line
- Link Geo-Point
- Link Geo-Region
Complementary tools were gradually added, including some to directly edit the Geospatial Administrator.
- Administrator to view features
- Topological Analysis
- See SAFT
- Browse Feature
- Convert Curve to LineString
- Create Features
- Create properties
- DBConnect configuration
- DBConnect Inquiry
- Edit feature Xfm
- Edit Project Xfm
- Remove Features Xfm
- Parcel identification
- Modify Symbology
- Over-write features
- Thematization by classes
- Thematics by drop-down list
- Xfm Utilities
2. Data: Migration from DGN to Spatial Basis: Oracle Buider or Bentley Map?
The most interesting challenge in this was that a controlled migration was required, and bearing in mind that the DGN files having been updated for more than 10 years might have topology problems - a real madness.
Indeed it was. The main problems of the maps are here:
- The modification of a parcel in the border of the file (sector or zone) implies that there must be modification of both, including the coincidence of nodes in cases like when in a sector is a single line but in the neighbor that line is segmented.
- There are files that after 300 maintenance transactions stored in the DGN history can be corrupted.
- There are more complex problems that are not controllable in a cabinet, such as when a property overlaps another neighbor in another file, for amounts that can not be resolved on the map, as it would involve doing field inspection to avoid affecting a third party.
- Bad practices, such as the inclusion of maps in different projections, in this case had sectors in NAD27, although the standard was WGS84. In extreme cases adjustments were made between data from different projections, to the perverse.
The solution was a Wizzard type tool for mass migration, which can migrate individually a map, several or even all of a municipality (town hall) or department.
Basically what the tool does take the data of the Geographics project and promote them to features of Benltey Map, then makes a series of validations, such as:
- One-to-one relationship between geometry and database,
- Validation of missing duplicates,
- Validation of area-centroid consistency,
- Validation of map objects with inactive objects in the database,
- Validation of topology with respect to existing topologies in the spatial base
After the validations, the panel allows to aggregate information in a massive way, such as measurement method and quality control standard of that data.
Finally, it posts to the database, finally generating a report. From the saying to the fact there is a tremendous stretch, but finally adjusted to the whims of Oracle Spatial that are not as far-fetched as those of Bentley and his way of seeing complex estates or many plots of vertices.
3 The publication: Geoserver or MapServer? OpenLayers or Leaflet?
A viewer was built using OpenLayers and some plugins. For the first time after 10 years of abandonment of the development of the space part, was visible a new viewer that replaced the ActiveX of GeoWeb Publisher. We used the MapFish code for the expression, geojson to control the side tree, since Geoserver served the OracleSpatial served layers.
Finally, the replacement of technologies was done according to the following graph. As you can see, a combination of free code, maintaining the database and land management using proprietary software.
4. Construction and editing, direct to Oracle Spatial. Bentley Map or QGIS?
This is another story. Bentley Map edits native in the spatial base, which generates conflicts if you will not work with a Web Feature Service Transactional (WFS). The conflict is:
How to solve a rule of not allowing overlap of topology, if it is being edited and when wanting to post reports that the object affects itself?
This is solved by versioning before, editing directly and validating that when posting, if something fails the versioning is recovered leaving the transaction complete but in a failed state.
Another problem that had to be solved is the massive data entry, considering that the users had to stop using Geographics and there were several projects raising massive cadastre.
This was easy because only a tool similar to the one used to integrate the data in Microstation Geographics was made, facilitating with the potentialities of BentleyMap and with a more controlled assistant.
The image shows how this tool was developed, with some particularities, such as the creation and registration of vertices and the inclusion of the Puntoparcela, as list functionality in case the method of measurement of some vertices did not meet a certain quality standard.
Definately this flow was very good, because users knew which tools most often used. It was necessary to make them change their mentality between the step of multiple features to management by levels, promoting new benefits so that they forgot the archaic Microstation V8 2004, such as the WMS service, the transparencies and native recognition of DWG files of recent versions; Not to mention interoperability with kml, shp and gml for the more astral.
Tools were also made for cadastral maintenance, having the option of editing directly in shapes or lowering them to arc-node for complex cases.
5. Client for municipalities via GML. QGIS or gvSIG?
QGIS. But that, is another story to tell later.