Valeria and I are in Tucson, participating in the inception workshop of the ISE-NIFA project on a global Knowledge System for Range Lands.  we will report life from the workshop.  Photos are at 

http://www.flickr.com/photos/johanneskeizer/sets/72157624909516446/.

The workshop has participants from University of Arizona, UC Davis, University of Idaho and an external partner is also Rangeland Australia.

After introduction of the participants (jk explains FAO's interest in Rangeland Global)  Barbara Hutchinson gives an introduction and an overview of the history of the Rangeland West information systems

 Interesting presentation by University of Davis on Goolge Earth tours to show Rangeland issues. You can create own Google earth databases by uploading maps and photos on the google infrastructure.  A related service is Gigapan. www.gigapan.org.  This could be very interesting for TECA, but also to link from AGRIS records.

next presentation from University of Idaho and  Montana State University on Rangeland Science Information System.  Karen points out the enormous necessity of selection in a  situation of Information Overload. (The 200 best range land articles in our area).

At Idaho they also have project for higher education for reusable Learning objects.

Last presentation of the morning is by Matt on the state of the Rangeland West project.

----

Matt and Wolfgang presented the status of the current version of the portal, still under development. It is built on Drupal and has a nice faceted search interface. There are two collections in the system at the moment, one including all the resources that were present in the old Rangelands West portal and one including articles from the Journal of Rangelands Management.

Points for discussion: a) how to import/harvest contents with very diverse metadata formats and vocabulariees; b) how to provide a common search interface between generated contents and harvested/collected contents that use different vocabularies and have different quality of metadata?

In the afternoon, Johannes gave a presentation on Linked Open Data and how to semantically link contents from different sources; he also demonstrated alternative ways to do this like the current use that the Agris portal makes of Google and the Agris metadata to link a resource to additional related information. It was decided that, even if semantic technologies in the Rangelands portal will be implemented gradually, being a Linked Data provider and consumer should be part of a Roadmap for the project.

John presented the VIVO project and demonstrated how RDF triples from any source can be queried and displayed into any web page through SPARQL queries, Json responses and javascript, or simply by using the Drupal RDF modules: he showed examples using the FeedAPI module, the Feeds modules, the SPARQL endpoint module, the RDF CCK module.

The discussion went on for so long that the last presentation was postponed to tomorrow.

---

On the second day, Valeria presented the status of the AgriDrupal "reference tool" (assembling the most useful functionalities for agricultural information management) and the status of the AgroTagger service. Special attention was dedicated to the things that most interest the Rangelands project: the integration of Agrovoc, the OAI-PMH interface, and the path to best exploit RDF in Drupal. It was agreed that these solutions will be adopted by the Rangelands West portal.

Then, some technical Drupal-related issues were discussed, like the option of using two content types to best accommodate the need of a nice input interface and an advanced content model on one side and the need to flexibly import and harvest content with lower quality metadata or different vocabularies on the other. Important related issues like vocabulary management, importing the old Rangelands database and implementing a common search / browse interface on locally generated contents and imported/harvested contents were also discussed.

Other issues discussed were: a) how to best integrate the learning modules developed by the University of Idaho and by UC Davis (learning modules and learning objects would be two additional document types and resources could either be uploaded/hosted or linked remotely); b) how to adapt the former Rangelands navigation based on a long and deep classification system; c) how to implement the State rangelands portals in the new version, in particular the Arizona rangelands portal, that is currently completely driven by the global rangelands database.

The most important conclusion was the agreement on a RoadMap that implements the foreseen functionalities in phases, based on an Agile approach: each phase would comprise one or more iterations. Tentative deadlines for three main phases could be:

1) 1st January 2011: RW version 0.5, built on Drupal 6 and building on the current prototype, adding the Agrovoc functionality, a basic OAI-PMH data provider interface using the DC metadata set, the Arizona RW State portal. Regarding contents, this version woudl include the former Rangelands West database and the archive of the Journal of Rangelands Management.

2) August 2011: RW version 0.8 (possibly built on Drupal 7), adding new sources, exposing an RDF store, implementing more export formats also in OAI-PMH and implementing some form of harvesting on a certain number of sources. This version would include a workflow to create State portals aggregating state-related contents from the global database.

3) August 2012: RW version 1.0, implementing all planned functionalities.