ESRI\'s Jack Dangermond gave this interview to our UK edition editor, Maria Pellegrini, while she was at the ESRI User Conference 2008 in San Diego, 13-17 July this year.
GeoConnexion: What is ESRI doing to support multidimensional GIS?JD: I have several responses to this. First, if by âmultidimensional,â you mean â3D,â then with ArcGIS 9.3, 3D features are fully supported, and it is possible to edit and visualize them. However, within version 9.3 the analytics support is only available to developers via an API. We plan in version 9.4 to develop the user interface to enable 3D analytics to be available to users, which will be the equivalent to current 2D geometric analysis, but in 3D.If the extra dimension you are interested in is time, at our earlier 9.2 release, we added time stamping on features so that a transaction or an update of the feature also updates the time stamp. Basically, this involves storing feature histories and allowing time-based query and visualization using GIS-based tools. This history can be accessed in our query tools to allow the reconstruction of maps for the past and the analysis of any changes over a selected period of time.Our time support includes animation; we have spent a lot of time on 2D and 3D animation, and our users are now creating animations both through past time and simulations of possible futures. However, time is more complex to deal with and not well defined in GIS science.However, to support the scientific communityâs multidimensional work, we do support netCDF files as data sources (read and write) in ArcGIS through our geoprocessing environment. NetCDF is a powerful format for sharing data sources that have a temporal dimension; for example, when helper applications cannot interpret temporal storage in a DBMS, as ArcGIS does, you can export to netCDF.We are now working with PES, a Dutch company we acquired that does dynamic simulation for modelling the past, present, and future. Simulating the future requires a model of the present to project forward. This is done by combining various overlays that describe or model future states of geographic phenomena. To support these modelling efforts, we continue adding analytic tools, such as spatially weighted regression, to help users better develop formulas for simulating the future.At version 9.4, we will be incorporating some of those dynamic feedback modelling tools into the core functionality. In 9.2, ModelBuilder could be used to create time series iterations. Using spatial analysis tools, you could then develop simulations. So between these dynamic feedback modelling and spatial analysis tools, we will have predictive modelling facilities in 9.4.One other thing to mention is developments within military applications, which they call Visualization Simulation Systems. The military currently uses condensed, specialized 3D data models for visualization. The GIS models operate more slowly because they are general purpose and not so highly optimized for visualization. Gradually, this is changing as the increasing speed of hardware compensates for the less optimized structures, and it will allow GIS to do the visualization. Also, GIS can do geospatial query and analytics as part of the visualization system, which enhances its capabilities.In conclusion, as the science develops in multiple dimensional databases, we are engineering the tools that will support them inside GIS.GeoConnexion: What is ESRIâs attitude toward âopen sourceâ and its views on future relationships with it?JD: This is a popular question, and no doubt of interest to your readership. Generally speaking, we very much respect the open-source community. Clearly, closed-source and open-source solutions, licensing, and business models can coexist in the same software ecosystem. ESRI demonstrates that every day with our open system strategy. For example, just to name a few projects, we have engineered our technology to run on Linux, we support the enterprise geodatabase on PostgreSQL and PostGIS, we use the Apache open-source Web server, we embed GDAL [the open source raster library], and we have made part of our spatial analysis to be open source by having it written in Python [the scripting language]. As such, we encourage our customers to take advantage of open-source solutions that enhance and complement the developer and user experience with ESRI software. We live in an open-source âsandwich.âCurrently, ESRI is a member of the OpenAjax Alliance and the 52˚North open initiative for spatial data infrastructures [SDI]. As for future developments, ESRI will continue to participate in and sponsor open-source geospatial conferences such as FOSS4G, the Free and Open Source Software for Geospatial Conference. Obviously, we are a closed-source company, but that does not exclude our interest and involvement in the open-source movement, which is broad and diverse, in order to engage in dialog with open-source geospatial developers and users, eliciting feedback that could eventually be reflected in future ESRI products and business directions for the benefit of all.Fundamentally, our users depend on us to be a strong business and provide ongoing and consistent product evolution and strong technical support.
Author: Maria Pellegrini
Bio.: Editor, GEO:connexion UK
For more information visit:
Visit ESRI's UC 2008 web site for more information