There’s still hope for Australia’s spatial industry

The ASIBA last week release a set of recommendations for the Australian government. These recommendations were based on a economic study evaluating what the spatial industry in Australia is worth. A staggering 6-12 billion dollars is the figure mentioned.

But lets look at the recommendations: I quite like the aims of some of these: government initiatives to commercialise OZ spatial information R&D; government policy that improves the quality, quantity, currency and accuracy of spatial data; management of digital rights and information sharing and a whole-of-government approach to licensing of geodata.

There are a number of initiatives already in the pipe-line to address some of these issues for example to address constraints on information sharing, OSDM is developing a whole-of-government approach for licensing of spatial data. There is also a push to use existing standards for discovery of spatial data where applicable such as OGC‘s CSW whose most well known implementation is FAO’s GeoNetwork. There are also projects that look to make taxpayer funding research data free. The question now is how long will it be before we start seeing the fruits of all the good intentions from the government and traditional Spatial industry. How long is a piece of string?

Meanwhile neogeography projects such as OSM continue to flourish and so do the related standards such as GeoRSS. Another recommendation that will hopefully see greater support for such open and community driven initiatives in Australia is that

… priorities are user driven, not ‘producer’ driven

Hopefully this means that the OZ government’s inacceassable silo’s of geodata that’s currently begging to be freed may one day become available in formats that ease their use online and amongst the developer community. This would no doubt lead to the emergence of new industries, a responsibility that governments have towards their nation.

Mapserver output as KML

Alessandro Pasotti has authored a utility to output mapserver layers as KML for display on Google Earth. This is written in PHP mapscript so will require mapscript to be installed with your standard Mapserver installation. Check it out here.

Although its a useful tool, it would be nice to have kml support in mapserver (C++) out of the box. Apparently expected is the release of libkml in 2008. So mapserver may have to wait for google to release libkml before adding native C++ support, provided the licensing is favorable.

Geospatial @ RailsConf Europe 2007

railsconf_speaker Some great news for those interested in the convergence between Ruby/Rails and Geospatial applications. Kashif and I will be giving a tutorial at RailsConf in Berlin. There is going to be 3 hr tutorial titled Rails GIS Hacks, where we will be going through some pretty nifty stuff such as RESTful geospatial applications development. Hope to see you there!

Pakistan Wetlands and Ramsar listed sites

Ramsar Sites in Pakistan Page
Wetlands are regarded by nature conservationist as unique ecosystems. They often exhibit delicately balanced environmental conditions and support a wide range of exotic plant and animal species. Wetlands are also often protected by various environmental laws and conventions. Last year when I was visiting Pakistan I learned about the Pakistan Wetlands Project that is under way. The aim of the project is to build capacity for improved wetlands management at the Federal, Provincial and Local government level, as well as the private sector in Pakistan. I thought it would be nice to be able to see all 19 wetlands in Pakistan that are listed under the Ramsar convention on GoogleMaps. This site was built using RubyonRails + GeoRuby. http://pakistan-wetlands.dyndns.org/.  Many thanks to Guilhem for his help during trouble shooting.

Trying out Geoserver + Googlemaps

karachi_small
With a dearth of street data for Pakistan on the default mapping resource of the world (Googlemaps) I decided to try putting up some of the data I had lying around the house. Using Geoserver as a WMS layer on top of Google maps results were quite impressive. Here's Karachi and the tomb of the founder of Pakistan (M. A. Jinnah).

One can get some impressive results using the SLD support in Geoserver WMS. The site is not live yet but it will be soon. I am just trying to come up with something a bit more useful than pretty roads.

Open Source Developer’s Conference 2006, Melbourne, Australia

OSDC 2006 OSDC 2006 starts tomorrow in Melbourne with a keynote from Randal Schwarts. Today I attended the pre-conference tutes which included a most usefull session by Jonathan Oxer. The talk titled "Large Scale Webapps: DevTeam Infrastructure" gave a run down on how software development process at his company is automated and streamlined. The automation process includes documentation, automated migration from development to testing, to packaging and releasing. You can find a more detail here. Earlier today I gave a tute on GIS Python Hacks based roughly around the workshop by the same name from Howard Butler and Shean Gillies at OSGIS'05. I'll put my slides online by tomorrow. The topics covered GDAL/OGR, GEOS and WFS with hands on snippets of code. We also did some promotion for the newly formed Open Source Geospatial Foundation.

UPDATE: Download the tutes: 
WFS_tute.zip, OGR_tute.zip, GDAL_tute.zip

PostGIS & Oracle Spatial Price Comparison

Ken Lord emailing to the annual “Who’s using PostGIS and for What?” inquiry wrote in detail about his experienes with Oracle and also a justification for decision to chose PostGIS over Oracle Spatial. I thought it was quite insightfull for anyone who are doing a cost/benefit analysis of choosing a spatial database for their next commercial project. To summarize it costs CAN$60,000 to use Oracle Spatial per CPU for web applications. Support & Maintainance is an additional CAN$13,000.
This is because inorder to use Oracle Spatial you need to buy Oracle Enterprise. Plus if you are going to use it for Web Applications then you will have to pay the per CPU license instead of the per user license.

PostGIS on the other hand seamlessly works with Mapserver (webifying your Spatial DB) and has most if not all of the functions offered by Oracle Spatial. Not to mention the supportive community who contribute by submitting bug fixes and answering newbie to advance questions. The documentation for PostGIS is also quite extensive.

There have been a few interesting comments to the annual inquiry from the developers of PostGIS. One of my favourites was: “We can have a new feature added to PostGIS for the license fees of a commercial solution

This is not to say that one would never consider Oracle Spatial for a project, since Oracle Enterprise with its myriad of features maybe able to fullfill the requirements of a project better than PostgreSQL/PostGIS. However, as far as Spatial Databases are concerned PostGIS does make economic sense.