It has been a little over 25 years since ESRI released its first product ArcInfo. Since then, it has rolled out a path-breaking suite of software that has made geospatial technology an indispensable part of almost every business and government around the world. Now, it’s ready with its latest version, ArcGIS 9.3, which is being incrementally released in a series of service packs with emphasis on refining and improving workflows. The bulk of the product will be released next year.

ESRI unveiled some of the additions during the recently held user conference. Version 9.3 promises pretty maps that can be printed, map tips, which enable the user to add information directly on a map, enhanced interoperability, Rest API and Java Script API support, caching and performance and security for web applications. It also makes the job of AutoCAD users much easier offering them spatial data management and cartographic capabilities. Further, the ArcGIS Server 9.3 includes a mobile client with a Windows Mobile Platform.

Basically, ArcGIS 9.3 packs include ways to improve geographic science such as Gaussian geostatistical simulation and rich error messaging and improving mapping and labeling, scripting, modeling and WYSIWYG graphic editing. Getting into the specifics, ESRI will provide an ArcGIS JavaScript library that extends the Microsoft Virtual Earth Software Development Kit. The users can use the extended SDK to mashup content from any ArcGIS 9.3 Server via the REST API for ArcGIS Server with Virtual Earth maps and imagery. A live demo was provided by the Virtual Earth team at the conference exhibition. Some believe this is a huge step for ESRI to port ArcGIS 9.3 data into Virtual Earth.

As for geoprocessing, there will be a number of functionality enhancements to the geoprocessing framework in ArcGIS 9.3. There will be new business graphs and charts including matrix scatter plots that reveal data characteristics. Scatter plots show slices of different correlations (disposable income, age, ethnicity, race, etc.). The plots are interactive such that each data theme can be selected to graph on a map. A user can select a chart type, draw a box around a group of points on the plot and have that specific set of data points displayed on the map. Regression analysis will allow users to look at the correlation between data. Users will be able to build a regression model and then apply the model to other geographic areas.

Besides, spatial regression models will allow users to understand how data compare with predicted circumstances or events. If a user determines that certain spatial patterns are not random, there may be some underlying geographic relationships to explore.

The Geographically Weighted Regression Model (GWR) fits a separate regression equation to each data point. It's possible to predict where other phenomena will occur and draw comparisons. Users will be able to develop geoprocessing tools on their own using Python scripts, ESRI's current scripting language. ESRI is seeking partners to add Python-written models to ArcGIS. Other spatial statistical calculations such as mean distributions, quartile ranges and spatial autocorrelation (Global Moran's I) will also be included.

ESRI will publish a new Web site named the Geoprocessing Center to coincide with the ArcGIS 9.3 release in an effort to promote the sharing of geoprocessing models and scripts among the ArcGIS user community. This new resource center is part of the ESRI support site.
Last but not the least, ESRI will be supporting Microsoft’s spatial yet to be developed spatial type in SQL Server. ArcGIS 9.3 will support this capability.