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

RE: [newsml-g2] Re: How to 'Lift' metadata out of the news content

Expand Messages
  • Kelvin Holland
    Hi Lino, Sorry for the delay in replying. First, ERRATUM. I was mistaken in telling you that @rel is a mandatory property of . It is optional and if not
    Message 1 of 13 , Jun 2, 2010
    • 0 Attachment

      Hi Lino,

       

      Sorry for the delay in replying.

       

      First, ERRATUM. I was mistaken in telling you that @rel is a mandatory property of <facet>. It is optional and if not present, the implied (default) relationship is “is a”.

       

      Yes, you can add the additional properties of the concept  as facets in the way you have outlined below (at Power Conformance Level), OR,

       

      If the information must be carried inline, it may be helpful to place a snapshot of the information about the concept in a dedicated structure, the <assert> wrapper, that can be referenced by all properties in the document that use this concept ID.   

       

      The advantage of using <assert> is that it can hold all of the properties your customers need to know about the companies in one place, but you can refer to this information from multiple places in the document. So:

       

      <newsItem>

      ...

      <contentMeta>

                                      ...

                                      <infoSource qcode=”foo:companyA” />

      ...

                                      <subject type="cpnat:organisation" qcode=" foo:companyA" />

                                      <subject type="cpnat:organisation" qcode=" foo:companyB" />

      <subject type="cpnat:abstract" qcode="transtype:acquisition">

      <name>Company Acquisition</name>

      <facet rel="trel:acquirer" qcode="foo:companyA" />
      <facet rel="trel:acquiree" qcode="foo:companyB" />
      <facet rel="mrel:acqtype" qcode="bar:agreed" />

      </subject>
      ...

                      </contentMeta>

                      ...

      <assert qcode=”foo:companyA”>

                      ...

      <name role="nrole:short">ANG</name>
      <name role="nrole:full">Angloplats</name>
      <sameAs qcode="isin:ZAE123"/>
      <sameAs qcode="tidm:abc"/>
      <facet rel="crel:IsRegisteredIn" qcode="iso3166-1a2:ZA"/>
      <facet rel="crel:HasListingStatusOf" qcode="lstatus:listed"/>

      ...

      </assert>
      <assert qcode=”foo:companyB”>

                                      ...

                      </assert>
                     

      Take a look at Chapter 18 of the Guidelines – this provides a full explanation and more examples, also some advice on using local Ids (@literal), which you may find useful.

       

      Anyone else care to jump in here? I’m pondering the Primary Company, Related Company issue and how best to express that.

       

      Lino, thanks for sharing your example -  it’s an interesting use case.

       

      Regards

       

      Kelvin

       

       

      Kelvin Holland

      Point House Media Ltd

      01825 743340

      07966 448437

       

      From: newsml-g2@yahoogroups.com [mailto:newsml-g2@yahoogroups.com] On Behalf Of lino_sg
      Sent: 26 May 2010 18:15
      To: newsml-g2@yahoogroups.com
      Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news content

       

       

      Hi Kelvin,

      Thanks for the feedback and clarifications on some of the finer details. Kelvin, if I take your definition for a company:

      <subject type="cpnat:organisation" qcode=" foo:companyA">
      <name role="nrole:short">ANG</name>
      <name role="nrole:full">Angloplats</name>
      </subject>

      Companies have multiple identifiers - could I then express these as follow, instead of my previous attempt:

      <subject type="cpnat:organisation" qcode=" foo:companyA">
      <name role="nrole:short">ANG</name>
      <name role="nrole:full">Angloplats</name>
      <sameAs qcode="isin:ZAE123"/>
      <sameAs qcode="tidm:abc"/>
      </subject>

      But how do I go about adding additional properties for CompanyA besides the name? Examples of these properties, include:

      - CountryOfRegistration (e.g. ZA, UK, US using ISO codes)
      - ListingStatus (listed, unlisted)

      Would I need to create a type of relationship which expresses these attributes, e.g.?

      <subject type="cpnat:organisation" qcode=" foo:companyA">
      <name role="nrole:short">ANG</name>
      <name role="nrole:full">Angloplats</name>
      <sameAs qcode="isin:ZAE123"/>
      <sameAs qcode="tidm:abc"/>
      <facet rel="crel:IsRegisteredIn" qcode="iso3166-1a2:ZA"/>
      <facet rel="crel:HasListingStatusOf" qcode="lstatus:listed"/>
      </subject>

      In a nutshell, I'm trying to express the follow XML in NewsML format:
      An regulatory announcement has:

      1. An Announcement number, e.g. 0001A;
      2. The announcement has "Primary Company" responsible for the actual
      announcement, E.G. ANG
      2.1. The primary company has few identifiers and attributes which
      are required by end-users, e.g. ISIN, MIC and TIDMs
      2.2. The primary company has other attributes, e.g. whether it is
      a listed company of not.
      3. The announcement may relate to other companies (called Related Companies") this happens often when the announcement is around mergers and acquisitions.
      3.1 The "Related Company or companies" have like the primary
      company, their set of identifies and attributes that need to be
      expressed.

      I think I've got most of the concepts of how to translate this to NewsML, except on how to assign a number of attributes and identifiers to acompany/organisation, besides the IPTC default of <name>.

      Legacy XML:
      ----------
      <sens:AnnoucementMetadata>
      <sens:AnnouncementNumber>001A</sens:AnnouncementNumber/>
      <sens:PrimaryCompany>
      <sens:CompanyType>Listed</sens:CompanyType>
      <sens:CompanyShortName>SAB</sens:CompanyShortName>
      <sens:ISIN>ZAE1234567891011</sens:ISIN>
      <sens:MIC>XJSE</sens:MIC>
      <sens:CountryOfRegistrer>ZA</sens:CountryOfRegistrer>
      </sens:PrimaryCompany>
      <sens:RelatedCompany>
      <sens:RelationshipID>1</sens:RelationshipID>
      <sens:CompanyType>listed</sens:CompanyType>
      <sens:CompanyShortName>ABC</sens:CompanyShortName>
      <sens:ISIN>ZAE123456789</sens:ISIN>
      <sens:MIC>XJSE</sens:MIC>
      <sens:CountryOfRegistrer>ZA</sens:CountryOfRegistrer>
      </sens:RelatedCompany>
      <sens:RelatedCompany>
      <sens:RelationshipID>2</sens:RelationshipID>
      <sens:CompanyType>unlisted</sens:CompanyType>
      <sens:CompanyShortName>XYZ</sens:CompanyShortName>
      <sens:ISIN>ZAE123456789</sens:ISIN>
      <sens:MIC>XJSE</sens:MIC>
      <sens:CountryOfRegister>US</sens:CountryOfRegister>
      </sens:RelatedCompany>
      </sens:AnnoucementMetadata>

      Thanks for the assistance - much appreciated!
      Regards
      Lino

      --- In newsml-g2@yahoogroups.com, "Kelvin Holland" <kholland@...> wrote:

