« INtegration of the Zachman Framework | Main | Great Article on Simplification »
November 4, 2005
Metadata Types
Metadata can be stored and defined in multiple ways. Do you agree with these classes?
Context Metadata
Context metadata is embedded into the asset itself. Semantic metadata is hidden inside the document or database and in many cases must be interpreted by higher level solutions. Inside document or web pages, we can extract metadata like subjects, topics, keywords, titles, headers, and context. Analysis of databases create metadata such as data content, data types, BI, and statistics based metadata.
Embedded Metadata
Embedded metadata is attached to the asset and specifically designed for the business of metadata. Most desktop applications allow and utilize metadata that is updated in the file properties. Property metadata enables records management and search technologies. In the database world we can retrieve metadata embedded in the Oracle tables, field names, logical elements, etc.
Associated Metadata
Associated metadata attaches metadata to an asset outside of the asset. Library card catalogs collect, store, and deliver metadata to the end user but this metadata is not inside of the book. Many collaboration and document management applications allow you at add metadata fields that enable other services to be utilized. Many metadata repositories will associate structured and unstructured metadata to the embedded and context metadata.
Thoughts?
Posted by Todd at November 4, 2005 12:01 PM
Comments
If I understand you correctly, I would rename the first one from Context Metadata to either Structural Matadata or Intrinsic Metadata. It is the metadata that can be discovered or derived directly from the content of the object.
Your second category is metadata intentionally added to the object.
The third category is metadata stored outside the object, with pointers referring to the object.
I like it.
Posted by: James Melzer at November 5, 2005 12:39 AM
You are correct James...
Posted by: RTodd at November 10, 2005 11:49 AM
Hi Todd,
What are your thoughts about these metadata classes:
Descriptive Metadata - Metadata that supports the discovery of an object. It describes the intellectual content of an object in order to aid indexing, resource discovery, and retrieval. Example elements are audience, creator, title, subject, etc. As you can see, it's similar to your thoughts on contextual metadata.
Administrative Metadata - Metadata employed to effectively manage a given object’s life-cycle. Administrative metadata may also include rights and access information that govern the object.
System Metadata - Metadata that describes the objects existence within a physical entity. In other words, the file name and size of a document on a computer. Transient in nature, system metadata is rarely described in great detail vis-à-vis having all of its elements described within a metadata application profile.
Structural Metadata - I mainly use DCMI's view.
There is one final class I'm struggling to articulate. It deals with the management of a metadata record. The reason is I'm working on a CMS project where the metadata record of an object will last longer than the actual object it describes. For instance, say a report is transferred from the CMS to another system (e.g. a records management system). A copy of the metadata record along, with some notification of what happen to the report, will have to remain the CMS in order to tell a user where to go if someone searches for it. Overtime then, the metadata record would be eventually deleted since the object is being managed in another system.
Any input would be greatly appreciated.
Mike Bannister
Agriculture and Agri-Food Canada
Posted by: Mike Bannister at November 18, 2005 4:12 PM
Hi folks - interesting Blog, Todd. I'm doing my PhD thesis right (data collection complete, analysis underway). Todd was kind enough to participate in the pilot phase. I'm looking at end user meta-data (its influence on user attitudes and level of use of a warehouse)in a data warehouse/BI context. I've come up with the following general categorization for meta-data that might useful to 'information consumers' (users) of a warehouse.
Definitional
Definitional metadata is any information that conveys the meaning of data in the warehouse (or, for example on reports or OLAP cubes) to end users. Included in this category are business definitions, calculations, business rules and allowable values. Definitional metadata answers the question: what does this data mean, from a business perspective?
Quality
Quality metadata advises users about the currency (freshness), accuracy, validity or completeness of the data in the warehouse (or on reports, queries or OLAP cubes). Quality metadata answers the business user's question: does this data possess sufficient quality for me to use it for a specific purpose?
Navigational
Navigational metadata provides users with a means to search for data (or other resource, such as a report). In other words, navigational metadata lets users 'Google' the data warehouse to search for what they need and to get an understanding of relationships between data objects of various types. Navigational metadata answers the question: where can I find the data I need?
Lineage
Lineage information tells users about the original source of data in the warehouse (or on a report) and describes what has been done to the data (for example, cleansing, transformation or aggregation) prior to being loaded into the warehouse. Lineage information answers the user's questions: where did this data originate, and what was done to it prior to being loaded into the warehouse?
I know that this is a narrow classification (due to BI context) but would be interested in any feedback.
Posted by: Neil Foshay at November 18, 2005 6:03 PM
I guess I'm the blog stopper.
Posted by: Neil Foshay at December 2, 2005 8:40 PM
