[xtm-wg] Re: An Approach to the Semantic Web

[scott.tsao@boeing.com]

--- In xtm-wg@y..., Eric van der Vlist <vdv@d...> wrote:
> Reading your summary, I realize that I haven't answered to your
> requirement about uses cases from a user's perspective.
Thanks very much, Eric, for coming up with those use cases.  It helps 
me a lot in understanding what a person (or machine) can benefit from 
using your "semantic web site."

> A. Current
> 
> 1) Site summary
> 
> When accessing a web site, users expect to see a hierarchical 
structure
> similar to a disk directory with lists of pages per category.
> 
> This structure, on XMLfr, is expressed by a set of RSS channels.
> 
> This could have been done in XTM by defining specific topics (1 
topic
> per RSS channel).
I would call the actor for this use case an "Onlooker" who will 
quickly browse your website and see if there is anything worth 
reading from the site summary.

> 2) Syndication
> 
> I assume that our users can benefit from finding the headlines of 
XMLfr
> on news portals.
> 
> This is one of the main applications of RSS.
> 
> This could technically be done through a piece of topic map, but I 
am
> not aware of any project in that direction.
I would call the actor for this use case an "Aggregator" which will 
extract the headlines and display them on its own website.

> 3) Key words
> 
> When reading an article, users are often in need for more 
information on
> the topics mentioned in the article.
> 
> On XMLfr, this is done using RSS and the taxonomy module. The 
Taxonomy
> module provides an identifier for each topic and the association 
between
> the topic and the articles.
> 
> This is similar to XTM's topics and occurrences.
> 
> In both cases, this can lead to lists of keywords and lists of 
resources
> related to a keyword.
I would call the actor for this use case an "Explorer" who likes to 
search for documents that are related to a particular set of topics 
of interest.

> 4) Discovery
> 
> When a user wants to start studying a new topic, he needs to find 
out
> the definition of the topic, the resources related to this topic 
and the
> links between topics.
> 
> On XMLfr, I have only a partial implementation of this, relying on 
RSS
> and the taxonomy module: what is still lacking is a description of 
each
> topic and the categorization of the relations between topics.
> 
> This could of course have been done through XTM.
I would call the actor for this use case an "Investigator" who is 
already involved in the investigation of a partticular subject of 
intense interest and would like to expand the reach of his/her 
knowledge through the discovery of related information on the Web.

So, we now have four use cases for four different actors:
1) "Site summary" for Onlookers
2) "Syndication" for Aggregators
3) "Key words" for Explorers
4) "Discovery" for Investigators

I think the first two use cases provide essentially the same business 
value, i.e., a (hierarchical) list of (linkable) headlines with a 
short summary description for each.  The main difference is where 
they appear: "Site summary" at XMLfr.org; "Syndication" at all other 
web sites that seve as the Aggregator.

While the first two use cases can hardly be considered an 
implementation of the "semantic" web, the last two use cases do offer 
features that one would expect from the semantic web.  One could 
build semantics based on features provided by the RDF or XTM models 
for advanced search and/or navigation.  Again, the main difference is 
where they are performed: "Key words" at XMLfr.org; "Discovery" at 
other semantic web sites (or applications).

> 4) Conclusion
> 
> One of the purposes of my article has been to show that the 4 use 
cases
> I have currently implemented could have been done using either 
RDF/RSS
> or XTM.
For the first and third use cases, as the implementor of XMLfr.org 
you do have full control of which technology to use, RDF/RSS or XTM.  
However, for the second and fourth use cases, it will be dependent on 
the people (or machine) playing the role of Aggregator and 
Investigator, respectively.  For Aggregators such as O'Reilly's 
Meerkat the RDF/RSS-based news stories may be the preferred (or the 
only) choice, but for an Investigator like me I would prefer to 
receive XTM-based news articles.  For example, suppose I have already 
built an XTM-based application to study the subject of "semantic 
web", I would be able to easily merge your article (i.e., "Building a 
Semantic Web Site" with metadata serialized in XTM) into my web site 
(or application) by creating topics, occurrences, and associations to 
express facts such as:
a) <Eric van der Vlist> is the <author-of> an <article> <en-titled> 
<Building a Semantic Web Site> which is <published-by> <XML.com> at 
<http://www.xml.com/pub/a/2001/05/02/semanticwebsite.html>.
b) This article <describes> the use of <RDF/RSS> and <XTM> in 
constructing a <semantic web site> at <http://xmlfr.org> <in-
language> of <French>.

Regards,

Scott Tsao
The Boeing Company


------------------------ Yahoo! Groups Sponsor ---------------------~-~>
Become a member FREE at http://www.searchdatabase.com?Offer=yahoo5 today.
Save time finding the best Database Development and
Management information available on the Internet.
Our targeted search engine and dedicated editorial staff "cut out the clutter" by aggregating the best links and the most relevant database news for you
every day.
http://us.click.yahoo.com/_w3RtC/rGiCAA/bT0EAA/2n6YlB/TM
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/