Report to AuSIG
Summary of the OGC TC meeting
Hosted by Galdos Systems & Natural Resources
Canada
Vancouver, Canada, 3rd-7th
December 2001
Author:
Rob Atkinson (Social
Change Online)
The 38th
OGC technical and planning committee was hosted in Vancouver, Canada by Galdos
Systems and Natural Resources Canada. Around 100 participants from member
organizations attended. Notes and
slides from the meeting are available on the OGC members’ site at http://feature.opengis.org/members/0112tc/index.htm.
This
message provides a number of insights and viewpoints to supplement the
materials to be found on the OGC Web Site. A number of sessions run in parallel
at each TC, and the practical technical issues of the OpenGIS Web Services
initiative (OWS-1) were a focus both for the TC for the correspondent.
The AuSIG
enabled the author to attend the meeting through a contribution to the travel
cost, for which the author would like to express his gratitude.
If you have
any further questions on OGC and SDI related issues, please don’t hesitate to
contact us: rob@socialchange.net.au
The most significant outcome of the TC meeting was the approval of the WFS 1.0 and WMS 1.1.1 specifications. Especially the Web Feature Service (WFS) specification has been long awaited, and will enable vendors and data providers to create standards based access to vector data.
OGC members can view the specifications online at http://feature.opengis.org/members/pending.htm
A common model for describing real world observations has been under intense development and was agreed upon by the OWS "Sensor Web" working group. This has wide ranging import in its ability to reduce modelling costs and increase interoperability of many real world applications. The models will eventually be available as both UML and XML-schema (GML) templates for re-use.
Of note:
i. Simon Cox (CSIRO, Minerals and Exploration, Perth) is leading the modelling effort and authoring a discussion paper
ii. Reviews of scientific literature and standards efforts have been undertaken
iii. The end result can be encoded with GML (see the following points)
iv. The model is robust enough to cover all sorts of real world cases, including groups of different sensors, homogeneous arrays of sensors (eg remote sensing), fixed and moving mounts, post-processed archives etc.
Author's note: this is a significant breakthrough in both semantic interoperability and opportunity to build SCOTS solutions. The results of this work will be exploited in an on-the ground project to create a distributed water quality archive system for NSW DLWC.
A means of reconciliation of Coverage and Feature Collection models has been agreed. This allows data to be delivered in XML as both features or as a compact format (coverage). This is great news as it means that it will be relatively easy to build technologies that can deliver either format.
There is a move to shortcut the ISO timeframes by creating a new work item to refine the process of implementing a "model driven architecture". This would allow UML models to be predictably and automatically implemented as GML application schemas. There may be some syntactical refinement of GML because of this work, however IMHO it is likely to be a medium term issue. A real world project should probably model its data properly anyway and then create interim GML schemas pending opportunities to standardise across similar projects. This simply improves standardisation opportunities in the long term.
The Web Feature Server 1.0.0 has been agreed as an implementation specification. This specification supports access to Feature Collections if you know how to interpret the data. OWS-1 is developing through the "common architecture" thread how to refine this so that the WFS can advertise well known shared data schemas and better information about how to extract meaningful views of available data.
Natural Resources Canada is developing a national data set using Landsat 7 data, with a rigid quality control methodology and distributed sub-contractual approach. This will be made available for free using tile download and OGC Web Map Server access protocols as a national framework data set, to assist in improving the quality of other data sets.
OGC has committed some resources to developing a set of tests for WMS - these have to be encoded coherently into a database and distributed. Two organisations are currently implementing database-driven testing platforms; Logicon-TASC for US classified environments and Social Change Online for online access through the OGC Network.
The themes for the forthcoming TC will include two areas where many activities are taking place within Australia:
i) State and Local government interaction
ii) Development of "enterprise architecture" - for large distributed problems including National SDIs
Upcoming OGC programs are focusing on "information community" based initiatives. At this stage there remains a significant amount of work to do refining the "policy" aspects of shared catalogs and services. At this stage no programs are currently sponsored to specifically address the design of SDI services.
The next TC meeting will be hosted in New York, February 4-8, 2002 (http://www.opengis.org/events/nexttc.htm)