[topicmapmail] tagging vs topics

[Lars Marius Garshol]

* peter@poorbuthappy.com
| 
| That's exactly what I mean :) The NewsML example is what I mean. You
| can see why someone would want to store stuff *within* the text of a
| document, right?

Of course! That's what I was talking about in my reply to you (but
obviously I didn't explain it very well).
 
| What I am trying to understand is how do you marry these two
| different approaches. How can you embed topics in a file, while
| still having a separate topic map.

The key issue is to stop thinking that you have to embed <topic>
elements. What you have to do is figure out how to turn the
information you have in the file into topic map data. And that is
actually quite easy.

For example, the system I built has a simple XML configuration that
turns the following

  <gcapaper>
    <front>
      <title>tolog</title>
      <subt>A topic map query language</subt>
      <author>
        <fname>Lars Marius</fname>
        <surname>Garshol</surname>
        <jobtitle>Development manager</jobtitle>
        <address>
  	<affil>Ontopia</affil>
  	<cntry>Norway</cntry>
  	<email>larsga@ontopia.net</email>
  	<web>http://www.ontopia.net/</web>
        </address>

into the following topic map data (somewhat simplified):

  [tolog : paper = "tolog - A topic map query language"
    @"file://.../tolog.xml"]
  {tolog, xml-paper, "file://.../tolog.xml"}

  [lmg : person = "Lars Marius Garshol"]

  written-by(tolog : work, lmg : author)

  {lmg, jobtitle, [[Development manager]]}

  [ontopia : company = "Ontopia"]

  employed-by(ontopia : employer, lmg : employee)

  /* and so on ... */

(This is in the Linear Topic Map (LTM) notation, BTW. [foo : bar =
"..."] means topic with ID 'foo', of type 'bar', with name "...".
foo(bar, baz) is an association, {foo, bar, ...} an occurrence.)

It also generates subject indicators for all the topics, so that if I
want I can provide topic maps that make additional statements about
these topics and have the two merged automatically. That allows me to
let these two forms of information work together.

Did this help?

| It would be silly to use two separate technologies, since you'd be
| duplicating topic data in separate places and that is of course
| evil.

Well, it would be silly *if* you duplicated data. If you don't it
might not be so bad to use two files.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >