I was privileged to spend a few days at FOSS4G in Denver. It’s quite the opportunity to have an international convergence of GIS developers handily just a bus ride up the highway. I could walk among the current luminaries of OsGeo, Arnulf Christl, Paul Ramsey, Frank Warmerdam, Simone Giannecchini, and so on and on, and trade a handshake with the hard working front range folks, Peter Batty, Brian Timoney (way to go Brian!), and Matt Kruzemark too. The ideas are percolating in every corner, but it strikes me in an odd way that this world has changed irrevocably. The very technology that makes Geo FOSS possible is also making it, of all things, mundane.
It wasn’t that long ago that OGC standards were brand new. Each release anticipated a new implementation for discovery. Now implementations proliferate helter skelter. OGC services? Take your pick. Or, slice, dice, and merge with MapProxy. Map Tiling engines line up like yogurt brands at Whole Foods. GeoFOSS has progressed a long way from naive amazement over OGC WMS connections. WFS and WCS have already passed their respective moments of glorious novelty.
This is the year of WPS, Web Processing Service, and the hope of constructing webs of analysis chains from disparate nodes. Each node adds a piece toward a final solution. Anyone with some experience using FME Transformers has some idea of what this looks like on the desktop. WPS moves analytic workflows into open standards and onto the web. Ref: 52north.org and GeoServer Ext
Fig2 – example of FME Workbench processing chain
On the floor at FOSS4G 2011
In a different vein, the next release of PostGIS will push raster analysis right into SQL. This is new territory for PostGIS and elevates a raster datatype to the same level as geometry. This PostGIS Raster Functions page gives some flavor for the raster functions soon to be available when 2.0 releases. Ever needed a raster reprojection?
raster ST_Transform(raster rast, integer srid, double precision scalex, double precision scaley, text algorithm=NearestNeighbor, double precision maxerr=0.125);
Algorithm options are: ‘NearestNeighbor’, ‘Bilinear’, ‘Cubic’, ‘CubicSpline’, and ‘Lanczos.’
(See even the slightly jaded can catch some excitement at FOSS4G.)
CouchDB/GeoCouch was present although “eventually consistent” NoSQL seems less pressing.
With over 900 attendees, the small world of FOSS seems all of a sudden crowded, jostling through accelerating growth. The Ordnance Survey was represented at a plenary session, “We have to learn to embrace open source. We want to use OS. We want to encourage OS.” Who’d of ever thought? FCC.gov opens their kimono in a big way with data transparency and OS APIs. Work shop topics include such things as: “Open Source Geospatial Software Powering Department of Defense Installation and Environment Business Systems,” or catch this, “The National Geospatial-Intelligence Agency OSS Challenge.” Is this the new Face of FOSS?
Such bureaucracies know little of agile. The memorable phrase “embrace, extend, extinguish” comes to mind. But who embraces whom? Is the idea of FOSS and its larger www parent, a trend that eventually extinguishes bureaucracy or does the ancient ground of bureaucratic organization trump connection? Byzantine bureaucracy has been around since … well the Byzantine Empire, and given human nature is likely enduring. But, what does a “flat” Byzantine Bureaucracy look like? How about crowd sourced Byzantia? Would an aging Modernity mourn the loss of “Kafkaesque” as an adjective?
Assuming growth continues, will success reduce the camaraderie of community as a motivation? Just this year we hear of OSM’s Steve Coast sidling into Microsoft, followed a little later by GDAL’s Frank Warmerdam beaming up into the Google mother ship. The corporate penalty, of course, is the loss of personal intellectual property. In other words, will Steve Coast’s imaginative ideas now fall under the rubric, “all your base are belong to us,” and Frank’s enduring legacy recede into the patent portfolio of “Do no evil?” FOSS4G halls are still filled with independent consultants and academics, but a significant corporate representation is evident by this apparent oxymoron, a presentation by ESRI, “Open Source GIS Solutions.”
The GIS nervous system grows connections around the world faster than a three year old’s brain. OK, maybe I exaggerate, after all, “During the first three years, your child’s brain establishes about 1,000 trillion nerve connections!” Really, how many connections are there in the World Wide Web. For all its impact on life, our beloved internet encompasses the equivalent of what, a few cubic centimeters of a single infant’s anatomy.
Open Geospatial and FOSS are just a small part of this minor universe, but it’s easy to forget that even ten years ago this all hardly existed. “The first Interoperability Program testbed (Web Mapping Testbed) appeared in 1999.” About the same time Frank Warmerdam started GDAL and the GFOSS engines started.
I still recall the wonder of downloading Sol Katz utilities from BLM’s ftp. The novelty of all this data free to use from the USGS was still fresh and amazing. Sol sadly died before seeing his legacy, but what a legacy. The Sol Katz award this year went to Java Topology Suite’s well deserving Martin Davis.