Present

Minutes

  • specification has been ratified
  • next steps are putting this onto the REFEDS website / zenodo / REFEDS wiki
  • Alex has asked Nicole if the WG needs to do anything else before the specification goes online
  • Zacharias asked about promotion: REFEDS mailing list and the next REFEDS meeting
  • Zacharias asked: what's next?
  • Albert: needs to actual type out the actual language of filtering
  • Albert has a document about what's missing
    • The specification needs to help with the interpretation of the rules, can i filter this and that, or filter this or this
    • These are all the values you can choose from, we need to specify which fields
    • We have to maintain the XML version, and do we also do it for json - we should define the stucture, we define mapping to what we're querying against
      • Maybe Xquery?
      • However we're gonna do this, we need to give examples of similiary
      • Give the ability to query without knowing the structure, trust info allows for that, but Xquery would work well for people who know the data structure very well
      • We can borrow some of the syntax from the Xquery, but we shouldn't
      • We can start with XML in the trust info language, and describe it in XML, and work with porting in a separate or next exercise
      • Both XML and JSON are both syntaxes - one could think about describing the stucture and
    • We had described how to submit, and now it goes through the federation which requires their participation, but that doesn't make it easy to start using because the sending and receiving party could have different ideas on how to do this
      • Should REFEDS, "MFA profile" explains how to send, the actual signal is protocol agnostic, the transmission mechanism is specified by the MFA profile specification - should we do the same?
      • We have defined one way of sending, but should we describe others?
        • Some SP:s might have to live in more than one discovery world, think EBSCO
        • SP:s want the specification to be specific and precise, to make it easier to integrate.
      • Maybe we don't have to make part of the standard, not defining everything, and describe things we know we haven't described.
  •  Could we agree on the fields, but not define the actual values 
    • And give example values but that isn't exclusive
  • Could we agree on how to combine the different filters - how big should the language be - we can also agree and specify this
  • If we can help both Pyff and MDA to set what the output will look like, then we have a possibility to give examples that works for most of the discovery service - but it isn't a standard written in stone.
  •  
  • No labels