Loading ...
Sorry, an error occurred while loading the content.

RE: [opml-dev] Re: OPML+RDF/XTM?

Expand Messages
  • Danny Ayers
    hmm - I must admit having a literal text as an attribute does seem a little unorthodox. However I m not convinced the choices made are really an obstacle to
    Message 1 of 9 , Apr 5, 2002
    • 0 Attachment
      hmm - I must admit having a literal 'text' as an attribute does seem a
      little unorthodox. However I'm not convinced the choices made are really an
      obstacle to including metadata. The <head> element could contain
      document-wide metadata in exactly the same fashion as HTML, and individual
      items could also be described like :


      <outline text="Heart of Glass.mp3" type="song" f="Blondie - Heart of
      Glass.mp3">
      <metadata>
      ... dc:creator="Debbie Harry" ...
      </metadata>
      </outline>

      or whatever. Minor changes to the DTD.

      I've not played with XTM, but you're certainly right about RDF - it can get
      a bit clunky when describing simple hierarchies. But the fact is that it
      can, whereas OPML can't describe arbitrary graph data (well, not without
      serious bending) - I think this could be a restriction on the ubiquity of
      OPML that Dave Winer and others appear to be anticipating.

      Cheers,
      Danny.

      ---
      Danny Ayers
      <stuff> http://www.isacat.net </stuff>


      >-----Original Message-----
      >From: christopher_ogden [mailto:christopher_ogden@...]
      >Sent: 05 April 2002 22:44
      >To: opml-dev@yahoogroups.com
      >Subject: [opml-dev] Re: OPML+RDF/XTM?
      >
      >
      >I've thought about combining OPML with RDF and XTM, because RDF and
      >XTM are not well suited for hierarchical content, like OPML is. But
      >it doesn't seem very feasible to connect OPML with RDF and XTM until
      >OPML (or its de facto successors) puts content into an element,
      >rather than into an attribute, like the rest of the XML world.
      >
      >If OPML were more orthodox, you could have RDF, XTM, and even SVG
      >graphics as a node element of an "OPML 2.0" outline. And I don't
      >think any RDF, XTM, or other XML tools, including browsers, would be
      >willing to implement an OPML outline as an element, or xlink to an
      >OPML node, until OPML comes within the XML fold.
      >
      >
      >
      >--- In opml-dev@y..., "Danny Ayers" <danny666@v...> wrote:
      >> Has anyone played with Outline Processor Markup Language (OPML) [1]
      >> alongside RDF or Topic Maps? Not thought this through much but I
      >would
      >> imagine there could be a lot of potential, not least for an RDF or
      >XTM
      >> editor, where you might have an outline view (backed by OPML) with
      >pluggable
      >> semantics (from XTM or RDFS).
      >>
      >> Cheers,
      >> Danny.
      >>
      >> [1] http://www.opml.org
      >>
      >> ---
      >> Danny Ayers
      >> <stuff> http://www.isacat.net </stuff>
      >
      >
      >
      >To unsubscribe from this group, send an email to:
      >opml-dev-unsubscribe@egroups.com
      >
      >
      >
      >Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
      >
      >
    • Logan, Patrick D
      ... XTM are not well suited for hierarchical content, like OPML is.
      Message 2 of 9 , Apr 5, 2002
      • 0 Attachment
        >> I've thought about combining OPML with RDF and XTM, because RDF and
        XTM are not well suited for hierarchical content, like OPML is. <<

        Here's an opinion of someone who's done a fair bit of application-specific
        graphical editor and other "information presentation" development, including
        "semantic net" tools for several frame-based reasoning systems. But precious
        little OPML work. (i.e. I wrote tree display for OPML once).

        Any arbitrary graph can be displayed as an outline by selecting the paths to
        be outlined. The outline is one of many possible views of the graph.

        graph ======= Translator to HTML ==> HTML
        + preferences


        graph ======= Translator to OPML ==> OPML
        + preferences

        graph ======= Translator to XTM ==> topic map
        + preferences

        graph N ======= Translator to graph ==> graph N+1
        + preferences

        I think OPML should be seen as a format some devices can display, rather
        than as a general purpose notation.
      • Danny Ayers
        Hi Patrick, ... Is any of this work available online? I ve been working (on and off) on related stuff myself (http://www.isacat.net/2002/svolgo/index.htm), and
        Message 3 of 9 , Apr 6, 2002
        • 0 Attachment
          Hi Patrick,

          >Here's an opinion of someone who's done a fair bit of application-specific
          >graphical editor and other "information presentation" development,
          >including
          >"semantic net" tools for several frame-based reasoning systems.
          >But precious
          >little OPML work. (i.e. I wrote tree display for OPML once).

          Is any of this work available online? I've been working (on and off) on
          related stuff myself (http://www.isacat.net/2002/svolgo/index.htm), and this
          is one field at least where seeing other people's approaches can save vast
          amounts of wasted effort.

