The announcements out of Where 2.0 from John Hanke of Google and Jack Dangermond regarding integrating neogeography with professional GIS (perhaps not the best terms, but I'm confident readers understand) are quite a lot to digest. (Video available
here.) But that's ok, both companies are resetting their visions with regard to the other, to data and to services and it's certainly time for that.
Here's the substance of the relevant announcements teased out of coverage from Where 2.0, where the two geotechnologists shared the stage yesterday.
- ArcGIS Server 9.3 (available in about 4 weeks, per Dangermond) will make its metadata service "scrapable" into KML and thus findable via Google's geographic search (once known as KML search). Further, ArcGIS Server will be able to publish not only that data as streaming KML (and GeoRSS) but also related services. Dangermond showed finding data from a Portland, Oregon service, visualizing it and then performing analysis, all from Google Earth. Said another way, all data and services served by ArcGIS Server could potentially be findable and usable in any Google mashup. Further, the resultant KML can be used in app that supports the OGC standard.
- ESRI has enhanced the API for ArcGIS Server 9.3 (JavaScript/Flash) to make it more conducive to plugging into other Web mapping properties in mashups.
- Google is making its geographic search available in its various APIs. To date it was only available via Google Maps and Google Earth applications. Now any Google developers will be able to do "local search" on explicitly tagged data (KML built via MyMaps or 3rd party apps like Platial and Flckr or your GIS!).
This is a huge step forward for geography (neo, paleo, and all the rest). It does indeed bring the hidden data and emerging Web services from the huge ESRI community out into the light of day. I'm very excited as
I tried to paint a picture something like this when KML search was announced last year.
These announcements have some important implications for the geospatial marketplace:
1) the use of and demand for ArcGIS Server should rise
2) any geo data or services provider who wants to play on the Web needs to look at how it will provide findability and usability of its data and services in this way
3) geodata-finding portals may, in time, become extinct - if the KML vision for search and distribution becomes a de facto one as well
4) the technology may be available, but the institutional barriers to data sharing may still be blocking the path (as noted in a question at the end of the session)
Finally, it's worth noting that a parallel project is underway in the open source community. The Open Planning Project (the folks behind the open source GeoServer) are showing off a
new version that also provides KML findability. Part of OPP called OpenGeo (not to be confused with the Open Geospatial Consortium or Open Source Geospatial Foundation) was showing it off at Where. That means more choices such that in time even more players will have the ability to serve up findable, usable data to be mashed, and make the world a better place.
Google geo-search does currently support search on Atom/RSS elements coded as georss:point - they are planning support for georss:polygon and georss:box within the year, allowing for content to be more readily found and placed based on its extent. It seems to me that the indexing of feeds to provide links to both fuller metadata and data access URLs (i.e. WFS, WMS, or KML) would be quite effective.
It'll take a brave GIS manager at the City or County level to open this new door to the taxpayer's database.
A) Google Earth/Maps users are not likely to even know what metadata is, let alone look at it.
B) A lot of gov't data has serious errors, which GIS-users on the whole uderstand; GE users may not be so forgiving.
C) There were VERY few local and state gov't attendees at Where 2.0 (compared to the more forward-thinking mobile app providers, marketing/graphics gurus, open source evangelists, and academia), which leads me to think that we may be a ways off before the local/state/fed data providers to share data in this format.
http://www.itopen.it/soluzioni/kml-map-server/
Adena,
As you alluded to at point 4, the institutional barriers will be the highest hurdle.
As discussed recently on the OSGeo Discuss list, most countries don't have the liberal data access arrangements that the US and Canada do. Many organisations have their budgets tied to sales of licenses to their data. There are also issues of security, and concerns of legal liabilty relating to the use (or misuse) of spatial data. These issues are not going to be resolved in the short term.
By the way, just because an organisation is using ESRI software for data maintenance, does not mean that they will use it for the maintenance of their spatial metadata or commit to ArcGIS Server for their web mapping, services etc. There are a number of other good alternatives out there, many of them are open source.
In addition, I doubt that this development will replace the need for good spatial metadata catalogues (e.g. OGC CSW). I know that when I'm looking at undertaking some spatial analysis, I want to have a good understanding of the quality and lineage of the data that I will use as it will have a significant impact on the quality of the results that I will end up with.
I think that this last point that is often missed in the KML for search debate.
Bruce