It’s been a couple of weeks now since Google announced that its main search engine is now able to search and parse KML files, the native file format for Google Earth. This was widely reported in the blogosphere but with little comment, I’m not sure most mainstream GIS users are even aware of the news..
They should be !!
It might not seem a big deal, after-all KML is a “Google” format, and you would expect it to be searchable in the same way that a pdf document is for example. But.. and its a big but, the Google search engine is parsing and understanding the geographical data within the KML and returning relevant results geographically in additional to all the every clever page rank stuff.
So if I chose to publish the KML file of my evening walk around Teddington on my web-site.., The Google spiders would find it and parse the content noting from the description tags that it is about teddington, but would also get the extents of the GPS track from the linestring co-ordinates.
Now anybody searching for content on Teddington would find the file and its content either from the term teddington, or if using Google Earth from it’s actual location encoded as geographic co-ordinates.
Ok now move beyond a simple walking track to a KML with a linked shapefile, or network link to an enterprise spatial database of agricultural information. The mechanism described would search and find this content just as well !
As Michael Jones points out in the excellent Directions interview, Google Earth and I guess potentially tools that understand KML like ArcGIS Explorer become browsers of Geographic content in the same way Firefox or Safari are browsers of document based content.
What does these mean for the GI industry – I think this is really important !!
To develop infrastructures of Geographic information (SDI’s) we are doing the “right thing” working hard on metadata standards, and discovery portals but it is taking a long time and may need a revolution in semantic techniques to actually work using even quite broad controlled vocabularies of terms.
But hang on… the rest of the web did not wait to develop metadata standards for page content, instead it could be argued they took the “dirty” route and chucked massive computing power and very clever search algorithms to solve the problem with great success – To Google is now mainstream language.
With INSPIRE now real, it’s interesting to think, is the solution to a practical and cheap to implement SDI, publishing KML files and a simple search box ?
Wake up everybody !!!
Written and submitted from home, using my home 802.11 network.