GEOSS Future Products Workshop: Session 5 – Interoperability and Resource Discovery NOAA, Silver Spring, MD 27 March 2013 Moderator: Steve Browdy Rapporteur: David Arctur
Session 5. Interoperability & Resource Discovery - 1
Session 5. Interoperability & Resource Discovery - 2
Session 5. Interoperability & Resource Discovery - 3
Session 5 – Some Conclusions GEOWOW – EU FP7 Project for Weather, Oceans & Water –Discovery and Access Broker – semantically enriched search, results ranking, harmonize results –Generic purpose portal – user customizable, SSO, evolution/rating of resource, add semantic search capabilities Linked Data –Government data portals moving this direction –Semantically enriching catalogs will improve discovery Unique IDs –Complicated by multi-system approaches –Caution: collisions, persistence Brokering –Essential addition to GEOSS Common Infrastructure Authentication, Single-Signon (SSO) –Will benefit providers to track accesses (aggregated) –OpenID & SAML2: Need implementations & testing
Session 5 – Discussion 1 Q: Discovery and access approaches: GEOWOW architecture vs. linked data & schema.org. If search engines are looking for best ways to make data discoverable, what can GEOSS do? A: Current architecture is valid; we need to use it better. A: Scientists need to be comfortable with what they get back from models. Model science & validity has to stand up to court trial. From a regulatory viewpoint, maintaining expanded & improved architectures which currently exist in GEOSS is advisable. But may need broader search capabilities from e.g. Google, and sensor-enhanced medical appliances. Q: What would it take to make the leap to finding things easier? A: Once you discover something in geoportal, and you try to get the data, that’s not so easy. Several formats, etc. Need better interoperability between components and services. A: Depends on users. Some users want to access data but don’t care about specific content as much as others might.
Session 5 – Discussion 2 A: 3 Answers: 1) Eye on Earth example: user wants data on a map for a barchart. 2) People create apps. 3) Kids today are often comfortable coding up solutions. Q: 3 kinds of GEOSS users; where are the citizens? What are our GEOSS users, and then how can we package the product? A: Open to suggestions for definitions of users. User could be scientist trying to discover flooding data; citizen user is hard to predict. Q: about Unique ID’s, there’s an issue about duplication and persistence. There have been ID collisions, what should be done? For persistence, do we need special registries? A: we need full documentation of recommendations to avoid collisions. At least IDs are more stable than URLs. A: W3C Linked Data WG has some recommendations. Generally considered bad form to make one’s own IDs where they are already curated, but there are situations where heterogeneous IDs are unavoidable. In EPA trying to take bridge sets to work on first. (?)
Session 5 – Discussion 3 Jeff DLB: NOAA will host unique IDs only for datasets registered in NOAA ID registry. Comment: Comments on how to use a given data page could be made at some other site for feedback, eg, at ESIP. George: Helper App concept from AIP-4: when you discover data, you also discover a helper you could use to read it. Has GEOWOW put the Helper App concept to use? A: yes, GEOWOW will apply this approach. Q: would the helper app be sensitive to different users needing to open a different application from each other? A: user would discover a dataset, as well as tools that could access the data SJSK: DOIs being done at NASA; a dataset DOI resolves to a landing page. Publishers would get a different DOI. SB: SIR/CSR now experimenting with passing MIME types; helpers could look at mime types to tell what to do next
Session 5 – Discussion 4 Q: how can I keep my broker from breaking because someone changed their input format? A: Provider contracts, notifications, governance oversight Q: Privacy laws prohibit collecting personal identification info in US. How will this play out over other countries. A: DSWG made decision there would be no identifying info from user or org that would be kept; and none of metrics data collected could link back to an individual. Q: How do we connect with the right people to provide input requirements to authentication procedures? A: Not easy to get people to contact SIF for help via form; want to talk to a human. Q: NASA using LDAP with OpenID, but it’s only for http, not so easy for ftp authentication. CEOS WGISS experience: people won’t submit form, only talk by phone
Session 5 – Discussion 5 Q: How has NSIDC broker implementation worked out? Performance issues? A: so far so good