Position Paper

from

Galdos Systems Inc

Workshop on Position Dependent Information Services

Feb 15/16 2000

--------------------------------------------------------------------------------

1.0 Topics we would like to see addressed:

We are a software developer, and a member of the OpenGIS Consortium, with a particular interest in the use of XML technologies for encoding, exchanging and transforming geographic information. We are particularly interested in the following specific areas:

Layered architectures for position/direction and other information to support mobile computing

XML Encoding for Geography (We have a first cut paper on this - see GML - this is an OGC RFC)

RDF and RDF/Schema foundations for XML encoding of geography - see GML/RDF. This includes spatial reference systems.

XML Schema Data Types for encoding geographic data (alternative to current GML DTD models)

Spatial Extension Functions for XSLT to handle geographic queries and transformations.

Layered Architectures:

In our view, the typical mobile user needs to have access to a variety of position dependent data/services each of which must in turn be derived from various local services and databases that vary in structure and content as the user moves from one location to another. While position/direction protocols such as POIX solve part of this problem we feel that additional layers are required to properly support the mobile user. POIX can for example provide an encoding for a route but is less than adequate to express requests respecting the geography of the users immediate environment (where am I ? what is around me?) nor for representing the response to such requests. Furthermore protocols which are directed at minimal encoding of position and direction do not provide the mechanisms to interface with existing data sets which may exists in a wide variety of formats and which may employ many different types of spatial semantics.

It is our belief that we need to develop a layered architecture in which position/direction encodings such as POIX are layered on other XML encodings (such as our proposed Geography Markup Language) of the real world geographic environment. This GML layer can then serve as both an interchange mechanism to legacy GIS systems, and as a storage format in its own right.

We believe further that these encodings need to be developed in some consistent and interrelated manner so as to minimize downstream maintenance problems. In the XML 1.0 environment this might involve the use of several physical DTD's with the appropriate use of external parameter entities to link one DTD to another.

XML Encoding of Real World Geography (GML):

A group of companies in the OpenGIS Consortium have proposed an XML encoding for geographic information. This is now available as a public RFC. The proposed encoding deals with geographic features, feature geometry and spatial reference systems. The current proposal is a fairly minimal encoding for geometry with many extensions proposed for subsequent versions. It is our intention that this support traditional GIS functionality such as overlay, routing analysis etc. as well as distributed update. We have produced a variety of maps using GML by using XSLT map styles to transform GML into SVG for viewing. If you have access to the Adobe SVG plug-in you can take a look at the demonstrations on our web site. We expect that XML encodings of geographic information will have a profound impact on GIS and most especially on the broad incorporation of geographic information into mainstream computing, long the holy grail of the spatial information systems industry. We anticipate in particular that GML can form the intermediate or source layer for data employed in PDA's and other mobile devices, either directly or through translation into POIX or simply SVG maps.

RDF and RDF/Schema:

We have started to explore the use of RDF and RDF/Schema in GML and we have written a white paper on how this might be done. This enables GML to be used to encode spatial information in RDF metadata records (e.g. the coverage property of Dublin core) and provides GML with a more formal foundation for feature types. We have also applied this to the description of spatial reference systems. We anticipate that such encodings can be used to define spatial reference systems which can then be shared over the internet and hence support automated coordinate transformation services that we feel are essential component of the mobile computing framework. We have prototyped such a coordinate transformation service using XSLT.

XML Schema Data Types for Geography:

We are currently developing an XML Schema version of GML. We will be proposing the basic element types of this schema (CoordinateList, Point, LineString, Polygon) for inclusion into the XML Schema Data Types. This would enable broad standardization of a few basic spatial primitives.

XSLT Spatial Extension Functions:

We have already shown that XSLT can be used as a powerful query, map styling and transformation tool. Using XSLT enabled servlets we have prototyped server side services for map generation (e.g. transformation of GML encoded data into SVG), coordinate transformation, and and spatial queries. It is our expectation that by standardizing on a set of spatial extension functions, XSLT can be transformed into a powerful spatial query mechanism essential for posing and responding to queries about the mobile user's environment. Our work with XSLT has also led us into how to phrase spatial queries in the distributed XSLT-engine enabled world. We will be making a proposal for an HTTP based query format for making spatial queries using XSLT and the proposed spatial extension functions.

2.0 Expectations of Workshop Outcomes

We anticipate the following outcomes from the workshop.

We hope to have some good technical discussions with other participants that will help clarify the requirements for XML encodings of geography and XML based (e.g. XSLT) geographic services for mobile computing.

We hope to have discussions which improve our understanding of the use of RDF, RDF/Schema and XML Schema to the spatial issues for the mobile user.

We anticipate finding partners who are interested in working with us on the topics outlined in Section 1.0, especially as it relates to mobile computing.

We wish to enhance the knowledge of the work of the OGC within the W3C community so that we can better integrate OGC develop standards with those developed by the W3C or by W3C related verticals.

3.0 Where we can contribute

We were one of the earliest members of the OGC (we joined in January 1994), and have many years experience in wide area spatial information systems, including both conventional GIS and integrated spatial database systems. We have a solid understanding of spatial data modeling, geometry representation and spatial reference systems. We believe we can contribute in all of these areas. We have also done a great deal of work on the semantic problems inherent in integration of wide area databases especially as relates to spatial information systems. We are part of a team which was recently awarded a contract to study semantic-based spatial schema translation by the U.S. Army. This will involve the use of XML based ontology representation mechanisms such as RDF/S and OML. Semantic based schema translation will be essential to any widespread provision of services to mobile users.

We have been in the XML arena for just over a year. We have successfully applied XML technology to the encoding of geographic information, to the generation of maps (GML transformed to SVG via XSLT map styles), to the provision of coordinate transformation services (XSLT) and to the implementation of spatial queries. Further development and exploitation of XML technologies for geographic information is a key component of our company's business.

We feel we can make significant contributions to discussions in a number of areas:

Application of RDF and RDF/Schema to position dependent services.

Application of XML/Schema to position dependent services

Development of architectures for spatial information frameworks to support mobile computing.

Application of XSLT to mapping and spatial services for mobile computing users.

Use of SVG for mapping.

Geographic data representation, spatial modeling, and geometry representation.