Re: subject indicators

Hi Magdalena,

At 09:39 24.02.2003 +0100, onet wrote:
> Currently I am writing my master thesis titled "Topic maps". When I read
> about subject indicators I feel a little bit confused :). There are
> two kinds of subjects: addressable and non-addressable.

Correct. Addressable subjects are information resources. Non-addressable
subjects are everything else.

> In case of
> these first ones, subject indicators are simply URLs to the resources
> which defined these subjects as unambiguously as possible.

No. In the case of addressable subjects (i.e. resources), *you don't need
subject indicators*, because every resource has an address (a URL) that can
be used directly to identify the subject.

For example, the XTM specification is a resource that resides at the
following address:

   http://www.topicmaps.org/xtm/1.0/

This URL can be used as an identifier for the subject "XTM specification",
as follows:

  <topic id="xtmspec">
    <subjectIdentity>
      <resourceRef xlink:href="http://www.topicmaps.org/xtm/1.0/"/>
    </subjectIdentity>
    ...
  </topic>

Note the use of <resourceRef>. This tells us that the URI is to be interpreted
as a *subject address* (i.e. the address of an addressable subject, or 
resource).

> When I want
> to assign an identity to the non-addressable subject I can use URIs (which
> contain also PSIs) or I can create my own definition inside or outside
> one of the topic maps.

In the case of non-addressable subjects, there *is* no address that can be
used to identify the subject. The Polish language, for example, is not a
resource, so we can't address it directly.

Instead, we address it *indirectly*, via a resource (a subject indicator)
that provides some "compelling indication" of the identity of the subject.
The subject indicator we use, since it is a resource, *does* have an
address (a URI), which we call the *subject identifier*. (The subject
identifier is simply the address of a subject indicator.) Here is an example:

  <topic id="polish">
    <subjectIdentity>
      <subjectIndicatorRef 
xlink:href="http://psi.oasis-open.org/iso/639/#pol"/>
    </subjectIdentity>
    ...
  </topic>

Here we use http://psi.oasis-open.org/iso/639/#pol as a subject identifier for
the Polish language. (Try following the link and you will find the subject
indicator - the resource that provides humans with an explanation of what the
subject is.)

Note the use of <subjectIndicatorRef> in this second example, instead of
<resourceRef>. This tells us that the URI is to be interpreted as a *subject
identifier* (i.e., the address of a subject indicator), NOT as a *subject
address*.

URLs can, in other words, be used in two ways:

* directly: to identify the resource that has that address
* indirectly: to identify whatever is signified (or indicated) by the resource
  that has that address.

Since a URI can actually be used in two ways, we need a way to distinguish
between the two usages. Topic maps provide the <resourceRef> and
<subjectIndicatorRef> element types for exactly this purpose. (Note that RDF
does not allow the same distinction to be made in syntax. This, I believe, is
one reason for what is sometimes called the "identity crisis " of the Web.)

By the way, the OASIS published subjects TC is currently working on
requirements and recommendations for published subjects. There is a draft
of the "Gentle Introduction to Published Subjects" available at

  http://www.ontopia.net/tmp/pubsubj-gentle-intro.htm

(If anyone has comments that would help make this more readable and/or
understandable, please let me know.)

I hope this helps, Magdalena. If you have any more questions, don't
hesitate to ask.

Steve

--
Steve Pepper, Chief Executive Officer 
<pepper-76mms6M3oqTR7s880joybQ@xxxxxxxxxxxxxxxx>
Convenor, ISO/IEC JTC1/SC34/WG3  Editor, XTM (XML Topic Maps)
Ontopia AS, Waldemar Thranes gt. 98, N-0175 Oslo, Norway.
http://www.ontopia.net/ phone: +47-23233080 GSM: +47-90827246