Discovery Hack-a-thon


The set of ESIP Discovery services encompass the overlapping conventions of Earth science federated OpenSearch, Collection Casting, Granule Casting, and Service Casting feed standards. To help lower the barrier of entry, we will provide a set of hands-on and simple approaches to using Discovery services. These include walking through some "low-hanging fruit" approaches to calling OpenSearch, Collection Casting, Granule Casting, and Service Casting. We will also cover a hands-on usage of the current Discovery Testbed based on the Esri GeoPortal server. See Discovery_Hack-a-thon_Details for additional resources.


What's the Plan? Get together to make some simple Discovery clients! All are welcome, no previous experience or coding skills necessary! Two back-to-back sessions on Tuesday July 17: 1:30 - 3:00, and 3:30 - 5:00.


Notes by Angela Murillo

Discovery Hack-A-Thon

OpenSearch servers

  1. Perl: Open search module
    • Application ideas: search and fetch
  2. Python
    • Use built-in XML
    • Parse OpenSearch
      • ElementTree
      • DOM
  3. XSL
    • One of easiest way is to digest XSL – XSLT
    • Can generate, cvs,
    • Saxon uses command line
  4. Java
    • parsing and publishing Atom
    • OpenSearch Query engine Parses OpenSearch

Example from Doug

  1. web app from scratch
  2. KML file to Open Search
  3. Extract bounding box


  1. have break out groups so we can go to different rooms and screens so people can see examples on screen
  2. chat interface or separate webex’s for each group
  3. attainable goal/preplanning
  4. produce results in short timeframe, having something prepackaged
  5. provide a package of things that can be downloaded
  6. less prep more coding
  7. use case that is already set
    • adding search and access into the use case
    • people ready to work on problems in real time
    • what should that challenge use case be: ex. Analysis and search, search and something else.


Within discovery how do you ingest it, consume it, from end to end lifecycle

  • are there certain segments that this lifecycle needs to be more focused on
  • What should we focus on as a community group?

Discover: both Search and Access & Advertising

  • do we want another hackathon or do we want to extend beyond that
  • whatever we work on should include both casting and searching
  • planning meeting on Friday, should discuss



Hua, H.; Lynnes, C.; Discovery Hack-a-thon; Summer Meeting 2012. ESIP Commons , June 2012


<b>Notes</b> <ul><li>No coding necessary! Many ways to interact with ESIP Discovery services (coding and non-coding)</li> <li>Chris shows a wget example (documented in the slides)</li> <li>What other tools are possible to replace curl or wget?<ul> <li>POSTman, ChromePOST, Fiddler, etc.</li></ul></li> <li>Christine talks about Geoportal, there are "test" accounts set up feeds and OpenSearch</li> <li>Ruth discusses browsers and news readers</li> <li>View Discovery resources page (</li> <li>Ruth discusses the question, "How do you find Casts and OpenSearch services"</li> <li>At NSIDC, they are writing crawlers to aggregate the casts</li> <li>From Peter Fox:<ul> <li></li> <li></li></ul></li></ul> <b>Results & Feedback</b> <ul><li>Few people were actually able to "build a client"</li> <li>Try doing a VM, as suggested by James Gallagher (with checkpoints)</li> <li>Walk through one example, then walk through example together, then give resources for further homework.</li> <li>A canned project to work on (LMS)</li> <li>Have a clear bitly link (a zip file with all the resources you need)</li> <li>Have a survey before the session to find out what the audience is interested in</li> <li>Any barriers to discovery?<ul> <li> The URL length is limited... (but OpenSearch </li> <li> There is no uniform response, even from supposed implementers </li> <li> What about information overload? There's lots of burden on the client to perform the Discovery... </li></ul></li> <li> Todd Park CTO data jams (come up with a really few cool ideas over a finite period after the "hack-a-thon"), i.e., use the "hack-a-thon" as an idea generation engine</li></ul> <b>Questions</b> <ul><li>How do we find the URL templates? - the OpenSearch description document</li> <li>Is there something fancy going on, besides just string matching? - the ESIP specs are mostly focused on standardizing the search mechanisms independent of server implementations...</li> <li>ESIP Discovery aims towards "a search infrastructure that enables [discovery] applications"</li> <li>What are some of the underlying technologies that make this work? - XML and Atom format</li> <li>What are the dependencies for datasets?</li> </ul>

Read notes at: