Friday, May 18, 2007

Quick review of GeoTec

I spent a couple of days this week at the GeoTec conference in Calgary, and it was a good event. Matt Ball said they had 1200 people registered - I don't think there were quite that many in the opening session, but it was good crowd still. Ed Parsons gave a good keynote on "once there were maps, now there is neogeography", focused on the impact of Google, Microsoft and Yahoo on the industry. I'll just mention a couple of snippets. He mentioned the momentum of open source data in Europe, something we haven't seen as much of in North America - Open Street Map is one of the main initiatives in this area. He also mentioned how geospatial standards had evolved from “proprietary standards” like DWG, DGN, Shape and Geodatabase, to “industry standards” like SOAP/UDDI, OGC WxS, and ISO 191xxx, to “lightweight standards” like GeoRSS, KML and REST. There's quite a lot of debate flying around at the moment about relative merits of KML and GML - I won't get into that at the moment, but may come back to it in a later post!

He also talked about Google MyMaps and highlighted that these maps are searchable via Google just like any other data. While he didn’t mention this specifically, any KML file online is now indexed by Google and can be searched for using both keywords and geospatial constraints. Later in the opening panel which followed the keynote, in which I participated, Sam Bacharach of OGC asked whether the panel thought that the more unstructured search approach of Google would be how people found spatial data in future, versus more structured approaches like those which had been worked on by the traditional geospatial community for the past ten years. There was some discussion around this, but overall I think the general opinion was probably yes, in most cases. There was also some discussion both on the panel and afterwards about concerns around metadata (in the OGC/ISO sense rather than the more generic KML sense), and being able to determine the quality of geospatial data. I think that as Google works with OGC on trying to harmonize KML with GML, it would be logical to allow the optional addition of OGC/ISO metadata into KML files, and then Google searches could use metadata when available to help determine where a given file should rank in response to a given search. In general there is a battle going on between more formalized approaches to being able to search for “structured” data, which generally fall under the umbrella of the “semantic web” initiatives advocated by Tim Berners Lee and others, and less structured approaches such as tagging (as used by flickr and many other “Web 2.0” sites), which are less theoretically elegant but much easier to get going in practice, and which may provide “good enough” results in many cases.

Another strong theme at the conference was open source, which really seems to have strong momentum in Canada. When I was at Intergraph, I got many more inquiries from colleagues in Canada about the impact of open source geospatial solutions than from anywhere else. I enjoyed spending some time with fellow blogger Jason Birch, and attended a very interesting presentation he gave with Paul Spencer of DM Solutions about the work that City of Nanaimo is doing using open source Internet mapping solutions (there is a lot of new work in progress which isn't on their public site yet). I am seriously considering going to the FOSS4G conference, which is in Victoria in September - it sounds as though it should be pretty interesting (plus Vancouver Island is always a nice spot to visit!).

I also attended a couple of interesting presentations on the use of Google Earth in more "traditional" geospatial applications, just reiterating the trend that we are seeing towards this. All in all it was a good conference.