ZedAI Periodicals Iteration1
From zedwiki
Background
The periodicals profile work in this iteration was channneled towards the exploration of possible approaches that could be traversed in the definition of the content model for the periodicals profile in iteration 2. to explore each of the approaches detailed in the next section, sample instance documents were created and circulated to all members of the wg. Electronic news feeds were used as the source of the information used for the sample instance documents, however, it was recognised that such a source of information could be semantically rather limited when compared to its printed counterpart, and to this end, it was established that the Periodicals profile work in Iteration 2 would rely upon both the electronic news feeds and printed newspapers to define the normative content model (through the associated prose and schemas).
Approaches explored for the definition of the periodicals Profile content Model.
the following four approaches were explored in this iteration for the definition of the periodicals Profile content model: 1. Importing concepts from the newsML g2 vocabulary directly into an xhtml 2.0 document - In this approach, newsML g2 concepts (aka classes, elements) were directly imported into an xhtml 2.0 document through the xml namespace mechanism detailed at: http://www.w3.org/TR/xml-names
The following example illustrates this approach:
<html ...> <head> ... </head> <body> <section> <h>News</h> <section> <newsml:headline>NO MORE GRIEF <br/> A month left, let's get our lowest road toll</newsml:headline> <newsml:by>MAL HYDE, POLICE COMMISSIONER</newsml:by> <newsml:dateline>2008-12-10</newsml:dateline> <p>ALL South Australians should be delighted with the reduction in road trauma so far this year.<br />Our road toll stands at 88, compared with 112 at this time last year. ... </p> </section> </section> </body> </html>
Although this approach proved both functionally and technically feasible, there was general consensus in the group that concepts such as headline, byline and dateline were pertinent to more types of publications e.g. university compendiums ... than just newspapers, which thus manndated the creation of elements in our own namespace, that could then be used to represent a much broader range of publications.
2. representing newsML g2 concepts using the xhtml role Attribute module - In this approach, generic xhtml 2.0 concepts were specialised into NewsML g2 concepts through the role attribute module.
the following example illustrates this approach:
<html ...> <head> ... </head> <body> <section> <h>News</h> <section> <h role="newsml:headline">NO MORE GRIEF <br/> A month left, let's get our lowest road toll</h> <span role="newsml:by">MAL HYDE, POLICE COMMISSIONER</span> <span role="newsml:dateline">2008-12-10</span> <p>ALL South Australians should be delighted with the reduction in road trauma so far this year.<br />Our road toll stands at 88, compared with 112 at this time last year. ...</p> </section> </section> </body> </html>
This approach was not deemed appropriate as it was held that the usage of the role attribute module to change the nature of the element that it specialises (in this context <span>) is not appropriate. furthermore, applicability to a broader range of publications i.e. University compendiums ... was another consideration that was not satisfied by this approach.
3. Representing newsML g2 concepts using the xhtml meta information module - In this approach, generic xhtml 2.0 elements were annotated with newsML g2 concepts using the xhtml 2.0 meta information module.
the following example illustrates this approach:
<html ...> <head> ... </head> <body> <section> <h>News</h> <section> <h property="newsml:headline">NO MORE GRIEF <br/> A month left, let's get our lowest road toll</h> <span property="newsml:by">MAL HYDE, POLICE COMMISSIONER</span> <span property="newsml:dateline">2008-12-10</span> <p>ALL South Australians should be delighted with the reduction in road trauma so far this year.<br />Our road toll stands at 88, compared with 112 at this time last year. </p> </section> </section> </body> </html>
This approach was not deemed appropriate either because the xhtml 2.0 meta information module has been designed to annotate elements with metadata; newsML g2 concepts such as headline, byline and dateline were not believed to be metadata in the context of the meta information module. Once again, applicability to a broader range of publications was another consideration that this approach could not satisfy.
4. Creating our own elements to represent NewsMl g2 concepts, and our own roles to specialise generic xhtml 2.0 concepts - In this approach, we created our own elements to represent newsML g2 concepts, and our own roles to specialise generic xhtml 2.0 concepts (e.g. Section to a News Section).
The following example illustrates this approach:
<html ...>
<head>
...
</head>
<body>
<section role="d:newsSection">
<h>1. Sektion</h>
<section role="d:newsItem">
<zai:sectionHead>
<meta property="position.section">1</meta>
<meta property="position.sequence">1</meta>
<meta property="section.name">1. Sektion</meta>
<meta property="dc:copyright">Weekendavisen 2008</meta>
<meta property="dc:subject">LEDER1</meta>
<meta property="dc:title">Løgnere</meta>
<meta property="dc:date" datatype="xsd:date">04-12-2008</meta>
</zai:sectionHead>
<h role="d:headline"></h>
<zai:byline></zai:byline>
<p>Article content...</p>
</section>
</section>
</body>
</html>
this approach allowed us the flexibility to create elements that could be used by other types of publications e.g. compendiums ... given that we already are well down the track with the definition of roles for the Leisure book profile through the RDF work that the wg has undertaken thus far, Periodicals specific roles such as NewsSection, Newsitem, article will only be a logical extension to our roles taxonomy. furthermore, this approach gives us the ability to create special purpose elements such as Sectionhead to organise and encapsulate specific pieces of information e.g. metadata.
recommended approach
We recommend approach 4 for the definition of the periodicals Profile content model in Iteration 2 for the reasons detailed above. Not only does this approach offer us greater flexibility, it has the potential to cater for a much broader range of publications than newspapers.
