Reading Notes: Metadata

Subject classification has long been a standard of library science.  Subject classification uses hierarchical relationships to describe objects and their correlation with similar items.  In contrast, metadata is any order-of-magnitude information about information (the description of a particular object) either in digital or physical format.

Digital metadata is used to describe digital objects (i.e. text documents, images, video, etc.).  Digital metadata is ideally embedded in the object it describes so that is is not displaced if the object is used.   Some metadata is static and never changes while some is dynamic and is used for updating information about an object and how it changes over time. Metadata can also be used to extract information about a text: language, acronyms and their meaning, names of people, time and date stamps, email addresses, phrase hierarchy, etc.

 In Objects of a Biographical System, Witten states that there is little need for subject classification in a digital world.   Isn't subject classification still useful for helping the user retrieve information and find other items that may be of interest?  Metadata can also meet this need but is there anything subject classification can do which metadata cannot?

Witten discussed the possible future of MPEG-7 files and their potential ability to evaluate a few notes of music and identify similar melodies, to retrieve graphics or logos from a few user-made digital brush strokes, identify the source of sounds from pitch samples, and to describe movements from actions in video files.  Have there been any like developments since this work was published? Are projects aiming to discern human gestures and postures from video recordings likely done using metadata from MPEG-7 files or some other technology?

Anne Gilliland discusses metadata's long-term benefits.  She says the following,

"What we do know is that the existence of many types of metadata will prove critical to the continued online and intellectual accessibility and utility of digital resources and the information objects that they contain, as well as the original objects and collections to which they relate. In this sense, metadata provides us with the Rosetta stone that will make it possible to decode information objects and their transformation into knowledge in the cultural heritage information systems of the future."

Is she referring to  digital objects as the future artifacts of our culture?  If so, what digital objects may be of importance to our predecessors?  I wonder if they will have interoperability issues or if technology will have solved problems like interoperability by then.

Weibel's discusses difficult and unanticipated problems in the creation of a universal standard for metadata creation.  He says the following,


" To borrow from the oldest joke of the Dismal Profession, put all the data modelers in the world end to end, and you won't reach a conclusion (we did, but it took ten years to manage it)."

What kind of problems is he referring to?  How do data couplers help to find solutions?  What problems have been resolved since this article was written (Summer 2005)? 

If professionals are unable to come to a consensus on metadata standards, is it possible to create software that can identify the different standards used and accommodate them simultaneously (or even convert them to a standard format)?

1. Ian H. Witten. “How to Build a Digital Library”. Morgan Kaufmann Publisher. 2002. ISBN: 1-558-60790-0.
2. Anne J. Gilliland. Introduction to Metadata, pathways to Digital Information: 1: Setting the Stage http://www.getty.edu/research/conducting_research/standards/intrometadata/setting.html   
3. Stuart L. Weibel, “Border Crossings: Reflections on a Decade of Metadata Consensus Building”, D-Lib Magazine, Volume 11 Number 7/8, July/August 2005 http://www.dlib.org/dlib/july05/weibel/07weibel.html
   

Week 2: Notes from Works About Interoperability, Extensibility, and the World Wide Web

In addition to establishing an accepted definition for digital libraries, there is a movement within the field to establish a protocol or set of standards for building the components of digital libraries.

Having a standard protocol for digital library development allows for both interoperability (the ability to use the same components with the same functionality across different interfaces) and extensibility (allowing for future development of the components without compromising functionality).  In order to develop a protocol, one must define and standardize the components of digital libraries.

Some of the components defined in these articles are:

• repositories where digital objects are stored 
  
• handles or identifiers for digital objects
• disseminators or processes associated with digital objects that provide increased functionality
• servlets or programs containing a collection of operations and associated with a specific disseminator types

Interoperability and extensibility allow developments and information to flow across multiple sources (as information flows across servers in the World Wide Web) and allows them to be dynamic and open to improving functionality as technology improves.  An established set of protocols for building components of digital libraries would allow those components to become mobile and would allow creators of digital libraries to customize a set of tools already established (instead of re-creating the wheel with each new library).  This is not a new concept - programming languages (such as java) use an established set of tools to build and customize sophisticated programs which can be used across many interfaces.

The Digital Library Research Group at Cornell University and the Corporation for National Research Initiatives (CNRI) have been working together to establish infrastructures that provide real-life examples of interoperability and extensibility through open architecture.     They tested the interoperability of their product across repositories, disseminators, and the behaviors of their digital objects.

Their tests show that:

• Interoperability across repositories can be established using RAP (Repository Access Protocol)  interface
• The same tasks can be accomplished in separate repositories with the same outcome
• Extensibility can be established, which allows community standards to be integrated into existing digital libraries

Future work of the Cornell/CNRI group will focus on:

• Allowing more variation between repositories without compromising interoperability
  
• Creating a measurement system for interoperability
  
• More testing for interoperability
• Issues of security

William Arms discusses the history and components of the World Wide Web in his chapter, "Overview of the World Wide Web".   This chapter applies to the development of digital libraries because the web can serve as an example of a well-established network of information with both  interoperability and extensibility.  Like the architectural concepts of digital libraries described by Arms et al., the World Wide Web is supported by a set of well-defined components which are functional across systems and allow information to flow freely across networks.  Arms would argue that the world wide web is under-credited (because of it's perceived exclusion from the library world) as being significant to digital library development.


References:
1.Hussein Suleman and Edward A. Fox. “A Framework for Building Open Digital Libraries”, D-Lib Magazine, December 2001. Volume 7 Number 12. http://www.dlib.org/dlib/december01/suleman/12suleman.html.
2.William Y. Arms , Christophe Blanchi, Edward A. Overly. ”An Architecture for Information in Digital Libraries”. D-Lib Magazine, February 1997. http://www.dlib.org/dlib/february97/cnri/02arms1.html.
3.Sandra Payette, Christophe Blanchi, Carl Lagoze, Edward A. Overly. “Interoperability for Digital Objects and Repositories, The Cornell/CNRI Experiments”, D-Lib Magazine, May 1999, Volume 5 Issue 5. http://www.dlib.org/dlib/may99/payette/05payette.html.
4.ARMS, Chapter 2, or “An Overview of World Wide Web”, http://www.cio.com/WebMaster/sem2_home.html