G2 CR00145: new attributes for keyword, subject

IPTC G2-Standards: Change Request document

Title: New attributes for keyword and subject

Change Request number: CR00145
Standard to be changed (delete the inappropriate ones): NewsML-G2
Standard’s version to be changed: 2.9

 
Change Request History
Version Date Author(s) Notes
1 2011-09-25 Michael Steidl on behalf of the NAR-dev group  
2 2011-09-30 Dave Compton, modified, NAR-dev group agreed  

 

Typographic note
Changes to existing documents/specifications are highlighted in yellow.
Editing notes for the spec documents are highlighted in cyan.
In drafts only: Text highlighted in magenta needs further discussion

Description of the change: (free-text)
Add a few attributes to better quantify how the keyword(s) and/or subject (s)relate to the content.


Rationale for the change:
a) add the confidence and the relevance attributes which already exist for the subject property to the keyword property. As keyword and subject have similar use cases their use will be more consistent.

b) add a how attribute to subject and keyword to explain the means by which the property value was extracted from the content.


Reference to related IPTC discussion(s) (on Yahoo groups etc):
http://tech.groups.yahoo.com/group/iptc-news-architecture-dev/message/7348



Formalised change request: (spec table etc.)

 

A) add attributes to the keyword and subject property

Header

Value

(XML) Data Model

Element

Namespace (prefix)

nar

Name

keyword / subject

Title

 

Definition

 

User Note(s)

 

Implementation Note(s)

 

XML Schema Spec

At: PCL

Datatype

 

Internally Ctrl Values

 

Externally Ctrl Values

 

Attributes

optional attribute group name

Name Cardinality; Datatype; Definition
{keyword only} confidence, 0..1; QCodeType; The confidence with which the metadata has been assigned.

{keyword only}

 relevance, 0..1; QCodeType; The relevance of the metadata to the news content to which it was attached.

{keyword and subject}

how, 0..1; QCodeType; Indicates by which means the value was extracted from the content
Note {subject only}:

- When @why has a value of http://cv.iptc.org/newscodes/whypresent/direct: @how SHOULD be used to define how the subject was extracted.

- For all other @why values: @how SHOULD be omitted (and ignored in processing), as it has an implicit value of http://cv.iptc.org/newscodes/howextracted/tool.

 

Child elements

 

XML Schema Note(s)

 

Example(s)

Example 1: Subject with @how (@why="why:direct" is implicit)

<subject qcode="N2:FR" confidence="80" relevance="50" how="how:tool"/>

 

Example 2: Keyword with @confidence, @relevance, @how

<keyword confidence="100" relevance="50" how="how:person">LANDSCAPE</keyword>

 

 

 

 

B) Define a CV to be used with the @how attribute

SchemeURI: http://cv.iptc.org/newscodes/howextracted/
Code - definition:
person - The value was extracted by a person
assisted - The value was extracted by a person assisted by a tool
tool - The value was extracted by a tool