Notes from Tuesday Jan 9, 2007 telecon.
Peter Fox, Rob Raskin, Liping Di, Meixia Deng, Karen Moe.
Also posted on the ESIP SW wiki: see http://wiki.esipfed.org/index.php/Semantic_Web#SW_Cluster_Telecon_Notes
- Introductions -
- Followup from ESIP meeting in Portland (notes posted on cluster wiki).
Rob: suggested a focus on joint projects - developing ontologies and and tools for The Earth Information Exchange (EIE). Meixia: suggsted a focus on ontologies merging, especially technologies to automate this work. There are some tools semi-automatic software tools for mapping and matching -and for concept alignment Most are light weight but show potential matches, and suggest We agreed that an activity for the cluster would be to try out these tools and post a survey of concept decomposition tools on the wiki Karen: noted that there was very good support for the development of use cases and documenting them, a small group is working in the use case template - which will be on the wiki All agreed: two phases - immediate demos and then use cases that provide an application for using and demonstrating semantic technologies. Peter: schedule at least an hour - for cluster meeting in summer
- Set agenda for monthly telecons
Topic 1 - development and agreement on a service ontology
Currently, the OGC ISO 19119 provides a high level class of service - not too useful for general use, since there is a clear ESIP need for service chaining that covers grounding, model, profile at type level modeled, but they do not do have grounding at type level, but do have it at instance level. GMU has a prototype for the use case of wildfire prediction. However, the data are not available as a service.
Topic 2 - datatype ontology
Data-types are critical, ISO 19119 provides specifications for image manipulation, geo-location, etc. but ISO 19115 does not have data types, it has metadata. GCMD has data taxonomy but is it encoded.... Luis Bermudez has been working on an encoding of these data types.
However, there are multiple classification schema - based on scientific use/meaning - e.g. nearIR band... geo-rectification and format conversion. This is relatively little comprehensive work out there. The OGC work is driven by test beds and there are a few starting to appear, The W3 activity has started. It was suggested that the military, and private companies are likely to have implementations but not to share.
SWEET does feature data types for use with sciflo. e.g. level 2 and 3 and differs from swath and grid.
Topic 3 - community process for ontology development, using http://www.planetont.org
In addition, there is SCOOP - MMI project for ontology matching, no use cases are available yet. They are currently working on instruments but other work is specific to marine biology. Tools are available to convert a taxonomy to owl, ut for mapping there are no tools. John Graybeal is key in this effort.
UAH - also has the beginnings of data mining/service ontologies they are willing to share and with NOESIS now in beta release they have experience with using SWEET and will contribute to community discussions and evolution for SWEET.
Topic 4 - workshop 1/2 or 1 day at ESIP summer on ontol - before or after.
Topic 5 - demonstration projects
+ Thetus - company in Portland, on-line demo, focus on SW, KM, lot military, Danielle Forsythe agreed to donate software to Federation. Webinar on Jan 18. sees fed as a set of SW resources rather than a place to sell.
+ GMU projects - need a catchy name!
Topic 6 - EIE semantic components - purpose to classify product and services as well as knowledge/skills science/technical interests, in an ontology for discovery (first), and description so they could be used, services can be invoked, e.g. chaining or otherwise providing visualization search.
Framework for classification. what other products beyond data products. operations, read, subset, re-project,
Smart use of services. Do MORE than what GOS does. Create new info, set up workflows, etc. Rob is the ESIP lead on this.
Karen asked about GCMD and Melanie indicated that semantics is a paradigm shift for them, ontology is much more general and formal than their structured vocabulary but they are providing information in an OWL export format. Lola Olsen was suggested as a liaison (or to appoint one) to this cluster.
Rob - Who else to invite, encourage? Suggestions were : MMI, Benno Blummenthal, Siri Joda, and would need an association with a federation member to participate. NOAA data centers would be welcome NODC, NCDC, NGDC, and also USGS?
Tuesday, February 13th , 2007 - check the wiki for call in information and times.