[topicmaps-comment] situated context,multiple subjectindictor ref handling

[David Dodds <drdodds42@hotmail.com>]

In 1987, the conference proceedings (Vol 851) for Space Station Automation 
III, hosted by SPIE, contained two papers written by me. Both of these 
papers were about advanced robotics software for operation of the space 
station robotics systems. The focus of these papers was the use of *situated 
actions and machine perceptions*. The key was that they were *situated*. I 
provided computer implementable means of representing situated [knowledge-] 
perspectives, such that the robotic control software would operate the 
manipulators in a more "organic" way, rather than the somewhat-clued-out or 
vacuous ways that the original "Shakey" was programmed. A decade and a half 
ago I was showing situated software control systems in conferences.

    In July of this year (2001) WROX Press published a book XML Meta Data, 
in which I have written several chapters. One of these is an example of a 
line-drawing schematic electronic-circuit, written in XML SVG 1.0, which 
contained RDF metadata IN THE PICTURE file itself. A description was 
provided of how a program could examine the SVG picture and its embedded 
metadata and produce an XML Topic Map semantic graph about the circuit 
drawing. The topic map is downloadable from
    WROX at ftp://ftp.wrox.com/professional/4516.zip, it is in chapter 11 as 
circuittm.xml, and will be available soon along with other meta data book 
things at openmeta.org. The picture is also in the ch 11  directory as 
"svgcircuitpath.svg".
    (You may need to get a free SVG viewer from Adobe download, or Jackaroo, 
or CSIRO in order to view the SVG.)
    In the Meta Data book I discuss, with accompanying computer code, how a 
program can examine a diagram (the visual content itself) and its embedded 
meta data (RDF) and use SITUATED stance to develop a context with it, and a 
topic map graph.

    It becomes reasonably clear how to do selective reification of *visual 
content* of a picture file, (not merely of the filename) such that aspects 
of the subject can be derived (through transformation, or what I call 
"conveyance") into XTM representation. An extended version of the topic map 
(circuittm.xml) is being presented at XML 2001 Orlando where it will be 
discussed how *situated contexts* can be specified by the computer and used 
to provide a context (or "scope"-like) for subjectindicators and other 
representations of the characterizations of topics.

    In XML2001 and in KT2002 I discuss and provide computer code the 
reification of situated context (processing) and the use metaphor/analogy as 
a representational mechanism for bridging the 'knowledge-as-mind-stuff and 
"knowledge" as computer-representa' gap.

also I have been posting an ongoing series at KMCI virtual chapter 
yahoogroups, on situated context processing and also on metaphor/analogy. 
anyone bothering to look at these will see that I have actually posted 
computer code as well as discussion in support of my work on situated 
context processing. XML SVG diagrams and animations
(of Edward deBono diagrams) have been included and I am in the process in 
the series of explaining how to combine topic maps technology with picture 
content/metadata via situated context.

David Dodds



>From: "Steven R. Newcomb" <srn@coolheads.com>
>To: drdodds42@hotmail.com
>CC: topicmaps-comment@lists.oasis-open.org, topicmapmail@infoloom.com
>Subject: Re: [topicmaps-comment] Notions have existence .....
>Date: Wed, 17 Oct 2001 18:08:02 -0500
>
>[David Dodds:]
> > I am interested in the potential problems with
> > identifying the subject of a topic ina wya that
> > everybody can understand what the CHARACTERIZATION of
> > the subject is.  If you and I talk about extinct dodo
> > bird eggs and design a subject indicator "extinct
> > dodo bird eggs", how does someone who speaks only
> > Swahili or Irdu supposedly to know , oh yea, its a
> > topic about the specifica subject "extinct dodo bird
> > eggs". I mention a native speaker of Irdu or Swahili
> > because in a sense that where the standard computer
> > is coming from.  IT DONT SPEAK ENGLISH, and the
> > string "extinct dodo bird eggs" has no more meaning
> > that does 1234-6789-43-55-49 as a subject indicator
> > for the same thing. Sure the computer can compare
> > strings and number-symbols and "look for" ones that
> > are indentical, but is that all we awant to mean by
> > understanding the subject indicator?!
>
>It's true that the Swahili-only speaker will find the
>string "dodo bird eggs" useless as a subject indicator.
>However, it's possible that exactly the same subject
>will appear in a topic map whose subject indicators are
>expressed in Swahili.  What the topic maps paradigm
>brings to the table is the possibility for someone who
>speaks both English and Swahili to contribute value to
>the situation by creating a *third* topic map that
>specifies the merging of our topic map with the one in
>Swahili.  This third topic can, for example, contain a
>topic that has two subject indicators, one of which is
>our subject indicator (the string "dodo bird eggs"),
>and the other of which is the expression of the same
>thing in Swahili that was used as a subject indicator
>in the Swahili-only topic map.
>
>The bilingual person who writes the third topic map is,
>of course, expressing her opinion that the two subjects
>are identical.  Topic maps make such opinions
>representable, interchangeable, and useful.  All we
>need computers to do, here, is to recognize that any
>two topics that share any single subject indicator must
>be merged, and merge them.  The virtue of the paradigm
>is that the value of the work of creating the third
>topic map (no matter how it was created, by machine or
>by human, or both) can be preserved, at least to a
>large extent, even when we (or our Swahili-only topic
>map author) produce new versions of the first two topic
>maps.
>
>The big value of the topic maps paradigm is that it can
>be used to delegate the work of recognizing
>relationships between subjects, and to preserve the
>value of that work.  People can make businesses out of
>recognizing these relationships, and everyone can
>benefit from their contributions.
>
> > Another way of me stating my quandry is that I dont
> > see how we topic mappers are avoiding the infinite
> > regression of homunculi / "understanders" in our
> > means to specifying/conveying subject indicator.
>
>No humunculi are required; just topic map authors that
>we happen to trust, and our own intelligence.
>
>-Steve
>
>--
>Steven R. Newcomb, Consultant
>srn@coolheads.com
>
>voice: +1 972 359 8160
>fax:   +1 972 359 0270
>
>1527 Northaven Drive
>Allen, Texas 75002-1648 USA


_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp