Last Monday and Tuesday Consortech , one of the most creative providers of geospatial solutions I have been fortunate to know in North America, hosted conferences in Quebec City and Montreal. Pierre Genest, President of Consortech, arranged for Sam Bacharach of the Open Geospatial Consortium (OGC), Don Murray of Safe Software, Xavier Lopez of Oracle , Carl Pelletier of the Ministere des Ressource Naturelles et de la Faune de Quebec, and myself to speak at the conference. Between speakers, Denis Piche of Consortech demonstrated the technology discusses in the presentations.
Sam Bacharach (OGC)
Sam is one of the few people who can make a presentation on standards interesting. His main point was that the motivation for standards is economic, and that it makes financial sense for OGC members to support standards. The OGC was founded in 1994, and Sam said that Oracle and Autodesk joined in October and November, respectively, of that year.
There were two of Sam's slides in particular that I found interesting. The first is the list of organizations with which the OGC has alliance partnerships, which includes well known organizations like the IEEE, W3C, ISO, and GITA. But the list also includes organizations that geospatial folks may not have heard of, like the International Alliance for Interoperability (IAI). The IAI is responsible for a standard called the Industry Foundation Classes (IFC). which defines standards for classes of objects used in buildings like heating and ventilation, utilities, and structural components. One of the interesting outcomes of the alliance between the IAI and the OGC is the upcoming IFC2x3g standard which allows building coordinate systems to be georeferenced to a known geographical coordinate system. This is a major turning point because it implies that architects and engineers now need to be concerned about where on the globe the building they are designing is situated. The UK's "right to light", Noise abatement programs in the EU , and the City of Vancouver's Vancouver View Cones are examples of how important it is to know where a building or infrastructure is going to be located geographically. Sam said that Microsoft and Google probably spend around half a billion dollars annually on data including increasingly 3D, but it is clear that they cannot collect and maintain this data all by themselves. Clearly we are going to need new standards to democratize the creation and maintenance of geospatial data including buildings and utility and telecommunications infrastructure.
The other slide of Sam's that is worthy of wider dissemination is the list of currently supported OGC standards. This is quite an impressive list because a high proportion of these standards including the Simple Feature Specification for SQL (SFS), the Web Mapping Services (WMS), Geographic Markup Language (GML), and the Web Feature Service (WFS), have become the standard for geospatial data exchange for governments worldwide. I would also suggest that a recently adopted standard, Web Geoprocessing Service (WPS), is very worthwhile taking a look at.
An area that is notably not well represented in the OGC standards is 3D, probably for the reason that this is an area where the technology churn is still very high. However, LandXML is an example of a widely adopted standard for roads and highways, the IAI's Industry Foundation Classes is an example of a well-defined Building Information Model (BIM) standard and the latest IFC standard, IFC2x3g, includes location. Another standard called CityGML is going through the review process at the OGC, and is already supported by some organizations such as Oracle.
Xavier Lopez (Oracle)
There are two topics that I felt were particularly important to note in Xavier's presentations. The first was that geospatially-enables relational database management systems support the concept of a single point of truth, and allow the exchange of information between geospatial products from different vendors. Because of the wide adoption of the OGC's Simple Feature Specification for SQL, ESRI, Intergraph, Autodesk, MapInfo (Pitney Bowes Software), Bentley, and others all support storing geospatial data in an RDBMS. This means that you really can have a single copy of geospatial data accessed by multiple applications from different vendors. The advantage of using an RDBMS as opposed to a simple file like Shape, for example, is that data access is standards-based, SQL and ODBC or JDBC.
But if you want to stay with files and have the flexibility to move to a spatial RDBMS in the future, the OSGEO FDO API is designed to support the concept of single point of truth, but supporting a wide range of file formats and well as spatial RDBMSs. FDO differs from the widely used OSGEO GDAL/OGR API in that it is designed as a read/write API for geospatial data stores, including files and spatially-enabled RDBMSs.
The other part of Xavier's talks that I found interesting were that while Oracle 11g supports 3D coordinate systems, 3D spatial data types, TINs, and point clouds, none of the spatially-enabled RDBMs support the data types that are necessary for architectural and engineering applications, whether 2D or 3D. In my opinion this is a significantly larger market than traditional GIS, and because building and infrastructure models are becoming much more widely adopted and much larger, I would suggest that being able to store architectural and engineering information in a queryable database is becoming essential in the future, especially as we are now able to model entire cities, including the inside of buildings and underground infrastructure. Xavier suggested that Oracle is moving in this direction.
Don Murray (Safe)
Don Murray from Safe Software, who develop and maintain FME, which is the de facto standard for geospatial data translation software, made several important points. First of all Don showed the latest version of one of my favourite slides which I have plagiarized several times. The number of file formats supported by FME has reached 220, and there are no signs of abatement in the world's creation of new file formats. Don's important point about this is that the challenge is not format, but data models, and that (in my words) the single largest contribution to the progress of world civilization made by Safe is what I would call their metamodel, because the data contained in all of the formats supported by Safe flows through this model.
A very important point that Don made is that although FME users do use FME to translate between different file formats, therby creating multiple copies of the same data, Safe's preferred architecture is what Don calls on-the-fly data translation which avoids the creation and storage of redundant data in new files. For example, FME supports FDO, which means that users have on-the-fly read/write access to all of the file types that FME supports.
I can't help adding that Don also said that FME now supports AutoCAD Map 3D Object Data, which is something that will make all Map3D users and other folks needing to access data in DWG files very happy.
Another very interesting point that Don made and that is a harbinger of things to come is is that Safe is moving to support 3D data formats including BIM standards such as IFC. This means that the line between architectural and engineering data, which is also spatial, typically includes the insides of structures, and where precision is essential and geospatial data, which is typically outside of structures, is blurring.
Carl Pelletier (MRNF)
A speaker coming from the trenches as it were was Carl Pelletier who works for MRNF. One of his slides was so clear in presenting what the problems are with silos or islands of information that I'm including the original here with my own attempt at an English translation. At MRNF Carl said that they currently support 120 different systems maintaining about a terabyte of data.
Impact of islands of information
- Many different applications with many different data formats
- Data redundancy
- Problems finding data
- Information exchange is based on the import/export of paper or files
I think that Carl has captured the fundamental problems arising from silos or islands of information and the reasons that many folks are now looking at ways to implement a single point of truth within their organization.