[unrev-II] Fw: [informit-taxonomy] PSIs of general communication - check the m out!

From: Jack Park (jackpark@verticalnet.com)
Date: Thu Feb 22 2001 - 06:56:21 PST

  • Next message: Jack Park: "[unrev-II] Visiting ZigZag Again"

    Good morning folks. Here, I am shamelessly forwarding a thread from an
    e-list being run by Sunthar Visuvalingam of Informit
     http://www.informit.com/) , a member of the Pearson Education group.
    Sunthar is serving as developmental editor on my book _XML Topic Maps:
    Creating and Using Topic Maps for the Web_, which will be published by
    Addison Wesley in September of this year. My book is the work of a total of
    16 contributing authors.

    Bernard, who contributes to this list, has recently joined the book project.
    The topic of the thread I am forwarding is that of the Public Subject
    Indicator (PSI). A PSI, in XTM-speak (XTM being the new XML markup standard
    for Topic Maps), serves as a kind of grounding mechanism for Topics. A
    Topic is supposed to be a container for information related to a particular
    Subject -- one Topic, one Subject, and lots of Occurrences of that Topic and
    lots of Associations with other Topics, a kind of semantic net or graph
    structure. Bernard's insights into the area of semantic grounding is of
    great importance to the OHS project. Of course, I believe that there
    remains plenty of room for a really great IBIS-like discussion along the
    lines this thread runs. Please read this, and enjoy. You might want to start
    at the bottom.

