Notes on use of Dublin Core by NewsAgent

NewsAgent for Libraries

Notes on the use of Dublin Core

These notes relate primarily to the usage of the Dublin Metadata Core Element Set (DC) by the NewsAgent eLib project - more specifically they relate to the embedding of DC in Web pages by content providers (e.g. publishers). Where content is delivered to NewsAgent in alternate formats (e.g. e-mail) a sub-set of the list of the metadata attributes that follows will be generated based on object content.

DC specification is still undergoing modification at time of writing (97-98).
Further information about the Dublin Core is available from:
<URL:http://www.ukoln.ac.uk/metadata/resources/dc.html>
NewsAgent attributes have been mapped onto the Dublin Core element set. Where no mapping is possible, NewsAgent specific META tags will be used. Additional attributes will also be used internally, e.g. to hold copyright information.
Proposed usage of elements follows. In each case, where it is proposed to make use of an element, an example of the current method of embedding DC in HTML is given - but see 1) above, this syntax is subject to potential change.
- Title - Full title, to be truncated on display for single line headlines. Mandatory
```
<META NAME="DC.Title" CONTENT="BORGES: Lessons Learned from Developing and Delivering the BORGES Information Filtering Tool">
```
  This information may be obtained from the <TITLE> tag if the above is missing:
```
<TITLE>BORGES: Lessons Learned from Developing and Delivering the BORGES Information Filtering Tool</TITLE>
```
- Subject - free keyword list. Use keywords for product names if identified by publisher. Highly Desirable
```
<META NAME="DC.Subject" CONTENT="BORGES, Information Filtering, Web, WWW, USENET">
```
  This information may be taken from the <META NAME="keywords"> tag if the above is missing:
```
<META NAME="keywords" CONTENT="BORGES, Information Filtering, Web, WWW, USENET">
```
- NewsAgent Topic - controlled keyword list. See the separate NewsAgent Topic classification scheme page for more information.
```
<META NAME="NewsAgent.Topic" CONTENT="Research, Information Retrieval Systems">
```
- Description - Short abstract.
```
<META NAME="DC.Description" CONTENT="Alan Smeaton describes an information filtering service for Web pages and USENET news">
```
  This information may be taken from the <META NAME="description"> tag if the above is missing:
```
<META NAME="description" CONTENT="Alan Smeaton describes an information filtering service for Web pages and USENET news">
```
- Contact - such as supplier at end of news item, for user followup links. Types name, email and Web address. Desirable
```
<META NAME="NewsAgent.Contact" CONTENT="John Kirriemuir">

<META NAME="NewsAgent.Contact.Email" CONTENT="lisjwk@ukoln.ac.uk">

<META NAME="NewsAgent.Contact.HomePage" CONTENT="http://www.ukoln.ac.uk/~lisjwk/">
```
- Creator - Author. 2 types, name and email. Optional
```
<META NAME="DC.Creator" CONTENT="Alan Smeaton">

<META NAME="DC.Creator.Address" CONTENT="a.smeaton@somewhere.ac.uk">
```
- Date - Expiry date to override per source default set internally by NewsAgent. Uses ISO 8601 format, yyyy-mm-dd.Optional
```
<META NAME="NewsAgent.Date.ValidTo" CONTENT="1997-01-31">
```
- Type - Editorial Label. Optional
```
<META NAME="DC.Type" CONTENT="Text.article">
```
- Format - Format identified by IMT (Internet MIME type) - further investigation needed to see if this element is required. Optional
```
<META NAME="DC.Format" CONTENT="text/html">
```
- Identifier - Necessary where metadata is supplied separately from object. Optional
```
<META NAME="DC.Identifier" CONTENT="http://www.ukoln.ac.uk/ariadne/new/borges/">
```
- Coverage - Use pragmatic subset, e.g. UK, US, Europe, Other. Optional
```
<META NAME="NewsAgent.Coverage" CONTENT="Europe">
```
- Language - Use for future, but not yet useful as all in English. Optional
```
<META NAME="DC.Language" CONTENT="en-uk">
```
- Publisher - Not used (as separate source details record held, linked by item ID)
- Relation - Possible future use for directing the NewsAgent robot. This area needs more work.
- Rights - Possible future use??? Copyright will not be specified by the publisher at source but via the System Admin Data Source Management, and it has been agreed that additional details for publishers behind each source feed would be needed.
- Source - Not used.

A full example of embedded DC for NewsAgent follows:

<META NAME="DC.Title" CONTENT="BORGES: Lessons Learned from Developing and Delivering the BORGES Information Filtering Tool">

<META NAME="DC.Subject" CONTENT="BORGES, Information Filtering, Web, WWW, USENET">

<META NAME="NewsAgent.Topic" CONTENT="Research, Information Retrieval Systems">

<META NAME="DC.Description" CONTENT="Alan Smeaton describes an information filtering service for Web pages and USENET news">

<META NAME="NewsAgent.Contact" CONTENT="John Kirriemuir">
<META NAME="NewsAgent.Contact.Email" CONTENT="lisjwk@ukoln.ac.uk">
<META NAME="NewsAgent.Contact.HomePage" CONTENT="http://www.ukoln.ac.uk/~lisjwk/">

<META NAME="DC.Creator" CONTENT="Alan Smeaton">
<META NAME="DC.Creator.Address" CONTENT="a.smeaton@somewhere.ac.uk">

<META NAME="NewsAgent.Date.ValidTo" CONTENT="1997-01-31">

<META NAME="DC.Type" CONTENT="Text.article">

<META NAME="DC.Format" CONTENT="text/html">

<META NAME="DC.Identifier" CONTENT="http://www.ukoln.ac.uk/ariadne/new/borges/">

<META NAME="NewsAgent.Coverage" CONTENT="Europe">

<META NAME="DC.Language" CONTENT="en-uk">

A fairly minimal (and more realistic :-) ) example follows:

<META NAME="DC.Title" CONTENT="BORGES: Lessons Learned from Developing and Delivering the BORGES Information Filtering Tool">
<META NAME="DC.Creator" CONTENT="Alan Smeaton">
<META NAME="DC.Subject" CONTENT="BORGES, Information Filtering, Web, WWW, USENET">
<META NAME="NewsAgent.Topic" CONTENT="Research, Information Retrieval Systems">
<META NAME="DC.Identifier" CONTENT="http://www.ukoln.ac.uk/ariadne/new/borges/">

Note that even in this example the DC.Subject META tag could be missed out.

The Dublin Core qualifiers in the above are all taken from
<URL:http://www.loc.gov/marc/dcqualif.html>
A list of editorial labels could be taken from the list of Dublin Core Types at
<URL:http://sunsite.berkeley.edu/Metadata/types.html>

Page maintained by: Andy Powell, UKOLN
Status: Working draft
Last updated: 16-Feb-1998