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

RE: [semweb-dev] Developing an overall ontology for news

Expand Messages
  • misha.wolf@thomsonreuters.com
    Hi Jarred, I don t find the use of the word ontology particularly helpful in this context. The IPTC has a taxonomy of Media Topics and an (older) taxonomy of
    Message 1 of 3 , Oct 21, 2011
    • 0 Attachment
      Hi Jarred,

      I don't find the use of the word "ontology" particularly helpful in this
      context.

      The IPTC has a taxonomy of Media Topics and an (older) taxonomy of
      Subject Codes. Thomson Reuters has its own taxonomy of News Subjects.
      There are many other such taxonomies.

      In my view, the word "ontology" can be used helpfully to describe
      something like G2 (or rNews) itself.

      I don't see it being useful for describing taxonomies.

      And I think that mixing taxonomies and structures together would be a
      thoroughly Bad Thing.

      Regards,
      Misha


      -----Original Message-----
      From: iptc-semweb-dev@yahoogroups.com
      [mailto:iptc-semweb-dev@yahoogroups.com] On Behalf Of Jarred McGinnis
      Sent: 21 October 2011 13:02
      To: iptc-semweb-dev@yahoogroups.com
      Cc: Silver.Oliver@...
      Subject: [semweb-dev] Developing an overall ontology for news

      Hello all,

      http://data.press.net/ontology

      Stuart asked me to write to the group introducing the SNaP ontology that
      we developed at the Press Association. We developed it in the last six
      months as part of a green fields project to replace our current IT
      infrastructure. The purpose of the ontology was to provide a holistic
      representation of the news. We wanted to represent not only the assets
      produced by journalists but how those assets relate to current
      categorisation schemes both native and IPTC as well as a representation
      of the world which appears within news content (I.e. The people, places
      and events that are news worthy).

      My hope is that this ontology be the starting point (we know it works
      for at least one news agency) to develop a shared ontological (not just
      PA) model for news with the hope that it could become an IPTC standard.

      This IPTC news ontology could form the foundation for other news genre
      ontologies. For example, we can develop specialised ontologies for sport
      (the BBC's sport ontology being a good place to start there)
      entertainment or even more finely grained ontologies for football
      matches or film premieres by subclassing more general concepts from
      SNaP.

      A couple emails have already been passed around (thanks Jean-Pierre and
      Stuart for the questions and suggestions), but I think it would be good
      to put some focus on it.

      What do the others think? Any questions, comments?

      Best,

      Jarred

      This email is from the Press Association. For more information, see
      www.pressassociation.com.
      This email may contain confidential information.
      Only the addressee is permitted to read, copy, distribute or otherwise
      use this email or any attachments.
      If you have received it in error, please contact the sender immediately.

      Any opinion expressed in this email is personal to the sender and may
      not reflect the opinion of the Press Association.
      Any email reply to this address may be subject to interception or
      monitoring for operational reasons or for lawful business practices.




      ------------------------------------

      Yahoo! Groups Links




      This email was sent to you by Thomson Reuters, the global news and information company. Any views expressed in this message are those of the individual sender, except where the sender specifically states them to be the views of Thomson Reuters.
    • misha.wolf@thomsonreuters.com
      Hi Jarred, IMO, this is completely detached from reality. The logical flaws are so fundamental, and the conviction with which the idea is being pushed so
      Message 2 of 3 , Oct 21, 2011
      • 0 Attachment

        Hi Jarred,

         

        IMO, this is completely detached from reality.  The logical flaws are so fundamental, and the conviction with which the idea is being pushed so total, that I don’t have desire to debate the proposal in this forum.  I shall, howver, strongly oppose it within those IPTC groups in which I am active.

         

        Regards,

        Misha

         

         

        From: iptc-semweb-dev@yahoogroups.com [mailto:iptc-semweb-dev@yahoogroups.com] On Behalf Of Jarred McGinnis
        Sent: 21 October 2011 14:34
        To: iptc-semweb-dev@yahoogroups.com
        Cc: Silver.Oliver@...; newsml-g2@yahoogroups.com
        Subject: RE: [semweb-dev] Developing an overall ontology for news

         




        Hi,

         

        I’m not sure how I led you to be think this was all about taxonomies. From our perspective, we want to move away from their use, especially for ‘fine’ grained categorisations. But I’ll save that discussion for the end.

         

        I’m also not sure what you mean by mixing taxonomies with structures? A taxonomy is a structure. If your concern is the common abuse of skos’ narrower and broader terms to represent more robust relationships like geographical areas being ‘within’ others. I agree. Or just generally people using taxonomies when they really want ontologies. I doubly agree. But I am not suggesting either.

         

        G2 is certainly one way for representing ontologies. I think at first we should keep the discussion at the model level rather than the implementation though.

         

        rNews is a better example. Yes, something like rNews is what we are after but rNews is about news as it appears in web documents. We’re looking to expand on that. There is certainly overlap. rNews is about the presentation to do that you need to talk about assets. rNews has a very nice and robust representation of those assets but rNews isn’t about representing the workflow and what news looks like during the composition, packaging and delivery of news. That’s what we are doing. From cradle to grave. Additionally, rNews has a basic representation of the world that is the subject of news. It has a limited representation of the domain of news i.e. the real world. We’ve focused on making the representation much more fuller because when you have those relationships represented in triples with a layer of inference engine over it you get a lot of work for free. If it helps, here’s some slides from a presentation I gave last week (http://www.slideshare.net/TalisConsulting/semantic-news) about what we are up to.

         

        Taxonomies.

         

        Part of the news workflow is certainly taxonomies. As you say, you, us and everybody else use taxonomies to relate content. The problem is that we can’t make any links between them without manually doing the mappings for each and every taxonomy. Would it not be better for all of us to use something like skos (I’m not recommending skos. It has lots of nasty consequences once you start using inference over it.) for representing for our taxonomy and making the link via something like owl:sameAs or equivalentClass. Then because we’ve agreed on how to model taxonomies we can now share across? I know now what code or topic X that Reuters uses because they’ve said  X is sameAs IPTC Media Topic Y and PA also has an internal topic called Z that is sameAs IPTC’s Y.

         

        So while we have included the ability to holistically classify a news asset by some existing taxonomy (modelled in rdf), for example the IPTC taxonomies that you refer to. This classification ontology provides a transition and migration path for adopters of SNaP from the traditional old school world of taxonomical knowledge representation,  into  the contemporary world of semantic ontological based knowledge representation. The lions share and the raison d'être of SNaP is about the representation of news and the world it reports on ontologically NOT taxonomically. The taxonomical part exists so news publishers don’t immediately need to throw out their existing taxonomies when adopting SNaP. This is a good thing.

         

        Best,

         

        Jarred McGinnis, PhD

        Research Manager, Semantic Technologies

        PRESS
        ASSOCIATION

        www.pressassociation.com

        jarred.mcginnis@...

        T: +44 (0) 2079 637 198
        Extension: (7198)
        M: +44 (0) 7816 286 852 

         

        Registered Address: The Press Association Limited, 292 Vauxhall Bridge Road, London, SW1V 1AE. Registered in England No. 5946902


        From: iptc-semweb-dev@yahoogroups.com [mailto:iptc-semweb-dev@yahoogroups.com] On Behalf Of misha.wolf@...
        Sent: 21 October 2011 13:30
        To: iptc-semweb-dev@yahoogroups.com
        Cc: Silver.Oliver@...; newsml-g2@yahoogroups.com
        Subject: RE: [semweb-dev] Developing an overall ontology for news

         

         

        Hi Jarred,

        I don't find the use of the word "ontology" particularly helpful in this
        context.

        The IPTC has a taxonomy of Media Topics and an (older) taxonomy of
        Subject Codes. Thomson Reuters has its own taxonomy of News Subjects.
        There are many other such taxonomies.

        In my view, the word "ontology" can be used helpfully to describe
        something like G2 (or rNews) itself.

        I don't see it being useful for describing taxonomies.

        And I think that mixing taxonomies and structures together would be a
        thoroughly Bad Thing.

        Regards,
        Misha

        -----Original Message-----
        From: iptc-semweb-dev@yahoogroups.com
        [mailto:iptc-semweb-dev@yahoogroups.com] On Behalf Of Jarred McGinnis
        Sent: 21 October 2011 13:02
        To: iptc-semweb-dev@yahoogroups.com
        Cc: Silver.Oliver@...
        Subject: [semweb-dev] Developing an overall ontology for news

        Hello all,

        http://data.press.net/ontology

        Stuart asked me to write to the group introducing the SNaP ontology that
        we developed at the Press Association. We developed it in the last six
        months as part of a green fields project to replace our current IT
        infrastructure. The purpose of the ontology was to provide a holistic
        representation of the news. We wanted to represent not only the assets
        produced by journalists but how those assets relate to current
        categorisation schemes both native and IPTC as well as a representation
        of the world which appears within news content (I.e. The people, places
        and events that are news worthy).

        My hope is that this ontology be the starting point (we know it works
        for at least one news agency) to develop a shared ontological (not just
        PA) model for news with the hope that it could become an IPTC standard.

        This IPTC news ontology could form the foundation for other news genre
        ontologies. For example, we can develop specialised ontologies for sport
        (the BBC's sport ontology being a good place to start there)
        entertainment or even more finely grained ontologies for football
        matches or film premieres by subclassing more general concepts from
        SNaP.

        A couple emails have already been passed around (thanks Jean-Pierre and
        Stuart for the questions and suggestions), but I think it would be good
        to put some focus on it.

