• DataModel
• ExampleObjects gentofteCollection

Gentofte Revyerne

TODO: Remove references to the word 'Gentofte'

The purpose of this page is to document the metadata model chosen for the Gentofte Collection of Bil-, Båd-, and MC-Revys.

About the Gentofte Revys

• Project home page: http://www.statsbiblioteket.dk/gentofte/

• Old analysis from DOMS pilot: Årbøger

The Gentofte project is about collecting snapshots of articles present in a collection of Danish magazines. The subjects of these magazines are

• Cars (Bil Revyen)
• Boats (Båd Revyen)
• Hifi (Hifi Revyen) (Not included yet)
• Motor Cycles (MC Revyen)

The users can browse in a tree structure

  Magazine ->  Year -> Category -> Model

When a specific model is selected the user is presented with a scanned GIF or JPEG image with the option to download a PDF for printing purposes.

When a magazine has been selected the user has two other options:

• Get an alphabetically sorted list of manufacturers where she can browse each manufacturer for the known models
• Search on a model name

Technical Analysis

On the server side the data is stored in a a file system with the file path encoding all the metadata. The paths construct is as follows

  <revy>-revyen/<year>/<category>/<manufacturer>__<model>.{jpg,gif}

All files are accompanied by a pdf version. Earlier documents are gif files while later documents are jpg files.

Not all files are reviews. Examples include commercials, articles or pricelists. These will not have a manufacturer and the path will look like this. Technically, you can often decide if there should be an manufacturer based on the category.

  <revy>-revyen/<year>/<category>/<articlename>.{jpg,gif}

The central object is the specific product presented on each page. We shall refer to each such entity as an item. Ie, an item is a specific model car, boat, or motor cycle (or possibly an article or a commercial). From the above path we see that an item has the following metadata (in the case of articles and commercials, the first field are not relevant)

• Item Metadata
  Manufacturer	The company producing the item (if a review)
  Model Name	The model name as attributed by the manufacturer (if a review)
  Article Name	The title of the article (for reviews <manufacturer>_<model>)
  Parent Magazine	The magazine in which it appeared
  Publication Year	For what we know, the year in which the item appeared in the parent magazine
  Category	A human categorization in a non-controlled vocabulary
  Additional Media	The matching PDF file and image

Metadata Model

The obvious candidates for Fedora objects are the items themselves as well as the magazines. In addition to this it has also been chosen to let the manufacturer be an object. That way it is more easily extensible in the future should we want to add metadata to the manufacturers.

All the following TYPE objects has a "isObjectType" relation to doms:Type_type and an "extendsType" relation to "doms:Type_DOMS".

We use the revy namespace for relations.

• doms:Type_revy

  • PID

    • doms:<revyName>

  • Datastreams

    • DomsDC

      • dc:title=<revyName>

• doms:Type_review_item doms:extendsType doms:Type_Image and doms:Type_Text

  • PID

    • doms:<articleName>_<year>_revyItem

  • Datastreams

    • DomsDC

      • dc:title=<modelName>

      • dc:creator=<revyName>

      • dc:subject=<category>

  • Additional Relations

    • revy:isPartOfRevyYearbook (Type_revy_yearbook)

    • revy:hasManufacturer (Type_manufacturer) (optional)

• doms:Type_manufacturer

  • PID

    • doms:<manufacturer>_manufacturer

  • Datastreams

    • DomsDC

      • dc:title=<manufacturer>

• doms:Type_revy_yearbook

  • PID

    • doms:<revyname>_<year>_yearbook

  • Additional Relations

    • revy:isPartOfRevy (Type_revy)

  • Datastreams

    • DomsDC

      • dc:title=<revyname>_<year>

      • dc:creator=<revyName>

      • dcterms:issued=<year>

Example Diagram