      >
      > Hi Lino,
      >
      > I think the information you are expressing is more properly
      <subject>. The
      > property is "about" the company. You are likely to want to
      search on
      > <subject> rather than <infoSource>, which would return a
      rather narrow set
      > of results. (But of course if it is important to indicate the Information
      > Source AS WELL, you can)
      >
      > <facet> describes an "inherent characteristic" of the
      concept; ISIN and TIDM
      > are identifiers, not facets. (and by the way, note when using
      <facet> that
      > @relation is a mandatory property; <facet> represents an RDF triple
      "subject
      > predicate object".
      >
      > As mentioned above, the subject's @qcode should be an identifier for the
      > company, not its role in the transaction, so your subject block would look
      > something like:
      >
      > <subject type="cpnat:organisation" qcode="
      foo:companyA">
      > <name role="nrole:short">ANG</name>
      > <name role="nrole:full">Angloplats</name>
      > ...
      > </subject>
      >
      > or
      >
      > <subject type="cpnat:organisation" literal="
      companyA">
      > <name role="nrole:short">ANG</name>
      > <name role="nrole:full">Angloplats</name>
      > ...
      > </subject>
      >
      > Noting what you said about the volatility of ISIN codes, I think you need
      to
      > reflect on how your customers are expected to be able to unambiguously
      > identify your subjects now and in the future. I'm not an ISIN expert, but
      > from a G2 perspective It does not matter if new ISIN codes are created as
      > long as existing ISIN codes are not deleted or re-used to identify some
      > other company. Is this how ISINs work?
      >
      > On representing the transaction, again, use <subject>, and then you
      can use
      > inherent characteristics of the transaction - facets - to indicate its
      > properties:
      >
      > <subject type="cpnat:abstract"
      qcode="transtype:acquisition">
      > <name>Company Acquisition</name>
      > <facet relation="trel:acquirer"
      qcode="foo:companyA" />
      > <facet relation="trel:acquiree" qcode="foo:companyB"
      />
      > <facet relation="mrel:acqtype" qcode="bar:agreed"
      />
      > </subject>
      >
      >
      > Regards
      >
      > Kelvin
      >
      >
      > Kelvin Holland
      > Point House Media Ltd
      >
      > Editor, G2 Implementation Guidelines
      >
      > From: newsml-g2@yahoogroups.com
      [mailto:newsml-g2@yahoogroups.com] On Behalf
      > Of lino_sg
      > Sent: 24 May 2010 15:17
      > To: newsml-g2@yahoogroups.com
      > Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news content
      >
      >
      > Hi,
      >
      > Thanks to everyone that has provided feedback. I'm inclined now on using
      the
      > "ISIN:133", "TIDM:999" approach to define the concept
      around these codes.
      >
      > What I'm still undecided about is how to link this to the actual company
      buy
      > using 'facet', 'relation' or even 'role' attributes.
      >
      > In a regulatory announcement, a company may play a primary role( which I'm
      > think of defining as: "arole:CompanyPrimary") or a related role
      > ("arole:relatedcompany") e.g. by being a counter-party of a
      takeover.
      >
      > Only the primary company will actually create the regulatory announcement,
      > but the other other companies involved may have contributed to the source
      of
      > information and certainly are the subject of the announcement.
      >
      > What I'm undecided about is on how best to reflect the above semantic
      > information, e.g. as <infoSource> or <subject> items, e.g.:
      >
      > <infoSource type="cpnat:organisation"
      role="arole:primarycompany">
      > <name role="nrole:short">ANG</name>
      > <name role="nrole:full">Angloplats</name>
      > <facet qcode="isin:123"/>
      > <facet qcode="tidm:999"/>
      > </infoSource>
      >
      > or
      >
      > <subject type="cpnat:organisation"
      qcode="arole:primarycompany">
      > <name role="nrole:short">ANG</name>
      > <name role="nrole:full">Angloplats</name>
      > <facet qcode="isin:123"/>
      > <facet qcode="tidm:999"/>
      > </infoSource>
      >
      > What do you think reflects the role a company may play in an announcement?
      > Also note that I planning to using <subject> to define one of a
      finite set
      > of announcement types, e.g.:
      >
      > <subject type="cpnat:abstract"
      qcode="atype:acquisition" />
      >
      > Thanks and regards,
      > Lino
      >
      > --- In newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com> ,
      > <misha.wolf@> wrote:
      > >
      > > I suggest you use:
      > >
      > > qcode="isin:123"
      > >
      > > Regards,
      > > Misha
      > >
      > >
      > > -----Original Message-----
      > > From: newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com>
      > [mailto:newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com> ] On
      > Behalf Of lino_sg
      > > Sent: 24 May 2010 13:32
      > > To: newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com>
      > > Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news
      content
      > >
      > > Hi Misha,
      > >
      > > If I then use the ISIN schema below with an alias of ISIN
      > >
      > > Scheme URI: http://cvx.iptc.org/isin/
      > >
      > > how would define the qcodes as ISIN is open ended?
      > >
      > > Could I use a previous suggest from Kostas and do the following?
      > >
      > > <infoSource type="cpnat:organisation"
      role="comprole:PrimaryCompany">
      > > <name role="nrole:short">ANG</name>
      > > <name role="nrole:full">Angloplats</name>
      > > <facet qcode="isin:value" literal="123"/>
      > > <facet qcode="tidm:value" literal="999"/>
      > > </infoSource>
      > >
      > > Regards
      > > Lino
      > > --- In newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com> ,
      > <misha.wolf@> wrote:
      > > >
      > > > It is not a requirement that a KnowledgeItem exist for every
      taxonomy.
      > > >
      > > >
      > > >
      > > > The IPTC has put a lot of thought into how to handle taxonomies
      managed
      > by bodies which are not News Providers. For details, see:
      > > >
      > > >
      > > >
      > > > http://www.iptc.org/cms/site/index.html?channel=CH0215
      > <http://www.iptc.org/cms/site/index.html?channel=CH0215>
      > > >
      > > >
      > > >
      > > > and:
      > > >
      > > >
      > > >
      > > > http://cvx.iptc.org/ <
      href="http://cvx.iptc.org/">http://cvx.iptc.org/>
      > > >
      > > >
      > > >
      > > > ISINs are specifically addressed there. The scheme URI defined
      by the
      > IPTC for ISINs is:
      > > >
      > > >
      > > >
      > > > http://cvx.iptc.org/isin/
      > > >
      > > >
      > > >
      > > > Regards,
