One of the things that strikes me about Dublin Core is its flexibility. There are formatting rules, and guidelines, but there are often multiple ways to enter the same data. For this session, I’ve been using the Creating Metadata User Guide from the Dublin Core wiki (henceforth: “user guide”).
Firstly, each field is divided into a property and a value – the value describing the property of the resource (e.g., the value “A Christmas Carol” corresponds to the property “title”). Sometimes properties have what look like sub-properties to me, but are described as “creating a relationship between the described resource and a more detailed title description”. In effect (and using the example given in the user guide), “title” is the main property, and sub-properties (more detailed descriptions) are “in greek” and “in latin” for transliterations of a title. The values then correspond to the sub-titles, so “in greek” equals the title written out in Greek, and “in latin” equals the title written out in Latin, as you see below.
|in greek||(title in Greek characters)|
|in latin||Oidipous Tyrannous|
If I understand the table above correctly, it formulates a relationship (of parent to child, perhaps) between the general “title” property and the more detailed properties that describe the transliterations.
The flexibility of Dublin Core is demonstrated in the case of a “creator” property (a creator being the party “primarily responsible for making the content of a resource”). This type of relational table can also be expressed more simply.
The relational table (the user guide describes this as expressing the creator values like another resource) looks like this:
While the other option, simply dividing the two names (family, given) by a comma, can be seen in the following table.
Another important aspect of the flexibility inherent in Dublin Core: properties can be repeated to ensure interoperability* – allowing metadata creators to use terms from different controlled vocabularies simultaneously, thus increasing their findability. The example given in the user guide demonstrates different values possible for the property: “type”.
In this table, it appears that in the second column, “dctype” indicates the controlled vocabulary term used to determine the value “Software.”
The last example I’ll use to demonstrate the flexibility of forming content using Dublin Core is of the “extent” property of a resource. The user guide shows two ways to describe the extent: using text strings (as in the simpler examples above) and using “captions” that specify a numeric value, as below:
The text string table looks like this:
So far, Dublin Core metadata creation has been pretty easy to understand. The flexibility the schema allows must be an incredible boon in many cases, the way it enables the use of multiple (or no) controlled vocabularies, provides built-in interoperability, and increases findability by allowing different terms and vocabularies to describe the same resource or element. Bonus: I had too much fun creating these tables in HTML.
*I am intrigued with this interoperability aspect written into the metadata.
User Guide/Creating Metadata. [Stephanie Ruhle, Tom Baker, Pete Johnston. 6 January 2012 (Accessed 16 May 2014)]