Term: identifier scheme Class:  
 vernacular   (0%)
Created 9 September 2013
Last modified 19 September 2013
Contributed by Stephen Richard


Definition: A set of rules and conventions that define the scope of resources for which a collection of identifiers may proxy, the syntax for generating new identifiers, the kinds of representations that are recognized for the identified resources, and how an identifier is dereferenced to obtain the desired representation.
Examples: http://repository.usgin.org/uri_gin/usgin/dlio/331

Note after the RDA plenary 2:
I think the identifier scheme concept is very close to the 'PID type' concept discussed at the meeting, basically an identifier scheme is a definition of a PID type.
One other thing the identifier scheme should specify is a default representation type that is available for each identifier in the scheme, and how that default representation is obtained given the identifier.
Submitted 14 April 2016
by Stephen
Identifier schemes will be layered. http, ark, doi etc. are all identifier schemes (RFC3987 requires all identifiers to start with a segment that specifies a registered identifier scheme). What is missing is some convention for how to discover the identifier scheme for schemes that are 'stacked' on one of these base URI schemes, e.g. the handle system is layered on http. I suspect the best solution is to include a segment in the identifier to indicate the scheme, but a user still needs a priori knowledge to know which part of the id string specifies the 'id scheme profile'. Submitted 14 April 2016
by Stephen