      > > >
      > > > Misha
      > > >
      > > >
      > > >
      > > >
      > > >
      > > > -----Original Message-----
      > > > From: newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com>
      > [mailto:newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com> ] On
      > Behalf Of lino_sg
      > > > Sent: 21 May 2010 09:55
      > > > To: newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com>
      > > > Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news
      content
      > > >
      > > >
      > > >
      > > > Hi Kostas,
      > > >
      > > >
      > > >
      > > > Thanks for your feedback.
      > > >
      > > >
      > > >
      > > > 1) ISIN codes change virtually daily as new companies are listed
      and
      > delisted. The volatility of ISINs is perhaps comparable to ISBN. If this
      is
      > the case, wouldn't the proposed solution be cumbersome as the controlled
      > vocabularies would be changing daily, especially if I had to maintain it
      via
      > a G2 CV? Perhaps I'm assuming that "ISIN:somecode" would need to
      be resolved
      > by a G2 KnowledgeItem? Can I still specify the ISIN as
      "ISIN:somecode" as
      > you've suggested and let the user resolve the code using local databases
      of
      > received ISIN codes from the exchanges feeds, or some other issuing
      > authority?
      > > >
      > > >
      > > >
      > > > 2) If I have a long list of attributes about the company in
      addition to
      > the ones already mentioned (e.g. if I open a record on company Angloplat
      in
      > my reference database, I get a least 20 other attributes or properties of
      > the company), what is the recommended way of exposing this metadata
      without
      > having to constantly maintain a G2 Knowledge Item? How to a take your
      > previous suggestion to include company metadata items (n1, n2, n3...n20)?
      > > >
      > > >
      > > >
      > > > Thanks for you feedback
      > > >
      > > > Regards
      > > >
      > > > Lino
      > > >
      > > >
      > > >
      > > > --- In newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com> ,
      > Kostas Konstantoulis <k.konstantoulis@> wrote:
      > > >
      > > > >
      > > >
      > > > > Hi,
      > > >
      > > > >
      > > >
      > > > > A 5th option is to use ISIN codes as a controlled
      vocabulary as well
      > as any other code system:
      > > >
      > > > >
      > > >
      > > > > <newsItem>
      > > >
      > > > > <contentMeta>
      > > >
      > > > > <subject type="cpnat:organisation"
      qcode="isin:10F4">
      > > >
      > > > > <name
      role="nrol:short">AngloPlats</name>
      > > >
      > > > > <sameas type="cpnat:organisation"
      qcode="nyse:xyz">
      > > >
      > > > > <sameas type="cpnat:organisation"
      qcode="bloomberg:abc">
      > > >
      > > > > <sameas type="cpnat:organisation"
      qcode="reuters:qwerty">
      > > >
      > > > > </subject>
      > > >
      > > > > </contentMeta>
      > > >
      > > > > </newsItem>
      > > >
      > > > >
      > > >
      > > > > Kostas
      > > >
      > > > > ________________________________
      > > >
      > > > > From: newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com>
      > [mailto:newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com> ] On
      > Behalf Of lino_sg
      > > >
      > > > > Sent: Friday, May 21, 2010 2:31 AM
      > > >
      > > > > To: newsml-g2@yahoogroups.com
      <mailto:newsml-g2%40yahoogroups.com>
      > > >
      > > > > Subject: [newsml-g2] How to 'Lift' metadata out of the news
      content
      > > >
      > > > >
      > > >
      > > > >
      > > >
      > > > >
      > > >
      > > > > Hi,
      > > >
      > > > >
      > > >
      > > > > I have the following scenario:
      > > >
      > > > >
      > > >
      > > > > 1. A Regulatory Release by a Company which has a ShortName
      of
      > 'AngloPlats' and which is Listed on the Johannesburg Stock Exchange(JSE)
      on
      > amount of Dividends the company will be paying at year-end.
      > > >
      > > > >
      > > >
      > > > > 2. Issue: What would be the best practice to 'lift' some of
      the
      > metadata around the companies involved in the announcement, such as
      company
      > 'ShortName' and ISIN (International Stock Identification Number)?
      > > >
      > > > >
      > > >
      > > > > 3. I believe it would NOT be feasible to have controlled
      vocabularies
      > for company ShortNames or ISINs as the list of companies listed in the
      stock
      > exchange varies too much too much over time.
      > > >
      > > > >
      > > >
      > > > > I have come up with several options below of doing this,
      e.g. using
      > 'facet', 'relations' and extensions using a (jse:) namespace. Do you have
      > any recommendations, best practices or advise on this topic?
      > > >
      > > > >
      > > >
      > > > > Thanks
      > > >
      > > > > Lino
      > > >
      > > > >
      > > >
      > > > > (...)
      > > >
      > > > > <contentMeta>
      > > >
      > > > >
      > > >
      > > > > <--.... other metadata -->
      > > >
      > > > >
      > > >
      > > > > <subject type="cpnat:abstract"
      code="AnnounceType:DivPayment"/>
      > > >
      > > > >
      > > >
      > > > > <-- Option 1 -- >
      > > >
      > > > > <subject type="cpnat:organisation"
      literal="ANG">
      > > >
      > > > > <facet rel="relation:ShortName"
      literal="Angloplats"/>
      > > >
      > > > > <facet rel="relation:ISIN"
      literal="10F4"/>
      > > >
      > > > > </subject>
      > > >
      > > > >
      > > >
      > > > > <-- Option 2 -->
      > > >
      > > > > <subject type="cpnat:organisation"
      literal="ANG">
      > > >
      > > > > <related rel="relation:ShortName"
      literal="Angloplats"/>
      > > >
      > > > > <related rel="relation:ISIN"
      literal="10F4"/>
      > > >
      > > > > </subject>
      > > >
      > > > >
      > > >
      > > > > <-- Option 3 -->
      > > >
      > > > > <subject type="cpnat:organisation"
      literal="ANG">
      > > >
      > > > > <narrower rel="relation:ShortName"
      literal="Angloplats"/>
      > > >
      > > > > <narrower rel="relation:ISIN"
      literal="10F4"/>
      > > >
      > > > > </subject>
      > > >
      > > > >
      > > >
      > > > > <-- Option 4: Using a JSE namespace extension ->
      > > >
      > > > > <jse:company>
      > > >
      > > > > <jse:AlphaCode>ANG</jse:AlphaCode>
      > > >
      > > > > <jse:ShortName>Angloplats</jse:ShortName>
      > > >
      > > > > <jse:ISIN>JSE</jse:ISIN>
      > > >
      > > > > </jse:company>
      > > >
      > > > >
      > > >
      > > > > <contentMeta>
      > > >
      > > > > (...)
      > > >
      > > > >
      > > >
      > > >
      > > >
      > > >
      > > >
      > > >
      > > >
      > > >
      > > >
      > > > ------------------------------------
      > > >
      > > >
      >

