Requires Open Standards-Based GIS Software
Sharing Geographic Knowledge: Providing Open Access to Geographic Information
At Esri we have a vision of building technology that facilitates open sharing of geographic knowledge freely and easily so that the power of thinking geographically can be brought to bear on many of the world's pressing problems. For this vision to become a reality, it is essential that the billions of dollars of geographic data, geographic processing capability, and user expertise be easily available.
Esri has engineered its software products to support this vision. One of the main ways to facilitate this is by creating and implementing standard ways to store, access, and process geographic information. Our goal has been to help users collaborate by facilitating the free exchange of data and other GIS resources and the interoperability of GIS components.
The purpose of this article is to outline the nature of this work and give a report on the current status. The key requirements for open access to geographic resources are open platforms, open DBMS, open data, and open application programming interfaces (APIs).
A GIS platform is the hardware and operating system on which software runs and data is managed. Esri is unique among GIS software vendors in terms of the range and number of platforms we support. In the mid-1980s Esri committed to provide support for a wide range of operating system platforms. Today, Esri software runs on Digital Tru64 UNIX and Alpha Windows NT, IBM AIX, HP-UX, Microsoft Windows 95/98/NT/2000, NEC EWS-UX, SGI IRIX, and Sun Solaris. We are currently evaluating and testing Linux as a viable platform. Esri data sets are binary compatible across all these platforms, which means, for example, that coverages created on a Solaris platform can be used without translation on Windows NT.
This allows users to choose the hardware platform(s) most appropriate to them and to benefit from the price and performance advances of the different vendors. Historically, one of the major reasons for the success of our software has been its ability to run on nine different software platforms (including Macintosh and several kinds of UNIX and Windows). As users and the industry have focused on Windows for desktop tasks and Windows and UNIX for server functions, so Esri has followed.
As user databases have become larger, with more concurrent users, it has been a natural transition to use database management system technology to store geographic data. In the early 1980s Esri was the first GIS company to build a commercially viable product that accessed data stored in a DBMS. Initially this was attribute data only, but since then this has encompassed geographic data as well. Today, the combined geodatabase and ArcSDE technology in ArcInfo 8 is a world-class data management solution able to manage data stored in several different database management systems on multiple hardware platforms. For example, ArcInfo and ArcSDE are able to work with data in Informix, Oracle, Sybase, and SQL Server. Esri uses open, industry-standard APIs to communicate with the DBMS (see below for more details).
Open access to data in databases allows users to take advantage of DBMS technology to store and manage data, to support multiple users and applications concurrently on the same database, and to integrate heterogeneous data at the desktop. Using DBMS to store and manage data provides a superior solution for backup/recovery, replication, failover remote synchronization, and multiuser access.
We are often asked the question, Why are there so many geographic data formats, and why is it so difficult to use them? There are so many formats because no single format is appropriate for all tasks and applications. It is not possible to design a format that supports, for example, both fast rendering in command and control systems and sophisticated topological analysis in natural resource information systems. The two are mutually incompatible. Thus many different formats have evolved in response to diverse user requirements. A na´ve view of providing better access to geographic data is to ask that formats be published. While direct access to simple formats is a good thing, it is not recommended for advanced formats where applications are required to manage multiuser access, maintain data integrity (e.g., topological consistency), and support data encapsulation so that storage structures and indexes, for example, can evolve over time. This is why Esri has three main native data formats that have been adopted as de facto industry standards:
Given the high cost of creating databases, many people have asked for tools to move data between systems and to reuse data through open interfaces (APIs). In the former case the approach has been to develop software that is able to translate data, either by a direct read into memory or via an intermediate file format (there are advantages and disadvantages to both). In the latter case software developers have created open interfaces to allow access to data (see below for more details).
Esri pioneered the development of "direct read" technology with the introduction of ArcView GIS 2.1 when the ability to read the relatively simple AutoCAD DWG and DXF formats was added. Subsequently, direct read support has been added for several other formats in ArcInfo, ArcView GIS, and MapObjects. The list of supported formats includes AutoCAD DWG/DXF, MicroStation DGN/MGE, VPF, and many image formats. Unfortunately, direct read support can only be easily provided for relatively simple product-oriented formats. More complex formats such as SDTS and the UK NTF formats were designed as translation formats and require more advanced processing before they can be viewed (e.g., multipass read and feature assembly from several parts).
Esri software is able to translate more data formats than any other GIS software. The current list of supported formats in ArcInfo 8 is shown in the table below.
One way to support open access to data is to enable read/write capabilities via a programmable interface. Access via a clearly defined interface is at the heart of object-oriented and next generation object-component systems. Esri has long supported open APIs to our software and data formats. For example, the GRID/ArcView Spatial Analyst extensions to ArcInfo/ArcView GIS include a GRID I/O library that allows a programmer to relatively easily read and write the GRID data structure. Similarly, for many years ArcSDE has included a rich set of industry-standard APIs. First came the C API, then the CAD Client interface; and more recently Java, SQL, and COM interfaces. In addition, ArcInfo/ArcSDE 8 ships with an OLE DB provider (interface) for geodatabases and tabular DBMS access.
Historically, the GIS industry has been active in creating standards for data interchange. More recently, there have been attempts to begin to standardize other aspects such as geoprocessing, metadata, and projections (see, for example, the Info 2000 and OGC Web sites). This is an interesting but ultimately limiting direction. If GIS is to become part of the IT mainstream, then we must embrace IT standards and extend them into the geographic domain, not create our own industry-centric standards.
The four key parts of a modern desktop/enterprise GIS are Windows for user interface, Component Object Model (COM) for tools, Structured Query Language (SQL) for data access, and Transmission Control Protocol/Internet Protocol (TCP/IP) and Hypertext Transmission Protocol (HTTP) for network data transfer.
There are other important standards that also need to be extended in the geographic domain to support open access to geographic data and processing:
The Role of Standards Committees
There are at least 26 organizations involved in the standardization of various aspects of geographic data and geoprocessing (for a list see a site such as Info 2000. Several of these are country and domain specific. Esri tries to be active in the main groups that have the most significance to our users. In particular, Esri staff have been active in Open GIS Consortium (OGC) technical committees.
Esri was the first GIS vendor to become certified as having GIS-compliant software (SDE 3.0.2 SQL interface) (see www.esri.com/opengis). We were also lead authors for the Simple Features SQL standard submission and participants for the COM and CORBA submissions. More recently, we have been actively engaged in the efforts to develop Web mapping standards through the Web Mapping Test Bed (WMT) initiative.
Esri extended ArcIMS to support the OGC promulgated standard. In an attempt to further support OGC as it moves more into the marketing domain, Esri has recently increased our level of membership to the Principal Member level (the highest tier).
Esri also participates in the ISO SQL/MM (Multimedia [Spatial]) Committee, which is attempting to extend SQL into the spatial domain, and the other key ISO technical committees involved in geographic standards, TC 287 and TC 211. There are many other examples of geographic standards committees that Esri is working on behind the scenes. We see our role as supporting the good work of the many active community participants.
This article outlines the importance of standards for GIS, especially those that are built on the firm foundation of generally accepted IT standards. Esri's goal is to build open GIS technology, which really allows users to share geographic data and processing resources. Esri already has software interfaces and data formats that allow users to open up their data and systems to others if they wish. While we are conscious of the considerable progress we have made, we are also aware that more can and will be done to support our user community.