          >Any arbitrary graph can be displayed as an outline by selecting
          >the paths to
          >be outlined. The outline is one of many possible views of the graph.
          >
          >graph ======= Translator to HTML ==> HTML
          >+ preferences
          >
          >
          >graph ======= Translator to OPML ==> OPML
          >+ preferences
          >
          >graph ======= Translator to XTM ==> topic map
          >+ preferences
          >
          >graph N ======= Translator to graph ==> graph N+1
          >+ preferences

          Nice way of putting it.

          >I think OPML should be seen as a format some devices can display, rather
          >than as a general purpose notation.

          That makes sense, but including RDF payload would put OPML on the left-hand
          side of the translations above, in fact the 'translation' stage would be
          redundant for :

          OPML (+graph) => OPML-aware app
          OPML (+graph) => RDF-aware app

          Cheers,
          Danny.
        • patrickdlogan
          ... No. I did all this 15 years ago on Lisp machines and I never owned the work myself. I ve been reviving some related ideas, though, in JScheme. This is
          Message 4 of 9 , Apr 6, 2002
          • 0 Attachment
            >> "semantic net" tools for several frame-based reasoning systems <<

            > Is any of this work available online? <

            No. I did all this 15 years ago on Lisp machines and I never owned
            the work myself.

            I've been reviving some related ideas, though, in JScheme. This is
            getting off topic, though, so ob-OPML...

            > ...would put OPML on the left-hand side of the translations... <

            Sure. For my own interests I would draw it like...

            <format-x> => <intermediate-format> => <format-y>

            OPML can be on the left and right, but it's not my choice for the
            middle.
          • Danny Ayers
            ... JScheme? Not seen that yet - I played around with a bit of Scheme (Dr. Scheme) a while ago (trying to think declaratively) I normally work in Java so
            Message 5 of 9 , Apr 7, 2002
            • 0 Attachment
              >> Is any of this work available online? <
              >
              >No. I did all this 15 years ago on Lisp machines and I never owned
              >the work myself.
              >
              >I've been reviving some related ideas, though, in JScheme. This is
              >getting off topic, though, so ob-OPML...

              JScheme? Not seen that yet - I played around with a bit of Scheme (Dr.
              Scheme) a while ago (trying to think declaratively) I normally work in Java
              so JScheme sounds great.

              >> ...would put OPML on the left-hand side of the translations... <
              >
              >Sure. For my own interests I would draw it like...
              >
              ><format-x> => <intermediate-format> => <format-y>
              >
              >OPML can be on the left and right, but it's not my choice for the
              >middle.

              That much I certainly agree on, though with the kind of architecture I've
              been playing with, instead of an intermediate format as such you'd have an
              intermediate representation, within a model in an app.

              Any further thoughts in this arena would be welcome ;-)

              Cheers,
              Danny.
            • christopher_ogden
              ... individual ... That s great if your goal is merely to annotate an outline. But in OPML you can t readily create an outline of XTM topics or RDF
              Message 6 of 9 , Apr 9, 2002
              • 0 Attachment
                --- In opml-dev@y..., "Danny Ayers" <danny666@v...> wrote:
                > ... The <head> element could contain
                > document-wide metadata in exactly the same fashion as HTML, and
                individual
                > items could also be described like :
                > <outline text="Heart of Glass.mp3" type="song" f="Blondie - Heart of
                > Glass.mp3">
                > <metadata>
                > ... dc:creator="Debbie Harry" ...
                > </metadata>
                > </outline>
                >
                > or whatever. Minor changes to the DTD.

                That's great if your goal is merely to annotate an outline. But in
                OPML you can't readily create an outline of XTM topics or RDF
                statements. In other words, though you can annotate an outline, you
                can't make an outline of annotations and topics.
              • Danny Ayers
                ... I ve not spent enough time with TMs to know there, but you certainly can t go from arbitrary RDF - outline directly because of the graph - tree issue. I
                Message 7 of 9 , Apr 9, 2002
                • 0 Attachment
                  >That's great if your goal is merely to annotate an outline. But in
                  >OPML you can't readily create an outline of XTM topics or RDF
                  >statements. In other words, though you can annotate an outline, you
                  >can't make an outline of annotations and topics.

                  I've not spent enough time with TMs to know there, but you certainly can't
                  go from arbitrary RDF -> outline directly because of the graph -> tree
                  issue.

                  I think perhaps talking in terms of annotations may be misleading in this
                  context, annotations are usually intended to be human-readable and have
                  usually just a single purpose, but RDF is definitely intended for machine
                  consumption and usually XTM is too, and both have a vast number of potential
                  uses.

                  BTW, does an outline really have to be a tree anyway? Wouldn't an outliner
                  that supported graphs be a lot more versatile (a la MindMaps). You could
                  always use it for OPML-style outlines (as a special case, a flag maybe). But
                  would it still be called an outline, or a mindmap ;-)

                  Cheers,
                  Danny.
                Your message has been successfully submitted and would be delivered to recipients shortly.