      (Message over 64 KB, truncated)
    • lino_sg
      Hi Kelvin, Thanks for the suggestions. I ll be away for the next 3 weeks (took some leave to watch some world cup games! Go Portugal and SA!) I ll pick up this
      Message 2 of 13 , Jun 8, 2010
      • 0 Attachment
        Hi Kelvin,

        Thanks for the suggestions. I'll be away for the next 3 weeks (took some leave to watch some world cup games! Go Portugal and SA!) I'll pick up this thread when I'm back :)

        Regards
        Lino

        --- In newsml-g2@yahoogroups.com, "Kelvin Holland" <kholland@...> wrote:
        >
        > Hi Lino,
        >
        > Sorry for the delay in replying.
        >
        > First, ERRATUM. I was mistaken in telling you that @rel is a mandatory
        > property of <facet>. It is optional and if not present, the implied
        > (default) relationship is "is a".
        >
        > Yes, you can add the additional properties of the concept as facets in the
        > way you have outlined below (at Power Conformance Level), OR,
        >
        > If the information must be carried inline, it may be helpful to place a
        > snapshot of the information about the concept in a dedicated structure, the
        > <assert> wrapper, that can be referenced by all properties in the document
        > that use this concept ID.
        >
        > The advantage of using <assert> is that it can hold all of the properties
        > your customers need to know about the companies in one place, but you can
        > refer to this information from multiple places in the document. So:
        >
        > <newsItem>
        > ...
        > <contentMeta>
        > ...
        > <infoSource qcode="foo:companyA" />
        > ...
        > <subject type="cpnat:organisation" qcode="
        > foo:companyA" />
        > <subject type="cpnat:organisation" qcode="
        > foo:companyB" />
        > <subject type="cpnat:abstract" qcode="transtype:acquisition">
        > <name>Company Acquisition</name>
        > <facet rel="trel:acquirer" qcode="foo:companyA" />
        > <facet rel="trel:acquiree" qcode="foo:companyB" />
        > <facet rel="mrel:acqtype" qcode="bar:agreed" />
        > </subject>
        > ...
        > </contentMeta>
        > ...
        > <assert qcode="foo:companyA">
        > ...
        > <name role="nrole:short">ANG</name>
        > <name role="nrole:full">Angloplats</name>
        > <sameAs qcode="isin:ZAE123"/>
        > <sameAs qcode="tidm:abc"/>
        > <facet rel="crel:IsRegisteredIn" qcode="iso3166-1a2:ZA"/>
        > <facet rel="crel:HasListingStatusOf" qcode="lstatus:listed"/>
        > ...
        > </assert>
        > <assert qcode="foo:companyB">
        > ...
        > </assert>
        >
        > Take a look at Chapter 18 of the Guidelines - this provides a full
        > explanation and more examples, also some advice on using local Ids
        > (@literal), which you may find useful.
        >
        > Anyone else care to jump in here? I'm pondering the Primary Company, Related
        > Company issue and how best to express that.
        >
        > Lino, thanks for sharing your example - it's an interesting use case.
        >
        > Regards
        >
        > Kelvin
        >
        >
        > Kelvin Holland
        > Point House Media Ltd
        > 01825 743340
        > 07966 448437
        >
        > From: newsml-g2@yahoogroups.com [mailto:newsml-g2@yahoogroups.com] On Behalf
        > Of lino_sg
        > Sent: 26 May 2010 18:15
        > To: newsml-g2@yahoogroups.com
        > Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news content
        >
        >
        > Hi Kelvin,
        >
        > Thanks for the feedback and clarifications on some of the finer details.
        > Kelvin, if I take your definition for a company:
        >
        > <subject type="cpnat:organisation" qcode=" foo:companyA">
        > <name role="nrole:short">ANG</name>
        > <name role="nrole:full">Angloplats</name>
        > </subject>
        >
        > Companies have multiple identifiers - could I then express these as follow,
        > instead of my previous attempt:
        >
        > <subject type="cpnat:organisation" qcode=" foo:companyA">
        > <name role="nrole:short">ANG</name>
        > <name role="nrole:full">Angloplats</name>
        > <sameAs qcode="isin:ZAE123"/>
        > <sameAs qcode="tidm:abc"/>
        > </subject>
        >
        > But how do I go about adding additional properties for CompanyA besides the
        > name? Examples of these properties, include:
        >
        > - CountryOfRegistration (e.g. ZA, UK, US using ISO codes)
        > - ListingStatus (listed, unlisted)
        >
        > Would I need to create a type of relationship which expresses these
        > attributes, e.g.?
        >
        > <subject type="cpnat:organisation" qcode=" foo:companyA">
        > <name role="nrole:short">ANG</name>
        > <name role="nrole:full">Angloplats</name>
        > <sameAs qcode="isin:ZAE123"/>
        > <sameAs qcode="tidm:abc"/>
        > <facet rel="crel:IsRegisteredIn" qcode="iso3166-1a2:ZA"/>
        > <facet rel="crel:HasListingStatusOf" qcode="lstatus:listed"/>
        > </subject>
        >
        > In a nutshell, I'm trying to express the follow XML in NewsML format:
        > An regulatory announcement has:
        >
        > 1. An Announcement number, e.g. 0001A;
        > 2. The announcement has "Primary Company" responsible for the actual
        > announcement, E.G. ANG
        > 2.1. The primary company has few identifiers and attributes which
        > are required by end-users, e.g. ISIN, MIC and TIDMs
        > 2.2. The primary company has other attributes, e.g. whether it is
        > a listed company of not.
        > 3. The announcement may relate to other companies (called Related
