+ Free Format data model fields and extensibility mechanism

When a data provider cannot use the proposed value fields (for instance in data product type), how can he/she still provide some information to characterise his/her data sets?

  1. How could we allow for extensibility? If we need to consider new kinds of science data product , not yet defined in the ObsTAP fields , how should we proceed?
  2. Description of a data set - obs_title. In ObsTAP , it is not a mandatory field, it could become important in future releases if its usage is well-defined. At the moment it is just free-format, and that could be an issue.


Regarding extensibility, I think we should take the advice of MarkusDemleitner regarding DataProduct, and have the values added as IVOA-controlled and registered enumerations. If possible, both a hierarchical and yuxtapositional notation ( la UCD) should be used.

Regarding obs_title, I do not think we should rely on it for discovery. We should try to push for obs_creator_did, or at least for obs_publisher_did, as the ways for PIs to refer to VO datasets in their literature. Conversely, use of bib_reference should be allowed for discovery, and substituted by a VARCHAR(19) (length of a bibcode), or bigger if DOIs where to be allowed (something like VARCHAR(2048) should cover most ground for existing DOIs.

-- JuanDeDiosSantanderVela - 07 Mar 2011

As far as I understand obs_title is a free text description. Maybe usefull but the field name is confusing I think. Nothing to do with an ID.

-- FrancoisBonnarel - 15 Mar 2011

Topic revision: r7 - 2011-03-16 - FrancoisBonnarel
 
This site is powered by the TWiki collaboration platformCopyright © 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback