First, let's examine the business perspective.What is the value and return on investment for interoperability? Many of you have encountered the problem of operating electric appliances and computers across country boundaries.This problem can be solved using electrical adapters.Clearly it is much more economical for a traveler to use inexpensive adapters than to expect a country to change its entire electrical infrastructure to achieve full interoperability.It is interesting to note that the same type of consideration applies to GIS software vendors.The software infrastructure on which modern GIS systems are built cannot be replaced overnight.This is the reason that the GIS industry is seeing the emergence of adapter and interoperability kits to achieve multi-vendor interoperability.One example is the Oracle Interoperability Initiative (see image at right; click for larger graphic).Another example is the Open GIS Consortium (OGC) WMS standard.
Let's look at an example from a different industry.The music industry (content providers and electronic manufacturers) agreed very early on a format for music CDs.CD players are able to play disks independent of the manufacturer.This enabled CD technology to mature rapidly and to be universally adopted.This seems a very obvious example, but how about video games? Today the "big three," Nintendo Gamecube, Sony PlayStation, and Microsoft Xbox, do not have technology that is compatible.If you walk in your favorite store, you will see the same games in multiple versions. Game publishers have to develop cross-platform software, test on multiple hardware, release several versions of their games, and create a duplicate support infrastructure.I invite readers to reflect on the issues that have led to full interoperability of music CDs and completely proprietary games media.I suspect that they will discover issues that are comparable to the GIS software industry.GIS data producers are creating vector and image data in a variety of incompatible formats.Will we ever see convergence?
All of this in itself is quite complicated, but that is not where it ends.Interoperability often relies on some technologies being universally adopted.What about intellectual property? Organizations invest substantial amounts of time, effort, and money in innovative technology development. Corporations will have to see a return on their investment.What are the business dynamics that will help to decide between making technology public and available license-free or keeping technology proprietary for a competitive advantage? This is an issue that confronts business leaders on a continual basis and impacts standards organizations.In the October 2003 article in CIO Magazine, "The Battle for Web Services," C.Koch discusses some of the patent and intellectual property issues encountered by the World Wide Web Consortium (W3C) and the Organization for the Advancement of Structured Information Standards (OASIS) while developing Web services standards.
In the technical aspect of interoperability, I like to use the example of e-mail.A message will appear in two different manners on two different e-mail systems.
Welcome to Cancún
Welcome to Cancun
A very simple example, but it illustrates that technical interoperability is not a reality even at the simplest level of e-mail text messages.What can we say about a complex GIS network model or complex GIS geometry types? In the past several years, we have seen a significant amount of work to improve interoperability of GIS systems.The Open GIS Consortium (OGC) has been successful in creating standards for such complex subjects as geodetic coordinate systems, following the lead of the European Petroleum Survey Group (EPSG), and advanced GIS models in the creation of Geography Markup Language (GML).
There is still much to be done.The recent Oracle Interoperability initiative mentioned above found standardization to be lacking for a basic GIS data type, point features.Vendors have implemented different "point models." Each vendor has to deal with mapping "oriented" points to "plain" points using data from different vendors.This example illustrates the information loss incurred when mapping data models as different models support different levels of richness.
Moving into more complex models that include topological structures, feature relationships, connectivity, georeferenced images, and more becomes a much greater challenge for mapping and transforming data models.OGC found it was fairly easy to achieve interoperability of a Web Map Server (WMS) where maps are displayed as images.In this situation, interoperability issues revolve around image formats (transparency), resolution, coordinate systems and other technical problems that have been largely solved.
On the other hand, the Web Feature Server (WFS) deals with complex GIS models encoded in GML.Clearly a much harder problem as it involves making rich object models compatible.WFS standards are an extremely important first step toward GIS interoperability.But this is just the tip of the iceberg.
In addition to data model issues, interoperability proponents need to address the issue of software components working together independent of platform.Earlier distributed object systems required a high degree of integration, which forced deployment on a single operating system.Web services provide a framework that enables components to interact across platforms.Web services generally encapsulate data, security policies, and expose an API accessible simple XML message.As the IT industry creates standards for Web services, interoperability will benefit through the powerful technology this enables.Web services include the concept of services discovery. Discovery is the process by which a software component can search for complementary functionality, find it, and use it as needed.For example, if you need to geocode addresses, your GIS could search the Web for a geocoding service that can do the job.These concepts introduce significant challenges.How are Web services advertised so that services requestors can find them? Most often they are advertised through service brokers (or catalogues or electronic yellow pages).If services are found, who is authorized to use it? How much will it cost? What will the quality of the results be? In order to have fully functional Web services we will need a Web services infrastructure that addresses these questions and more.
Interoperability represents a vision that will play a very important role in the future of geospatial technologies.The GIS industry - vendors, user communities, government organizations, research institutions, OGC, and more - is taking a leading role in solving complex problems associated with geospatial interoperability.Service discovery is an immature technology, but it will be exciting to watch it mature along with interoperability. Much has been achieved, but there are great challenges still ahead of us.