        > Companies") this happens often when the announcement is around mergers and
        > acquisitions.
        > 3.1 The "Related Company or companies" have like the primary
        > company, their set of identifies and attributes that need to be
        > expressed.
        >
        > I think I've got most of the concepts of how to translate this to NewsML,
        > except on how to assign a number of attributes and identifiers to
        > acompany/organisation, besides the IPTC default of <name>.
        >
        > Legacy XML:
        > ----------
        > <sens:AnnoucementMetadata>
        > <sens:AnnouncementNumber>001A</sens:AnnouncementNumber/>
        > <sens:PrimaryCompany>
        > <sens:CompanyType>Listed</sens:CompanyType>
        > <sens:CompanyShortName>SAB</sens:CompanyShortName>
        > <sens:ISIN>ZAE1234567891011</sens:ISIN>
        > <sens:MIC>XJSE</sens:MIC>
        > <sens:CountryOfRegistrer>ZA</sens:CountryOfRegistrer>
        > </sens:PrimaryCompany>
        > <sens:RelatedCompany>
        > <sens:RelationshipID>1</sens:RelationshipID>
        > <sens:CompanyType>listed</sens:CompanyType>
        > <sens:CompanyShortName>ABC</sens:CompanyShortName>
        > <sens:ISIN>ZAE123456789</sens:ISIN>
        > <sens:MIC>XJSE</sens:MIC>
        > <sens:CountryOfRegistrer>ZA</sens:CountryOfRegistrer>
        > </sens:RelatedCompany>
        > <sens:RelatedCompany>
        > <sens:RelationshipID>2</sens:RelationshipID>
        > <sens:CompanyType>unlisted</sens:CompanyType>
        > <sens:CompanyShortName>XYZ</sens:CompanyShortName>
        > <sens:ISIN>ZAE123456789</sens:ISIN>
        > <sens:MIC>XJSE</sens:MIC>
        > <sens:CountryOfRegister>US</sens:CountryOfRegister>
        > </sens:RelatedCompany>
        > </sens:AnnoucementMetadata>
        >
        > Thanks for the assistance - much appreciated!
        > Regards
        > Lino
        >
        > --- In newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com> ,
        > "Kelvin Holland" <kholland@> wrote:
        > >
        > > Hi Lino,
        > >
        > > I think the information you are expressing is more properly <subject>. The
        > > property is "about" the company. You are likely to want to search on
        > > <subject> rather than <infoSource>, which would return a rather narrow set
        > > of results. (But of course if it is important to indicate the Information
        > > Source AS WELL, you can)
        > >
        > > <facet> describes an "inherent characteristic" of the concept; ISIN and
        > TIDM
        > > are identifiers, not facets. (and by the way, note when using <facet> that
        > > @relation is a mandatory property; <facet> represents an RDF triple
        > "subject
        > > predicate object".
        > >
        > > As mentioned above, the subject's @qcode should be an identifier for the
        > > company, not its role in the transaction, so your subject block would look
        > > something like:
        > >
        > > <subject type="cpnat:organisation" qcode=" foo:companyA">
        > > <name role="nrole:short">ANG</name>
        > > <name role="nrole:full">Angloplats</name>
        > > ...
        > > </subject>
        > >
        > > or
        > >
        > > <subject type="cpnat:organisation" literal=" companyA">
        > > <name role="nrole:short">ANG</name>
        > > <name role="nrole:full">Angloplats</name>
        > > ...
        > > </subject>
        > >
        > > Noting what you said about the volatility of ISIN codes, I think you need
        > to
        > > reflect on how your customers are expected to be able to unambiguously
        > > identify your subjects now and in the future. I'm not an ISIN expert, but
        > > from a G2 perspective It does not matter if new ISIN codes are created as
        > > long as existing ISIN codes are not deleted or re-used to identify some
        > > other company. Is this how ISINs work?
        > >
        > > On representing the transaction, again, use <subject>, and then you can
        > use
        > > inherent characteristics of the transaction - facets - to indicate its
        > > properties:
        > >
        > > <subject type="cpnat:abstract" qcode="transtype:acquisition">
        > > <name>Company Acquisition</name>
        > > <facet relation="trel:acquirer" qcode="foo:companyA" />
        > > <facet relation="trel:acquiree" qcode="foo:companyB" />
        > > <facet relation="mrel:acqtype" qcode="bar:agreed" />
        > > </subject>
        > >
        > >
        > > Regards
        > >
        > > Kelvin
        > >
        > >
        > > Kelvin Holland
        > > Point House Media Ltd
        > >
        > > Editor, G2 Implementation Guidelines
        > >
        > > From: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > [mailto:newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com> ] On
        > Behalf
        > > Of lino_sg
        > > Sent: 24 May 2010 15:17
        > > To: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > > Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news content
        > >
        > >
        > > Hi,
        > >
        > > Thanks to everyone that has provided feedback. I'm inclined now on using
        > the
        > > "ISIN:133", "TIDM:999" approach to define the concept around these codes.
        > >
        > > What I'm still undecided about is how to link this to the actual company
        > buy
        > > using 'facet', 'relation' or even 'role' attributes.
        > >
        > > In a regulatory announcement, a company may play a primary role( which I'm
        > > think of defining as: "arole:CompanyPrimary") or a related role
        > > ("arole:relatedcompany") e.g. by being a counter-party of a takeover.
        > >
        > > Only the primary company will actually create the regulatory announcement,
