Home Articles Geo-Web,GML Present and Future

Geo-Web,GML Present and Future

Ron Lake
Chairman & CEO,
Galdos Systems Inc.

What is the status of Geo-Web since its development, what are the plans for the future?

The GeoWeb is not one thing but more like a layer of the Internet which deals with the sharing/exchange of geospatial information in real time or near real time. We anticipate that the GeoWeb will evolve naturally as open standards like GML, WFS, geoRSS etc. are more and more widely adopted, and as people demand more current and accurate geospatial information regardless of where it might be sourced.

How will the Geo- Web influence the Spatial Data Infrastructures? With most of the countries already having SDI, will they be required to rework their SDI?

I think the growth of local SDIs will help move the GeoWeb forward. One could view the global integration of such SDIs, especially at the local, province/state level as crucial building blocks in the development of the GeoWeb. Of course one must see SDI as meaning the support for real time or near real time transactional sharing of geospatial information.

Could you throw some light on Universal SDI and how far there is a possibility in future for a single SDI to serve all the segments?

I don’t see there being much opportunity for a single SDI, meaning a single system to serve everyone’s needs, even in a restricted setting. SDI is starting to emerge as a common architectural framework based on open standards from the OGC/ISO /OASIS/W3C. The key OGC standards are of course GML (Geography Markup Language), WFS, WMS (FPS in particular), and the Web Registry Service. The latter is key to managing the rich collection of artifacts that arise in the context of SDI. Other standards like geoRSS (W3C), SAML/XACML (OASIS), ebRIM (OASIS) are also key. These specifications and the associated technology components provide the foundation for a universal SDI architecture that can be deployed in virtually all economic sectors.

What are the initiatives being undertaken by various Government and Private agencies for moving under common Geo-Web frame work?

Many governments are taking active positions to foster the development of the GeoWeb. These include the European Union with its INSPIRE initiative, security projects in the United States, Canada’s CGDI, and more recently many active projects in India. The key thing here is for governments to realize that the action is at the local government level and not at the national level. The role of the national government is more to institutionalize open standards. The real business cases are more at the local level and these will then bubble up to the national level. An SDI driven only by national data and national business cases is likely to fail.

Products like Google Earth, having a similar model as GML have revolutionized the market in the recent years. Is there any learning for the Geo-Web from Google Earth?

Yes, KML and GML share identical geometry representations KML is a copy more or less of GMLv2.0. Many people argued GML was too complicated for average developers to adopt. Google Earth has shown this is NOT the case. I think Google Earth has also taught us some lessons about how to teach XML languages to a broad market – e.g. start with instances and examples – avoid schemas except for the expert. Of course Google Earth is a part of the GeoWeb. Right now, Google Earth gets its data by buying it in bulk, which translates into big delays. Suppose 15 minutes after a road changed in a local municipality the change was available on Google Maps! That is the power of the GeoWeb.

Do you think GML which came up with a notion “without the need to purchase client-side software” has overlooked at this aspect of Unanimous software for GML?

It is hard to know what universal software for GML means. There are many kinds of consumers or clients for GML software. Styling engines are clients that create visualizations in things like SVG or of course KML. The Galdos WFS is both a client and a server for GML data. I think it is a matter of getting past the view from GIS that geographic information is something to be shipped around in files and looked at in a viewer. You can do that in GML, of course, but real emphasis is on moving data from one server to another.

What are your views on the usage and popularity of GML and KML?

I don’t think there is any question that KML is more popular. Of course these serve very different needs. KML is more for the hobby practioner and GML more for the professional data manager. GML and KML can co-exist of course, KML makes a nice visualization language for GML, a geographic complement to SVG. In some was KML is more a competitor to SVG than to GML. Of course one can write KML in GML and not the other way around – maybe a Google GML is in the works?

The integration of data under a common reference has always been a challenge. As there is a huge list of coordinate systems to be handled, how well this issue has been considered in the design of GML. And what are the challenges faced?

Yes, I agree this is a huge challenge and not one that is well understood. Even well known GIS software programs often contain mistakes in this area. In GML we have tried to do two things with respect to coordinate systems. First, is to make the coordinate reference system used by a particular geometry element unambiguous. This made possible by referencing the actual definition used. Second, we can express virtually every possible coordinate system in GML, thus enabling the above reference to point at a machine readable definition.

This lays the ground work for automated services that can convert from one coordinate reference system to another. Not the complete solution by any means, but GML provides some of the key building blocks. Galdos is process of deploying a world coordinate reference system through the OGP (Oil and Gas Producers Association). This will likely supplant the existing EPSG Database as one of the global machine readable sources for coordinate reference system definitions. OGP as formally approached the ISO to use this registry as the basis of a world register for CRS. A good starting point.

When the platforms from Google, Yahoo and Microsoft providing environments to incorporate maps and images in websites, how is the concept of Geo-Web going to create a place for itself?

I view these as a part of the GeoWeb. Of course they could become even more proprietary – but I don’t think that will happen. They need data and they current data. Satellite and aerial imagery are all well and good, but you can’t readily detect changes like one new bridge or a house addition, or buried pipes or the boundaries of a land parcel. Data needs to come from where it is created and do so quickly. GeoWeb technologies will help that happen.

With reference to your statement in your Blog, “Google Earth is an impressive achievement”, why do you think so?

There were a number of attempts at making things like Google Earth before (e.g. TerraServer) and none really caught the public imagination. What Keyhole did was realize a few key points that the others missed, namely: 1)You need global coverage 2)You need some way to add your own information locally 3) The visual experience needs to be smooth and exploratory. Google Earth does all of these. It is much a social achievement as it is a technical one.

What are the image formats the current GML version supports? What will be the time frame when GML will evolve to fully support the image formats that the GIS industry uses?

To anwer this question one has to understand how GML approaches imagery. Images in GML are coverages, and GML simply provides markup to describe the geometry (e.g. grid dimensions, post spacing etc.) and radiometry (what the pixel values mean) of these coverages. It then can point at an existing file that contains the pixel values – a sort of extended World File if you like, but one which is much more powerful. In this sense GML works very well with the image formats in use today by the GIS industry.

GML has also been used in connection with the JPEG 2000 standard. This new standard is called GMLJP2 and puts GML data (geometry, radiometry) inside the image data file.

This is like a souped up Geotif – not just geometry, but full radiometric description, extensible coordinate reference systems – even bundled features and annotations. I expect this standard will be widely adopted.

Page 1 of 1