        What do the others think? Any questions, comments?

        Best,

        Jarred

        This email is from the Press Association. For more information, see
        www.pressassociation.com.
        This email may contain confidential information.
        Only the addressee is permitted to read, copy, distribute or otherwise
        use this email or any attachments.
        If you have received it in error, please contact the sender immediately.

        Any opinion expressed in this email is personal to the sender and may
        not reflect the opinion of the Press Association.
        Any email reply to this address may be subject to interception or
        monitoring for operational reasons or for lawful business practices.

        ------------------------------------

        Yahoo! Groups Links

        This email was sent to you by Thomson Reuters, the global news and information company. Any views expressed in this message are those of the individual sender, except where the sender specifically states them to be the views of Thomson Reuters.





        This email was sent to you by Thomson Reuters, the global news and information company. Any views expressed in this message are those of the individual sender, except where the sender specifically states them to be the views of Thomson Reuters.
      • Evain, Jean-Pierre
        Dear all, I second Misha s position. Beyond the look of taxonomies, the RDF of rNews recently distributed shows that the model is not sound - e.g. some of the
        Message 3 of 3 , Oct 22, 2011
        • 0 Attachment
          Dear all,

          I second Misha's position.

          Beyond the look of taxonomies, the RDF of rNews recently distributed shows that the model is not sound -> e.g. some of the classes shouldn't be classes, etc..

          In my presentations, I used to show that NewML-G2 is aligned with other audiovisual models. But rNews is breaking this into pieces and I fear this will be worse when you eventually address video. Too bad.

          I wish I would have made you think about this but I don't have the time and energy to try to influence any further. There is much more going up of direct interest to EBU.

          Best regards,

          jean-Pierre

          ________________________________________
          De : iptc-semweb-dev@yahoogroups.com [iptc-semweb-dev@yahoogroups.com] de la part de misha.wolf@... [misha.wolf@...]
          Date d'envoi : vendredi, 21. octobre 2011 15:57
          À : iptc-semweb-dev@yahoogroups.com
          Cc : Silver.Oliver@...; newsml-g2@yahoogroups.com
          Objet : RE: [semweb-dev] Developing an overall ontology for news

          Hi Jarred,

          IMO, this is completely detached from reality. The logical flaws are so fundamental, and the conviction with which the idea is being pushed so total, that I don’t have desire to debate the proposal in this forum. I shall, howver, strongly oppose it within those IPTC groups in which I am active.

          Regards,
          Misha


          From: iptc-semweb-dev@yahoogroups.com [mailto:iptc-semweb-dev@yahoogroups.com] On Behalf Of Jarred McGinnis
          Sent: 21 October 2011 14:34
          To: iptc-semweb-dev@yahoogroups.com
          Cc: Silver.Oliver@...; newsml-g2@yahoogroups.com
          Subject: RE: [semweb-dev] Developing an overall ontology for news




          Hi,

          I’m not sure how I led you to be think this was all about taxonomies. From our perspective, we want to move away from their use, especially for ‘fine’ grained categorisations. But I’ll save that discussion for the end.

          I’m also not sure what you mean by mixing taxonomies with structures? A taxonomy is a structure. If your concern is the common abuse of skos’ narrower and broader terms to represent more robust relationships like geographical areas being ‘within’ others. I agree. Or just generally people using taxonomies when they really want ontologies. I doubly agree. But I am not suggesting either.

          G2 is certainly one way for representing ontologies. I think at first we should keep the discussion at the model level rather than the implementation though.

          rNews is a better example. Yes, something like rNews is what we are after but rNews is about news as it appears in web documents. We’re looking to expand on that. There is certainly overlap. rNews is about the presentation to do that you need to talk about assets. rNews has a very nice and robust representation of those assets but rNews isn’t about representing the workflow and what news looks like during the composition, packaging and delivery of news. That’s what we are doing. From cradle to grave. Additionally, rNews has a basic representation of the world that is the subject of news. It has a limited representation of the domain of news i.e. the real world. We’ve focused on making the representation much more fuller because when you have those relationships represented in triples with a layer of inference engine over it you get a lot of work for free. If it helps, here’s some slides from a presentation I gave last week (http://www.slideshare.net/TalisConsulting/semantic-news) about what we are up to.

          Taxonomies.

          Part of the news workflow is certainly taxonomies. As you say, you, us and everybody else use taxonomies to relate content. The problem is that we can’t make any links between them without manually doing the mappings for each and every taxonomy. Would it not be better for all of us to use something like skos (I’m not recommending skos. It has lots of nasty consequences once you start using inference over it.) for representing for our taxonomy and making the link via something like owl:sameAs or equivalentClass. Then because we’ve agreed on how to model taxonomies we can now share across? I know now what code or topic X that Reuters uses because they’ve said X is sameAs IPTC Media Topic Y and PA also has an internal topic called Z that is sameAs IPTC’s Y.