        > > but the other other companies involved may have contributed to the source
        > of
        > > information and certainly are the subject of the announcement.
        > >
        > > What I'm undecided about is on how best to reflect the above semantic
        > > information, e.g. as <infoSource> or <subject> items, e.g.:
        > >
        > > <infoSource type="cpnat:organisation" role="arole:primarycompany">
        > > <name role="nrole:short">ANG</name>
        > > <name role="nrole:full">Angloplats</name>
        > > <facet qcode="isin:123"/>
        > > <facet qcode="tidm:999"/>
        > > </infoSource>
        > >
        > > or
        > >
        > > <subject type="cpnat:organisation" qcode="arole:primarycompany">
        > > <name role="nrole:short">ANG</name>
        > > <name role="nrole:full">Angloplats</name>
        > > <facet qcode="isin:123"/>
        > > <facet qcode="tidm:999"/>
        > > </infoSource>
        > >
        > > What do you think reflects the role a company may play in an announcement?
        > > Also note that I planning to using <subject> to define one of a finite set
        > > of announcement types, e.g.:
        > >
        > > <subject type="cpnat:abstract" qcode="atype:acquisition" />
        > >
        > > Thanks and regards,
        > > Lino
        > >
        > > --- In newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com> ,
        > > <misha.wolf@> wrote:
        > > >
        > > > I suggest you use:
        > > >
        > > > qcode="isin:123"
        > > >
        > > > Regards,
        > > > Misha
        > > >
        > > >
        > > > -----Original Message-----
        > > > From: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com>
        > > [mailto:newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com> ] On
        > > Behalf Of lino_sg
        > > > Sent: 24 May 2010 13:32
        > > > To: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com>
        > > > Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news content
        > > >
        > > > Hi Misha,
        > > >
        > > > If I then use the ISIN schema below with an alias of ISIN
        > > >
        > > > Scheme URI: http://cvx.iptc.org/isin/
        > > >
        > > > how would define the qcodes as ISIN is open ended?
        > > >
        > > > Could I use a previous suggest from Kostas and do the following?
        > > >
        > > > <infoSource type="cpnat:organisation" role="comprole:PrimaryCompany">
        > > > <name role="nrole:short">ANG</name>
        > > > <name role="nrole:full">Angloplats</name>
        > > > <facet qcode="isin:value" literal="123"/>
        > > > <facet qcode="tidm:value" literal="999"/>
        > > > </infoSource>
        > > >
        > > > Regards
        > > > Lino
        > > > --- In newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com> ,
        > > <misha.wolf@> wrote:
        > > > >
        > > > > It is not a requirement that a KnowledgeItem exist for every taxonomy.
        > > > >
        > > > >
        > > > >
        > > > > The IPTC has put a lot of thought into how to handle taxonomies
        > managed
        > > by bodies which are not News Providers. For details, see:
        > > > >
        > > > >
        > > > >
        > > > > http://www.iptc.org/cms/site/index.html?channel=CH0215
        > > <http://www.iptc.org/cms/site/index.html?channel=CH0215>
        > > > >
        > > > >
        > > > >
        > > > > and:
        > > > >
        > > > >
        > > > >
        > > > > http://cvx.iptc.org/ <http://cvx.iptc.org/>
        > > > >
        > > > >
        > > > >
        > > > > ISINs are specifically addressed there. The scheme URI defined by the
        > > IPTC for ISINs is:
        > > > >
        > > > >
        > > > >
        > > > > http://cvx.iptc.org/isin/
        > > > >
        > > > >
        > > > >
        > > > > Regards,
        > > > >
        > > > > Misha
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > > -----Original Message-----
        > > > > From: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com>
        > > [mailto:newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com> ] On
        > > Behalf Of lino_sg
        > > > > Sent: 21 May 2010 09:55
        > > > > To: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com>
        > > > > Subject: [newsml-g2] Re: How to 'Lift' metadata out of the news
        > content
        > > > >
        > > > >
        > > > >
        > > > > Hi Kostas,
        > > > >
        > > > >
        > > > >
        > > > > Thanks for your feedback.
        > > > >
        > > > >
        > > > >
        > > > > 1) ISIN codes change virtually daily as new companies are listed and
        > > delisted. The volatility of ISINs is perhaps comparable to ISBN. If this
        > is
        > > the case, wouldn't the proposed solution be cumbersome as the controlled
        > > vocabularies would be changing daily, especially if I had to maintain it
        > via
        > > a G2 CV? Perhaps I'm assuming that "ISIN:somecode" would need to be
        > resolved
        > > by a G2 KnowledgeItem? Can I still specify the ISIN as "ISIN:somecode" as
        > > you've suggested and let the user resolve the code using local databases
        > of
        > > received ISIN codes from the exchanges feeds, or some other issuing
        > > authority?
        > > > >
        > > > >
        > > > >
        > > > > 2) If I have a long list of attributes about the company in addition
        > to
        > > the ones already mentioned (e.g. if I open a record on company Angloplat
        > in
        > > my reference database, I get a least 20 other attributes or properties of
        > > the company), what is the recommended way of exposing this metadata
        > without
