[Esip-discovery] Key outcomes from the Discovery Meetings at ESIP...

[Lynnes, Christopher S. (GSFC-6102)]

...with an emphasis on possible action items.  Here is what I think we captured:  Please confirm or deny, or modify:

(1) Link identification approaches:  nobody was happy with the ones I presented.  However, James Gallagher proposed an xlink approach that looks tractable.

AI:  Brian Wilson and James Gallagher to draft a DCP based on the xlink approach.

(2) OS/GEO Convergence:  Aside from #1, above, the main difference is the use of <dc:date> by OSGeo instead of <time:start> and <time:end>.

AI:  Chris Lynnes to draft a DCP to convert over to <dc:date>.  It would probably be good to lump this with #1 to form a single release.

(3) Specifying the valids and/or type of non-standard query parameters.  We saw a few different approaches (<opensearch:Query>, the GSAC, and the approach used in S2S?).  It looks like this topic warrants a followup dedicated telecon.  Yes?

(4) Best Practices for error handling need to be written; we seem to be all over the map.

AI: Any volunteers to start one up on the wiki?

(5) Faceting:  very useful for a number of use cases, well implemented in SOLR.  How do we get it into OpenSearch??

AI:  Hook to figure out a way to do this.

(6) Adding JSON encoding:  Useful for machine-to-machine, and indeed, OpenSearch is more than just Atom.

AI:  Hook(?) to propose a JSON encoding.  N.B.:  needs to solve the same link identification problem.

(7) Discovery testbed:  we need to get more active about telling ESRI what we want out of it, i.e., solid requirements and Use Cases.  We are also thinking about a hack-a-thon to register data and services into Geoportal, which might help with validation of our OpenSearch services.

Did I miss anything important??
--
Dr. Christopher Lynnes     NASA/GSFC, Code 610.2    phone: 301-614-5185