innovation in metadata design, implementation & best practices

DCMI Kernel Metadata Community

Join this Community...

Moderator: John Kunze
California Digital Library
Status: This community is active.
Last Update: 2010-09-27
Established: 2006-12-18
Charter: The DCMI Kernel Metadata Community is a forum for individuals and organizations interested in very lightweight representations of Dublin Core metadata. Such representations attempt to maximize utility and minimize cost of creation, maintenance, and exchange of Dublin Core and other metadata standards that interoperate with it. A specific sub-goal of the DCMI Kernel Metadata Community is to provide a place to discuss issues that arise in relation to the DC Kernel Application Profile.

News and Announcements

2004-12-13, The next F2F Working Group meeting for DC:Accessibility will be held in melbourne in the we February 7 - 11 2005. Those interested in attending in person or via telecommunications should contact Liddy Nevile.

Resources, Tools, and Demos

Mailing List

Forum Archive:

To subscribe:

Task Group and Kernel Application Profile (KAP)

Kernel Task Group Wiki:

The DCMI Kernel Metadata Task Group was set up to complete the Kernel Application Profile (KAP). This includes the following tasks:


Have you ever wanted a citation format that is both natural to read and simple to process?

     erc: Kunze, John A. | A Metadata Kernel for Electronic Permanence
          | 20011106 |

The citation above exemplifies the simplest kernel metadata object description, answering the basic who, what, when, where questions regarding the expression of a work. In this case, the expression in question is the publication of an article, coincidentally, about a streamlined metadata record description called ERC (Electronic Resource Citation) base on a kernel set of metadata elements. Although digital permanence is the central application covered in the article, the format works well for more general resource description.

The record design starts with the Dublin Core consensus and distills out a subset of four semantic buckets - a metadata kernel - that balances the needs for adequate identification of persistent objects and for low cost metadata generation. To minimize the burden of creating, understanding, and manipulating data in those buckets, a very simple record format was created.

The basic format can be parsed by two lines of Perl code. This fundamental simplicity suggests an alternative path for the ongoing development of schemes that wish to make metadata easy to create, maintain, and interpret. The Kernel record aspires to offer a way forward for Dublin Core metadata that is simultaneously simpler, and yet more complete, compact, extensible, and international than other syntactical representations.