RE: [topicmaps-comment] Challenge Part 1 - Variable Values / Challenge Part 2 : Metadata

["Bandholtz, Thomas" <thomas.bandholtz@koeln.sema.slb.com>]

Title: RE: [topicmaps-comment] Challenge Part 1 - Variable Values / Challenge Part 2 : Metadata

Bernard,

I highly appreciate challenges like these.

-----Original Message-----
From:   Bernard Vatant [SMTP:bernard.vatant@mondeca.com]
Sent:   Monday, November 19, 2001 7:30 PM
To:     topicmaps-comment@lists.oasis-open.org; topicmapmail@infoloom.com
Subject:        [topicmaps-comment] Challenge Part 2 : Metadata

>What shall we do with variable values, and singularly
>numeric values?
...
>What shall we do with metadata on assertions, and singularly
>metadata indicating quality and source of information?

I am working in a project for the German government implementing a TM of about 75.000 existing "subject indicators", meeting challenges like yours in many areas.

We decided very early that XTM1.0 DTD cannot be used for this project. Instead we followed an XML Schema approach similar to http://www.diffuse.org/TopicMaps/schema.html by Martin Bryan. I am still wondering why Martin's Schema is not discussed deeply in the TM lists.

We use two Schemas:
1. a core Schema defining abstract generic types like <topic> <association> etc.
2. a domain Schema defining the topic and association types used by the community. Those implement and extend the abstract types of the core Schema.

One example:
Many of the project's topic subjects are events of importance in the customer's field.
First we define <event> as a topic type, extending <topic>. 
As the TM has to know the temporal extent of each event, we extend <topic> by adding a <timeExtent> attribute to <event>.

Secondly, there are different types of events, like i.e. conferences or accidents. So we define a topic type <conference> as an extension of <event>. Now we can write (excuse me for some values in German, but note the usage of the xml:lang attribute):

<topic id="_4711" type="my:conference" path="/event/conference/">
                <topname>
                        <base val="COP7"/>
                        <display lang="de" val="Siebte Vertragsstaatenkonferenz des Rahmenübereinkommens
                                                                        der Vereinten Nationen über Klimaänderungen"/>
                        <variant lang="de" val="Klimakonferenz Marrakesch"/>
                        <sort lang="de" val="Klimakonferenz Marrakesch"/>
                </topname>
                <icon iref="Cop7.JPG"/>
                <identity>
                        <subjectRef lang="en" uri="http://www.unfccc.int/cop7/index.html"/>
                </identity>
                <occurs lang="de" title="Regierungen bereit zur Ratifizierung des Kyoto-Protokolls"
                        date="2001-11-10" uri="http://www.uno.de/presse/2001/unic416.htm"/>
                <property type="timeExtent" from="2001-10-29" to="2001-11-09"/>
        </topic>

Instead of writing <topic ... type="my:conference">, <my:conference ...> would be equivalent.

>"*According to Survey AAA*, it's *possible* that, by the end of
>2001, France Telecom will control more than 30% of the Swedish
>phone market"

Using the same approach you can end with an XML encoding like the following:

<association type="marketShareProphecyAssoc" id="XYZ" plausibility="possible" percentage="30">
        <timeExtent endOfYear="2001"/>
        <members>
                <company tid="FranceTelecom"/>
                <marketSegment tid="phone"/>
                <nation locode="SE"/>
                <source tid="SurveyAAA"/>
        </members>
</association>

Hope this is stimulating, Thomas

Thomas Bandholtz
XML Competence Center
SchlumbergerSema
Sema GmbH
Kaltenbornweg 3
D50679 Köln/Cologne
++49 (0)221 8299 264