PETER KINNE

The concept is familiar: take data from one environment to another without losing integrity, information or intelligence.
 
However, the integration of CAD and GIS environments continues to be an issue, especially for local government and agencies where information from designers needs to be placed in its true geographic context.
 
CAD design professionals focus on detailed information for a specific area or facility, which requires that integrity and completeness have priority over location. Information is annotated to reduce misinterpretation at the time of construction.
 
The CAD world is truly 3D in its drawing tools, storage and visualisation. The ability of design applications to cater for volumetric objects provides the designer with options to store data in a variety of ways – from 2D to full 3D, from blocks to full attribution.
 
However, GIS professionals have a different focus. In a typical GIS, completeness is the overriding aim. Spatial datasets such as the cadastre aim for adjacency over positional accuracy. Although the data originates from a surveyor, the priority is on visualisation and interfacing with corporate databases. This illustrates the fundamental difference between the disciplines of CAD and GIS.
 
The perception that CAD data is dumb and accurate while GIS data is smart, but less than accurate, is well entrenched in our industry. This is characterised by the delineation of the two sides of the IT workflow in an organisation. Frequently, it encourages CAD and GIS professionals to each perceive their data as part of a separate workflow, as it moves from one environment to the other.
 
Organisational structures home in on functional differences. CAD and GIS functions are often placed in separate departments, further entrenching demarcation lines and reducing the potential for change. Individuals tend to focus inward and usually find it easier to create work-arounds than to challenge the processes of another department, or push for integration of CAD and GIS.
 
Habits and conventions also create further obstacles to integration. Internal CAD designers see a plot as their output and a document as the culmination of their efforts. As a result, the annotation blocks they used ten years ago still provide the same desired hard copy output as more contemporary methodologies.
 
If a GIS department only supplies a CAD designer with non-georeferenced DXF files for feasibility or design work, this virtually ensures the lack of location and internal standards when the design data for the completed project is returned to the GIS department.
 
In summary, failure in the integration of CAD and GIS is not entirely the result of technical deficiency in either type of product. It is the result of the fact that the required outcome for the different professionals involved is similar, but not the same. For example, the storage file types for the technologies differ, based on the potential data that will be generated by the software. CAD caters for 3D and volumes, while GIS is fundamentally 2D/2.5D. A final issue is that CAD users see hard copy plots as the end result. In the GIS world, it rarely is. GIS professionals remove location when extracting data for CAD designers.
 
Why change all this? Current practices produce islands of information and redundant data. There is delay and duplication of effort as data is re-entered from one environment to the other, creating backlogs and reducing data completeness and quality.
 
The time taken to capture CAD data (up to 12 months) reduces the effectiveness of billing, maintenance and asset management activities.
 
Furthermore, digitising and attribution is a mundane task and not the most effective means of engaging a GIS professional. Inevitably, time and resource limitations mean that only those items deemed essential make it across to a GIS dataset from the CAD drawings.
 
There are a few solutions to this issue. Open standards databases allow CAD and GIS data to co-exist. Users can choose the application most suited to their discipline. This has been successfully implemented in Australia, enabling users to share data across CAD, GIS and web applications.
 
This solution has its limitations. Not all mainstream GIS and CAD products are truly compliant and external designers rarely have access to internal databases.
 
Another approach has been to use the ability of some GIS products to read CAD data directly as a layer. While it allows GIS products to read CAD data directly, users necessarily need to re-align the CAD layers into GIS layers and re-enter annotation as attributes. Additionally, this does not cater for the loop of data back into CAD for feasibility and design work.
 
More recently, CAD applications can now read and (in most cases) write GIS formats without translation. This demonstrates a recognition by CAD vendors that they are operating in a multi-vendor environment. It is also the approach that holds the most promise of advancing internal and external workflow.
 
All three approaches all focus on the technology. More fundamental changes in data standards and structures still need to be addressed.
 
For about ten years, designers have been able to store CAD data with location included. Additionally, the data structures have been available to store object data (attribute equivalent) with the spatial data.
 
Simply put, the mainstream CAD tools either include spatial capability as standard or offer it as an option.
 
CAD products are now GIS-ready (and have been for some time) and integration is available. This means that GIS departments can request attributed data in a specific projection and datum, as well as supplying data as GIS formats and structures.
 
The top tier CAD technologies are as enterprise-capable as their GIS counterparts. Designers are able to self-serve data from direct connection through to services such as OGC's WMS and WFS standards. Data level integration is an opportunity. The only barrier is resistance to change.
 
It is up to CAD and GIS professionals to recognise their common goals and consider the requirements of the downstream applications of their data. An understanding of the data value chain will enable them to capitalise on existing application investments and expertise to suit all users, regardless of where data is created or used.
 
The challenge is to understand the vendor's data level offerings that support the respective CAD/GIS professionals, and then work out how this inter-relates with workflow.
 
There is opportunity for benefits for professionals and the organisations they work for. Double handling of data is redundant. The entire process becomes a value cycle. The goal is geospatial; CAD and GIS data are managed as dependent resources.
 
Peter Kinne is with Karelcad in Brisbane.
 
Issue 34; April – May 2008