[Aust-NZ] Any comments on this FOSS GIS abstract?

[pcreso at pcreso.com]

Hi,

I'm scribbling out an abstract/summary (something under 1 printed page).... 

I was asked to participate in the panel at the end of a GIS/Fisheries symposium in Brazil last week, my topic: discussing GIS systems.

The following text pretty much follows my presentation.

I figured I'd run it past people here for comment, to help catch any factual errors (I do make the odd one or two occasionally :-) & to see if anyone has any suggestions to improve the wording or content.

Hopefully the text is largely self-explanatory, & you won't disagree much with the content, so despite not having been present, you may still have useful input on the content :-)



Thanks,

  Brent Wood

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Abstract for GIS Systems panel discussion:

Keywords: GIS, RDBMS, web services, OGC, spatial data.

This presentation discusses the nature of GIS & GIS systems, and suggests a broad definition of GIS: "A tool supporting interactions with data which have a spatial component." Such interactions include collecting, managing, querying, analysing, modelling, mapping, reporting, summarising, (etc), spatial datasets. GIS applications therefore may include a wide range of statistical, modelling, visualisation, etc. tools which are not normally regarded as GIS, but are often used to work with spatial data. 

Following this definition of GIS, a GIS system is defined as: "A software suite supporting one or more of these interactions to achieve a particular purpose."

Historically, it was common for spatial data management in a GIS system to be embedded within a  generalised GIS application. However, an ongoing change in the approach taken to spatial data management is the increasing role being played by spatially enabled RDBMS (complying with the OGC SFS standard), with separate GIS applications used for analysing & mapping these data directly from the databases. The whole suite comprising the "GIS system" defined above. Thus the architecture of more recent GIS systems is typically a suite of mapping/analysis/etc applications accessing data from a (often standalone) spatial database.

For example, within the ESRI GIS software suite, spatial and aspatial data were initially managed within a single GIS application. With the introduction of SDE, the spatial data management was carried out within the "GIS", but aspatial data management was carried out in a linked external database. Today, using the "geodatabase" approach, both spatial and aspatial data are increasingly being managed outside the traditional "GIS" application, using an external third party RDBMS.

Accompanying this change, perhaps in some ways driving it, is the use of the internet to access spatial data. This model further separates the functionality of spatial data management from other GIS functionality. Web access to spatial data (and metadata), via OGC web services (eg, CSW, WMS, WFS) is becoming increasingly common. Front end GIS functions 
(mapping, analysis, modelling, etc) are thus becoming (and in many cases, have already become) divorced from the foundation of data management. This approach to spatial data management (and access to these  data) is likely to address, at least in part, concerns regarding access to spatial data which were commonly expressed throughout this symposium. 

These changes are causing standards to become increasingly important to GIS systems, as compliance with effective common standards is the only way disparate applications can access externally managed data and become interoperating components of a working GIS system.

In the second of these symposia (2002) Open Source (OS) GIS software was discussed, but was not widely used, despite a high level of interest being expressed in it. Six years later, at the 2008 event, the majority of presenters had used some form of OS software in the work being presented. Those that used commercial GIS applications such as ArcGIS or Mapinfo, had often also used OS packages such as R:stats to analyse or model the data. Even those that used only ArcGIS v9.2 are now using OS software, as ESRI have included code from the OS GDAL (Geographic Data Abstraction Library) within their product. 

Apart from the obvious cost advantages, this growth in the use of OS GIS software is probably related to the change towards standards based applications described above. Historically, proprietary GIS software tools have had a commercial advantage in locking customers into a custom model, making it more difficult for other vendors' solutions to interoperate with their products. OS solutions, however, are written to perform in a more cooperative, standards based environment, and thus, at least for now, may offer some advantages for GIS users over more proprietary solutions.