Geospatial Workshop (A UK, Australia & New Zealand collaboration) Report

Hosted by the  SEE-GEO Project

This page has been created subsequent to the collaborative workshop in Service-Oriented Applications of Geospatial Information for Education and Research held in Canberra on 24 July 2008. The following information is available:

• Scribes report of discussion held during the e-Framework Geospatial Workshop - saved as an attachment (Word doc) which can be accessed by scrolling to the bottom of this page

• An e-Framework report (directly below) prepared by Dr. Lyle Winton who presented at the workshop

If you would like to comment or discuss this information please register on the SUMTalk discussion page

---

Report

Kristin Stock from Social Change Online, a consultant to EDINA and also researcher at University of Nottingham, chaired the workshop.

The day kicked off with an overview of JISC and EDINA by David Medyckyi-Scott. David talked about the  Digimap project, a flagship EDINA service (10,000 users every day, 50,000 active users). 90% of people surveyed mentioned this was critical for their research. David later mentioned that the key to digimap's success is that teachers could contextualise content to their specific area. On the e-Framework front, David explained that Universities are having to interoperate with this and all the other infrastructure that is appearing. So these services need to fit in as part of the overall picture. Each country (and perhaps each institution) will have specific issues. He felt there is need for more thinking on how this (the e-Framework) will work in the University sector and how it will interoperate with other sectors.

I then gave a short presentation on the background of the e-Framework, it's principles including the service oriented approach, the technical components, and very briefly what has been happening in Australia. One comment that arose was it was difficult to know what the touch points with the e-Framework were. Paraphrasing "As a developer, I want to know how to interact with something." It was pointed out that the technical components were essentially documenting this... as a developer or community how do I correctly interact with what's been built or proposed.

Kristin gave a presentation on  SeeGeo mapping to the e-Framework. She talked about how they represent the Open Geospacial Consortium (OGC) standards in the e-Framework components. The differences between the e-Framework and web service ontologies are that the e-Framework is human readable, includes abstract concepts, and is generally more of a network of concepts. Kristin reviewed a set of geospatial and generic behaviour, as well as existing  upper level ontologies to describe these behaviours in a similar way to e-Framework Service Genres.  SUMO was the best fit as it had a good set of existing processes and operations, and so was considered for the Service Genre layer. Kristin's investigation started with the WFS (web feature service) as a service expression or a service usage model. (While the Web Map Service is the most commonly used WFS is the most interesting.)

It was generally felt that the e-Framework might be able to fast track the Service Genre mappings by looking at existing SUMO like activities. However a question was raised as to the level of confidence these activities have in their own service mappings. A practical comment was that SUMO (or Service Genres) would only be useful if the collection of things like WFS mapped through SUMO would enable someone to quickly see if something was interoperable. But the difficult is how do we get communities to do this mapping.

Chris Higgins gave a presentation on EDINA. EDINA hosts one of the National Data Centres (MIMAS is another servicing the academic sector). EDINA also hosts the GIgateway (Geospatial Information gateway?) which provides access to a range of data products. Chris showed a presentation of their data through "Digimap". They are effectively warehousing the data and making it available to students but not others (who have to sign some access agreement). At the heart of the geospatial data infrastructure is a catalogue or registry (based on Z39.50) called Go Geo.

Chris also talked about the SEcurE access to GEOspatial services ( SEEGEO) project that ended in August. Chris talked about INSPIRE (an Infrastructure for SPatial InfoRmation in Europe) which is a top down legislation within the EU member states to use data specifications (ISO 19131) and make free of charge services for discovery and viewing spatial data sets (subject to certain specific conditions). Because Universities are public sector they may have to comply with this, but it's not obvious who would do this.

Rob Woodcock talked about where the scientist are and what they do. He showed some good examples of where the principle barrier to advancing research computation is a policy issue. He used the phrase "an insurmountable small barrier". Rob talked about NCRIS but focused on  Auscope (NCRIS capability 5.13, solid earth sciences). A deliberate strategy of Auscope has been to fund the ICT gap (between themselves and NCRIS Platforms for Collaboration) so they can fund development as needs arise. See the presentation for some interesting structure/activity diagrams and examples of built portals. At one level, the National Geospatial Reference Framework, they are mapping from community schema through WFS to provide GML/XMML (now call GeoSciML), translating data sources to data services. At the layers above they are providing access to this data, storage, management, simulations, analysis and collaboration. They have a lot of skills in the tools being used and are hoping to train those skills back into the communities. Part of what they are doing now is hardening service, GeoServer and GeoNetworks, to production grade services so they can be deployed confidently. There is a technical reference group for anyone wanting to influence how these services are deployed.

Robert Gibbs, from Landcare Research and SCENZ-Grid in New Zealand, was extremely enthused about the e-Framework Geospatial opportunities that might arise from this meeting. Standards implementation in NZ is overseen by the eGovernment unit, but adoption has been slow. However, there is a climate of change in New Zealand. The Government is now purchasing, housing and managing data, and there is now backbone funding to maintain and deliver services for free (across government bodies?). This is also the first instance of government agencies taking responsibility for geospatial data assets and their use. There is also an environmental data managers group in recognition of the needs of good management.

The New Zealand government would like a collaborative test case to come out of this meeting to demonstrate the value of the e-Framework to the academic community. Up until now e-Framework engagement has been internal to the Ministry of Education. This is also an area where there is a clear link from research to education. Chris Higgins stated that JISC approach AeRIC to consider a formal collaboration on making progress in integration in the geospatial community. One suggestion was to have a persistent test-bed in which international data is linked, even if this isn't the case in the production environments. Another suggestion was to look at the ANZ Harmonised Data Framework (?). It's been slow in realisation but there is potential and there may be some holes in the methodology, taking things from a framework to implemented services.

Points to Consider

• There was also mention of several frameworks on a national scale that might be related, pushes to shared data across jurisdiction to through frameworks, standards and data infrastructure. The new government social inclusion agenda may bring this to the fore (special information to deliver social services and data)

• We need ways to make better decisions about infrastructure that doesn't joint together well. This requires a coherent metadata model expression how things might be glued together. From data, to service, to discovery service. eg. With international Authentication what do we not have to worry about?

• Does the e-Framework provide a bunch of resources people can get? Such as documentation of something useful like Geo SE/SUM. We need to test it to make sure people can really use it. It's consumer demand. It's not enough for the e-Framework registry just to be there, it needs to be linked from the resources themselves. Quality of the technical approach will underpin that. "Is it the Google maps for that academic sector?"

• Perhaps the e-Framework could separate technical from policy integration (technical integration will give us policy options). It could help in achieving integration across cyberinfrastructures in the Geospatial domain. Currently there are a few examples of this but no way you can find them. The community shouldn't have to read the full documentation to get access to these. (Analogy: you shouldn't have to read the full ISO spec to use the standard.) They should know that the standards are important and who is using them. There is a need for knowing who is using what

• We need a practical example, showing where you hit the problems. Perhaps the real world AWDIP WFS description (Australian Water Data). What are the decisions made, who made those decisions? What do the service behave like?

• We can formally describe what we are going to build, but we get stuck with the infrastructure. There is no mechanism to ask about similar things, a place where these similar things can point too. (Service Genres or SUMO services?)

• If the e-Framework components are going to be like SOA contracts, then there must be governance for each of them. So there must be someone/community who "owns" the service expressions, maintains it, and keeps it up to date. There must also be a way for consumers to contact governance. If the e-Framework is to be effective there must be valuable service expressions in it that people want and so come to the e-Framework for. This should document the touch points, people shouldn't have to read the whole of the standards documents to get access to these (to use instances of these services).