    Jack

    From: Bernard Vatant <bernard@universimmedia.com>

    > Kathleen wrote :
    >
    > "Am I correct in understanding that PSIs are URLs which have been
    > 'approved' by a wider community to stand for that which is being
    discussed?
    > "
    >
    > That's exactly what PSI should eventually be! But the problem is they are
    > not really at the moment, and I'll explain shortly why.
    > First the very notion of PSI is a technical term recently introduced in
    the
    > Topic Map terminology. The core of TM paradigm is the representation in a
    > system of "subjects". A formal topic is a way to gather information on its
    > subject : names, occurrences in online documents, and relationships with
    > other topics. I recommend reading www.topicmaps.org/xtm/1.0/ where all
    that
    > is defined very well all along, and the notion of subject, subject
    > indicator, and PSI. But in this specification, which is a very new thing
    as
    > you will learn by reading at that address, the notion of PSI is more a
    > wishful thinking than explicit reference to any existing stuff.
    > The core of the problem is ontological in both technical and philosophical
    > sense of the term. What is a subject ? Who will define what subjects are ?
    > Who will maintain the repositories ? How will the problem of diversity be
    > tackled ? These questions are not particular to the Topic Maps paradigm,
    > they are ever standing ones since classification was invented. But they
    are
    > central and crucial ones in Topic Maps, because the "subject" is at the
    > core of technology, and its clear definition is a condition for
    > interoperability and knowledge interchange.
    > There are various approaches to this issue in the Topic Maps community
    now.
    > The most natural one is to start from existing ontologies and thesauri and
    > represent them as Topic Maps. There is an ongoing project for representing
    > Cyc ontology in XTM, for instance. There are several ones using Open
    > Directory taxonomy. But this last one being less than stable - I've been
    > in, and can tell how it is - I have some doubts it can yield something
    > sustainable in the long run.
    > The other approach is to build from scratch ontologies thought from the
    > very beginning in a Topic Map spirit. It's what you find at Seruba.
    > Although the approach is interesting, the problem is, as your equine
    > example shows, that it's impossible for a private technical team, working
    > by itself, to grasp accurately the ontology of every community ... The
    > process of trust and community approval is a bootstrapping one. I wonder
    if
    > the Seruba team is in strategic enough position in the Topic Maps
    community
    > to start this bootstrapping process ...
    >
    > The approach I suggest in the to-be released in my contribution to the
    book
    > Sunthar made an allusion to is in short :
    > - Any attempt to build an Universal Ontology is unsustainable, from any
    > technical, philosophical and ethical viewpoint.
    > - The creation of ontologies should emerge bottom-up from community work,
    > using tools like OHS or collaborative Topic Maps editors. It's the spirit
    > of the project you can see at www.universimmedia.com/semantopic.htm
    > - A PSI should have a stable, authoritative and trustable URL, but a
    > dynamic content based on an ongoing "conversation about the subject".
    >
    > Well, it's a long way to go. The main difficulties will be human ones, not
    > technical ones. The question of PSI is a question of setting rules.
    Setting
    > smart rules is not that difficult if you are smart. We have a lot of smart
    > people around now, and various smart teams setting various smart rules of
    > their own.
    > The problem is to know how they will converge "on the same page", to quote
    > our friend Jack Park.
    >
    > Yours
    >
    > Bernard
    >
    > > From: Smith, Kathleen (PTG)
    > >
    > > Am I correct in understanding that PSIs are URLs which have been
    > 'approved' by a wider community to stand for that which is being
    discussed?
    > >
    > > To check out the PSIs of general communication, I went to www.lex4.com
    > and typed "equine" into the search box. I then tried "horse" to see what I
    > would get. The results surprised me.
    > >
    > >
    > > The results of "horse" brought back these terms:
    > >
    > > horse jumping
    > > horse racing
    > > horse sport
    > > horse sports
    > >
    > > classic horsemanship
    > > cross-country skiing trailed by horses
    > > horseback riding
    > > horses
    > > riding on horseback
    > > rodeo on horseback
    > >
    > > Within these terms are some surprises. Horse jumping and show jumping
    are
    > synonyms, but that doesn't leave room for the kind of jumping that show
    > ring hunters do (based on style, not speed and tight turns as with show
    > jumping), or cross-country riders doing three-day eventing (jumping over
    > solid objects, test of speed, stamina, and boldness). Show jumping is
    > actually a specific type of jumping within the horse community. Also, show
    > jumping is classified as being part of modern pentathlon. Although this is
    > true, what about Grand Prix jumping competitions? Or three-day
    > evening/horse trials? These sports are much more widely practiced in the
    US
    > than modern pentathlon. Show jumping needs to be a type of horse jumping,
    > along with steeple chase and cross-country.
    > >
    > > Under "horse sports" are the following entries:
    > > showjumping
    > > trick riding
    > > rodeo on horseback
    > > polo
    > > steeple chase
    > > horse racing
    > >
    > > Dressage does not show up in here, unless the implication is that it is
    > under trick riding (synonymous with classic riding in this ontology, yet
    > classic riding is a term not really used in English--the correct term is
    > classical horsemanship/riding, also known as dressage.) Trick riding
    sounds
    > more like a circus than classical riding, and it is a term dressage
    > enthusiasts would find offensive.
    > >
    > > Other sports which do not appear are endurance riding, fox hunting,
    > vaulting, three-day evening/horse trials, showing hunters, and driving (a
    > problem because it's technically not a "riding sport" as these others are,
    > although it IS a horse sport, and in this ontology horse sports come under
    > riding sports).
    > >
    > > From: Visuvalingam, Sunthar
    > >
    > > Please check out Seruba's proposed ontology for use by implementers of
    > topic maps. Loosely put, Public Subject Indicators are addresses on the
    web
    > that a community of knowledge consumers/publishers can point to as a means
    > of indicating that they are really talking about the same topic. They'll
    > play a critical role in the general acceptance of topic maps, and their
    > social function is very well described by Bernard Vatant in his
    > contribution to Jack Park's forthcoming title on Topic Maps.
    > >
    > > Sunthar
    > >
    > > P.S. The site seems to be rather sluggish at times...
    > >
    > > From: topicmapmail-admin@infoloom.com
    > >
    > > Hi list !
    > >
    > > Our company is building an ontology about subjects of
    > > general communication and we decided to publish a
    > > set of our topics to be used as PSIs. We set up
    > > a small site at http://psi.seruba.com and are
    > > eager to discuss the way the PSIs are presented as
    > > well as the chosen set. We hope this will lead
    > > to the best possible value for the TM community.
    > >
    > > Jan Algermissen
    > > --
    > > Jan Algermissen
    > > Chief Programmer
    > > SERUBA GmbH
    > > Notkestrasse 13
    > > 22607 Hamburg
    > > Germany
    > > Tel: ++49 (0)40 41 360-211
    > > Fax: ++49 (0)40 41 360-100
    > > jalgermissen@seruba.com
    > > http://www.seruba.com

    ============================================================================
    This message is intended only for the use of the Addressee(s) and may
    contain information that is PRIVILEGED and CONFIDENTIAL. If you are not
    the intended recipient, dissemination of this communication is prohibited.
    If you have received this communication in error, please erase all copies
    of the message and its attachments and notify postmaster@verticalnet.com
    immediately.
    ============================================================================

    ------------------------ Yahoo! Groups Sponsor ---------------------~-~>
    eGroups is now Yahoo! Groups
    Click here for more details
    http://us.click.yahoo.com/kWP7PD/pYNCAA/4ihDAA/IaAVlB/TM
    ---------------------------------------------------------------------_->

    Community email addresses:
      Post message: unrev-II@onelist.com
      Subscribe: unrev-II-subscribe@onelist.com
      Unsubscribe: unrev-II-unsubscribe@onelist.com
      List owner: unrev-II-owner@onelist.com

    Shortcut URL to this page:
      http://www.onelist.com/community/unrev-II

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



    This archive was generated by hypermail 2b29 : Thu Feb 22 2001 - 07:08:49 PST