Geography
Publication
In 2007 we're seeing the first web tools for geo-referencing content. All the major search engines have or will soon have applications that let you create maps and export them to GeoRSS or KML. These maps will reference other documents, and thereby geo-reference them. By 2017, almost all new data or documents with a Earth context published on the web will have geo metadata. Some of that metadata will originate from surveyors, some from GIS analysts, some from GPS enthusiasts, some from "consumer" grade web maps. Some will come from geo-coding the text. Some publishers will have geo metadata standards, many will have none other than "created using google maps 2016-10-24-10:00:00", or "Canon PowerShot? GPS-3000 2015-06-21-11:15:33".
The data will be primarily published using KML or GeoRSS, but there will also be a bestiary of other formats: RDF, lat/long tags, microformats.
Discovery
The google bot began crawling and indexing KML in late 2006. By 2017, this will be routine, and practically all KML or GeoRSS resources on the Web will be found and indexed. Almost anyone will be able to perform simple spatial queries starting first with bounding boxes, then proximity to points, then proximity to other features. Competition between search engines will open up more free search tools, while they sell industrial-strength GIS access to their indexes. The search companies will also geocode resources that are not explicitly geo-referenced. In 2007, this is basically limited to web resources with street addresses or well-known place names, but by 2017 there may well be plausible locations inferred for just about any web resource.
Like today, users must search many databases and repositories. There will be no one-stop sites. The profusion of sources will be mitigated somewhat by better watchlists, filters, and profiles. Users will poll these agents using Atom or RSS ("pull" tech has won in 2017, at least for this stuff). Social tagging will exist alongside formal metadata and hypermedia links, forming a semantic triumvirate.
