Resource Service Description and Assumptions

Description

We need a better summary to describes the 4 services currently identified.

Resource Entity

A Resource is something that is tracked as an entity in the system, and which may have a fair amount of metadata and capabilities, but which is not one of the managed "objects" of a collection. These include images (in some cases, where images are first-class collection objects), documents, letters, etc. It is not that this is abstract, in the way that grouping (exhibits or collections) is. Neither are resources entirely external in the manner of the book described by a bibliographic reference. Rather, resources are things that have digital representation, that need to be tracked in the system, but that are not the objects under management in the collections. Examples include images, reports, scanned articles, etc. These may act as annotations to an object, to an event, to a person or organization. These may also be the object of annotations. These may support upload/download, viewing, transformation (to produce derivatives like thumbnails, etc.), indexing for search, deduplication.
Note that a reference to an external resource (e.g., an URL to a web-page or to an image hosted on, e.g., Flickr) is not modeled as a resource, as the representation is not managed by C-Space services. Cf. Reference entities.

ResourceAnnotation

This associates a Resource (e.g., an image) instance to an object, a person, an institution, etc. There may be a function or reason defined.

Reference Entity

This defines some basic metadata to support things like bibliographic references to books, films, articles, webpages, etc. It can include URL links to a web page. It may link/point to a Resource entity if the associated article has been scanned or saved in digital form in the CMS.

Issue: The distinction between Resource and Reference may be a little confusing, especially when we consider which to use for an annotation. If we have a digitized article as a resource, would we have bibliographic metadata as part of the Resource schema, or would we need a Reference entity to hold this? Perhaps Reference is always to external abstract things.

ReferenceAnnotation

This associates a reference (e.g., a bibliographic reference to a book, article, etc.) to an object, a person, an institution, etc. There may be a function or reason defined.

Both the Annotations raise the general questions of modeling Annotation associations. These seem to support the idea of having a general annotation association, with a type and possibly an additional information entity defined per type.

Object versus Resource modeling

We will likely need to make a clear distinction between a CollectionObject (which is commonly a conceptual reference to some real-world entity), and the representations, surrogates and documentation thereof. Images are representations, and so are resources. 3-D and other surrogates are resources for the same reasons. Papers and books that discuss or cite an object and are managed in the system as documents are similarly resources (linked to the object(s) as annotations). However, in some workflows, images collected with specimens may be first class objects as well, which raises some modeling complexity. When is something an Object and when is it a resource? Is it a question of function? Does this get too complex for folks to manage? Can things change function over time?

Assumptions

  • TBD

Key Concepts

  • Need to work on the different between Object, Resource and Reference

Dependencies

  • Likely to depend on utility services for media support

Background Documentation

  • Need links to notes from community design meetings.
  • Need links to relevant sections in Spectrum.

Issue: to the extent that we model Images of Objects as resources, we will likely want to model derivatives as resources (web-versions, thumbnails, etc.). Does this also imply that we will  manage logical derivatives like segments (temporal segments of time-based media, or spatial segments of an image)? Would we further manage structured sets of derivatives? For a set of tiles that support zoom and pan, I would think yes. For a video playlist that presents a summary of the original media object, this may be non-trivial to support.