Maybe if we change the semantics and talk of maps as views (in SQL terms) instead of models. Whenever anyone looks at a map they don't want to see everything everywhere - even if they do the brain can't comprehend that much.
So then the problem of delivering custom maps becomes one of creating custom views filtered on extent and content. If we start with masses of "raw" data, each element tagged with the entity it represents then an aggregation becomes a view resulting from a generalisation algorithm applied to the "raw" data in a similar manner to that in which raw words garnered by a web crawler becomes a set of search terms. If we had a metamodel and the appropriate ontologies for "tagging" the raw data then we would have the basics for building a semantic geoweb - or am I being too simplistic? Cheers AlanK -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Thursday, 3 July 2008 10:41 AM To: [email protected] Subject: Re: [Geowanking] Critical Theory Interesting points Alan and Landon, and I agree we have a long way to go handle the masses of geodata that exist today and the additional masses that will be available tomorrow (when Landon's next gen GPS comes to pass). Maps are inherently models or abstractions of reality (despite our acceleration towards mirror worlds). In many ways this gets back to Renee's MAUP problem - one of the easiest way to abstract data is to aggregate it to larger units (census tracts, zip codes, counties etc.). While the promise of always working at the raw data level is tantalizing we are still a long ways off of that on the GeoWeb. If we look at the limits of KML file size support in Virtual Earth and Google Maps today it is roughly around 2mbs. There are limits of what you can handle in memory and limits on how much data you can render on the map. So, while you could conceivably have massive spatial databases of information your ability to serve up that data and make it consumable by the public is still severely limited. Just think about how hard it is to display an average size GIS data file in a GeoWeb application, especially a browser based application. The GeoWeb has done a great job dealing with points of interest and segmenting them into manageable chunks, but I agree with Landon that going the next step is going to take an incredible amount of work. Although it is definitely work worth doing. I believe the upside will dwarf what we've done with mostly local point data to date. FortiusOne Inc, 2200 Wilson Blvd. suite 307 Arlington, VA 22201 cell - 202-321-3914 _______________________________________________ Geowanking mailing list [email protected] http://lists.burri.to/mailman/listinfo/geowanking
