Draft Work Plan (August 2004)

This is a work plan for the DCMI ERC/Kernel Working Group, the home page for which is http://dublincore.org/groups/kernel/ .

  1. To do: formally document kernel metadata, including the who | what | when | where elements, qualifiers, different values of ``missing'', and clarification of value syntax rules.

    * So far: draft is now available at [1] for expansion and review. This is a key piece. Please say if you're interested in joining a small subgroup to take up editorial responsibility.

  2. To do: define and implement a process for anyone to publish new vocabulary terms. New terms will be marked as 'draft'.

    * So far: initial process is to create an email message following the format described in [2] and submit via email; policy is that any conforming submission that passes the laugh test will be published.

  3. To do: define and maintain an ERC-equivalent external format based on XML and perhaps HTML/XHTML for the purpose of easy record exchange with other systems.

    * So far: not started -- volunteers welcome.

  4. To do: add a ``how'' element, equivalent to DC Description.

    * So far: done -- see [1].

  5. To do: create open-source software for manipulating kernel metadata.

    * So far: sample Perl module at [3] available, but just the display and formatting stuff for now with some date normalization thrown in. As with any format, it's easier to dish it out than to interpret, so interpretation code is pending. On the plus side, you can see this module in a production system by visiting any of the URLs:

       http://ark.cdlib.org/ark:/13030/tf0v19n804?
       http://ark.cdlib.org/ark:/13030/tf0n39n82v?
       http://ark.cdlib.org/ark:/13030/kt3489n8m4?

    Prototype ERC mappings are also viewable for biomedical literature using PubMed identifiers (use ``pm'' plus id) as follows:

       http://ark.nlm.nih.gov/ark:/12025/pm10611131?
       http://ark.nlm.nih.gov/ark:/12025/pm15066887?

    Remove the question mark to access a described object, and double it to see a larger record. Any such identifiers should return eye- and machine-readable ERC/Kernel metadata.

    * Wanted: disciplined hackers to create growing test suites.

  6. To do: create best-practices for applying who | what | when | where to different object types, e.g.,
       erc/book:  author | title | pub. year | ISBN or URL
       erc/photo:  photographer | caption | date taken | URL
       erc/map:  cartographer | region | mod. date | URL
       erc/article:  author | title | pub. date | article URL
            ( + in: journal name | volume / issue | journal URL )
       erc/painting:  artist | identifier | date created | URL
       erc/artifact:  discoverer | identifier | date found | URL
       erc/specimen:  discoverer | identifier | date found | URL
       erc/code:  programmer | title | mod. date | URL

    * Wanted: subject and application area specialists to think through and debug foundation mappings of kernel elements in their areas.

[1] Draft Kernel Metadata Specification (AuguUse of uninitialized value in concatenation (.) or string at E:\listplex\SYSTEM\SCRIPTS\filearea.cgi line 455, line 108. st 2004) http://dublincore.org/groups/kernel/spec.html

[2] Draft Vocabulary Proposal Form (July 2004) http://dublincore.org/groups/kernel/propose_term.html

[3] ERC.pm Demonstration Perl Module (May 2004) http://dublincore.org/groups/kernel/ERC.pm.txt

[4] A Metadata Kernel for Electronic Permanence http://jodi.ecs.soton.ac.uk/Articles/v02/i02/Kunze/