          So while we have included the ability to holistically classify a news asset by some existing taxonomy (modelled in rdf), for example the IPTC taxonomies that you refer to. This classification ontology provides a transition and migration path for adopters of SNaP from the traditional old school world of taxonomical knowledge representation, into the contemporary world of semantic ontological based knowledge representation. The lions share and the raison d'être of SNaP is about the representation of news and the world it reports on ontologically NOT taxonomically. The taxonomical part exists so news publishers don’t immediately need to throw out their existing taxonomies when adopting SNaP. This is a good thing.

          Best,

          Jarred McGinnis, PhD
          Research Manager, Semantic Technologies
          PRESS
          ASSOCIATION
          www.pressassociation.com<http://www.pressassociation.com/>
          jarred.mcginnis@...<mailto:jarred.mcginnis@...>
          T: +44 (0) 2079 637 198
          Extension: (7198)
          M: +44 (0) 7816 286 852

          Registered Address: The Press Association Limited, 292 Vauxhall Bridge Road, London, SW1V 1AE. Registered in England No. 5946902
          ________________________________
          From: iptc-semweb-dev@yahoogroups.com [mailto:iptc-semweb-dev@yahoogroups.com] On Behalf Of misha.wolf@...
          Sent: 21 October 2011 13:30
          To: iptc-semweb-dev@yahoogroups.com
          Cc: Silver.Oliver@...; newsml-g2@yahoogroups.com
          Subject: RE: [semweb-dev] Developing an overall ontology for news



          Hi Jarred,

          I don't find the use of the word "ontology" particularly helpful in this
          context.

          The IPTC has a taxonomy of Media Topics and an (older) taxonomy of
          Subject Codes. Thomson Reuters has its own taxonomy of News Subjects.
          There are many other such taxonomies.

          In my view, the word "ontology" can be used helpfully to describe
          something like G2 (or rNews) itself.

          I don't see it being useful for describing taxonomies.

          And I think that mixing taxonomies and structures together would be a
          thoroughly Bad Thing.

          Regards,
          Misha

          -----Original Message-----
          From: iptc-semweb-dev@yahoogroups.com<mailto:iptc-semweb-dev%40yahoogroups.com>
          [mailto:iptc-semweb-dev@yahoogroups.com<mailto:iptc-semweb-dev%40yahoogroups.com>] On Behalf Of Jarred McGinnis
          Sent: 21 October 2011 13:02
          To: iptc-semweb-dev@yahoogroups.com<mailto:iptc-semweb-dev%40yahoogroups.com>
          Cc: Silver.Oliver@...<mailto:Silver.Oliver%40bbc.co.uk>
          Subject: [semweb-dev] Developing an overall ontology for news

          Hello all,

          http://data.press.net/ontology

          Stuart asked me to write to the group introducing the SNaP ontology that
          we developed at the Press Association. We developed it in the last six
          months as part of a green fields project to replace our current IT
          infrastructure. The purpose of the ontology was to provide a holistic
          representation of the news. We wanted to represent not only the assets
          produced by journalists but how those assets relate to current
          categorisation schemes both native and IPTC as well as a representation
          of the world which appears within news content (I.e. The people, places
          and events that are news worthy).

          My hope is that this ontology be the starting point (we know it works
          for at least one news agency) to develop a shared ontological (not just
          PA) model for news with the hope that it could become an IPTC standard.

          This IPTC news ontology could form the foundation for other news genre
          ontologies. For example, we can develop specialised ontologies for sport
          (the BBC's sport ontology being a good place to start there)
          entertainment or even more finely grained ontologies for football
          matches or film premieres by subclassing more general concepts from
          SNaP.

          A couple emails have already been passed around (thanks Jean-Pierre and
          Stuart for the questions and suggestions), but I think it would be good
          to put some focus on it.

          What do the others think? Any questions, comments?

          Best,

          Jarred

          This email is from the Press Association. For more information, see
          www.pressassociation.com.
          This email may contain confidential information.
          Only the addressee is permitted to read, copy, distribute or otherwise
          use this email or any attachments.
          If you have received it in error, please contact the sender immediately.

          Any opinion expressed in this email is personal to the sender and may
          not reflect the opinion of the Press Association.
          Any email reply to this address may be subject to interception or
          monitoring for operational reasons or for lawful business practices.

          ------------------------------------

          Yahoo! Groups Links

          This email was sent to you by Thomson Reuters, the global news and information company. Any views expressed in this message are those of the individual sender, except where the sender specifically states them to be the views of Thomson Reuters.




          This email was sent to you by Thomson Reuters, the global news and information company. Any views expressed in this message are those of the individual sender, except where the sender specifically states them to be the views of Thomson Reuters.
        Your message has been successfully submitted and would be delivered to recipients shortly.