        > > having to constantly maintain a G2 Knowledge Item? How to a take your
        > > previous suggestion to include company metadata items (n1, n2, n3...n20)?
        > > > >
        > > > >
        > > > >
        > > > > Thanks for you feedback
        > > > >
        > > > > Regards
        > > > >
        > > > > Lino
        > > > >
        > > > >
        > > > >
        > > > > --- In newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com> ,
        > > Kostas Konstantoulis <k.konstantoulis@> wrote:
        > > > >
        > > > > >
        > > > >
        > > > > > Hi,
        > > > >
        > > > > >
        > > > >
        > > > > > A 5th option is to use ISIN codes as a controlled vocabulary as well
        > > as any other code system:
        > > > >
        > > > > >
        > > > >
        > > > > > <newsItem>
        > > > >
        > > > > > <contentMeta>
        > > > >
        > > > > > <subject type="cpnat:organisation" qcode="isin:10F4">
        > > > >
        > > > > > <name role="nrol:short">AngloPlats</name>
        > > > >
        > > > > > <sameas type="cpnat:organisation" qcode="nyse:xyz">
        > > > >
        > > > > > <sameas type="cpnat:organisation" qcode="bloomberg:abc">
        > > > >
        > > > > > <sameas type="cpnat:organisation" qcode="reuters:qwerty">
        > > > >
        > > > > > </subject>
        > > > >
        > > > > > </contentMeta>
        > > > >
        > > > > > </newsItem>
        > > > >
        > > > > >
        > > > >
        > > > > > Kostas
        > > > >
        > > > > > ________________________________
        > > > >
        > > > > > From: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com>
        > > [mailto:newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com> ] On
        > > Behalf Of lino_sg
        > > > >
        > > > > > Sent: Friday, May 21, 2010 2:31 AM
        > > > >
        > > > > > To: newsml-g2@yahoogroups.com <mailto:newsml-g2%40yahoogroups.com>
        > <mailto:newsml-g2%40yahoogroups.com>
        > > > >
        > > > > > Subject: [newsml-g2] How to 'Lift' metadata out of the news content
        > > > >
        > > > > >
        > > > >
        > > > > >
        > > > >
        > > > > >
        > > > >
        > > > > > Hi,
        > > > >
        > > > > >
        > > > >
        > > > > > I have the following scenario:
        > > > >
        > > > > >
        > > > >
        > > > > > 1. A Regulatory Release by a Company which has a ShortName of
        > > 'AngloPlats' and which is Listed on the Johannesburg Stock Exchange(JSE)
        > on
        > > amount of Dividends the company will be paying at year-end.
        > > > >
        > > > > >
        > > > >
        > > > > > 2. Issue: What would be the best practice to 'lift' some of the
        > > metadata around the companies involved in the announcement, such as
        > company
        > > 'ShortName' and ISIN (International Stock Identification Number)?
        > > > >
        > > > > >
        > > > >
        > > > > > 3. I believe it would NOT be feasible to have controlled
        > vocabularies
        > > for company ShortNames or ISINs as the list of companies listed in the
        > stock
        > > exchange varies too much too much over time.
        > > > >
        > > > > >
        > > > >
        > > > > > I have come up with several options below of doing this, e.g. using
        > > 'facet', 'relations' and extensions using a (jse:) namespace. Do you have
        > > any recommendations, best practices or advise on this topic?
        > > > >
        > > > > >
        > > > >
        > > > > > Thanks
        > > > >
        > > > > > Lino
        > > > >
        > > > > >
        > > > >
        > > > > > (...)
        > > > >
        > > > > > <contentMeta>
        > > > >
        > > > > >
        > > > >
        > > > > > <--.... other metadata -->
        > > > >
        > > > > >
        > > > >
        > > > > > <subject type="cpnat:abstract" code="AnnounceType:DivPayment"/>
        > > > >
        > > > > >
        > > > >
        > > > > > <-- Option 1 -- >
        > > > >
        > > > > > <subject type="cpnat:organisation" literal="ANG">
        > > > >
        > > > > > <facet rel="relation:ShortName" literal="Angloplats"/>
        > > > >
        > > > > > <facet rel="relation:ISIN" literal="10F4"/>
        > > > >
        > > > > > </subject>
        > > > >
        > > > > >
        > > > >
        > > > > > <-- Option 2 -->
        > > > >
        > > > > > <subject type="cpnat:organisation" literal="ANG">
        > > > >
        > > > > > <related rel="relation:ShortName" literal="Angloplats"/>
        > > > >
        > > > > > <related rel="relation:ISIN" literal="10F4"/>
        > > > >
        > > > > > </subject>
        > > > >
        > > > > >
        > > > >
        > > > > > <-- Option 3 -->
        > > > >
        > > > > > <subject type="cpnat:organisation" literal="ANG">
        > > > >
        > > > > > <narrower rel="relation:ShortName" literal="Angloplats"/>
        > > > >
        > > > > > <narrower rel="relation:ISIN" literal="10F4"/>
        > > > >
        > > > > > </subject>
        > > > >
        > > > > >
        > > > >
        > > > > > <-- Option 4: Using a JSE namespace extension ->
        > > > >
        > > > > > <jse:company>
        > > > >
        > > > > > <jse:AlphaCode>ANG</jse:AlphaCode>
        > > > >
        > > > > > <jse:ShortName>Angloplats</jse:ShortName>
        > > > >
        > > > > > <jse:ISIN>JSE</jse:ISIN>
        > > > >
        > > > > > </jse:company>
        > > > >
        > > > > >
        > > > >
        > > > > > <contentMeta>
        > > > >
        > > > > > (...)
        > > > >
        > > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > > ------------------------------------
        > > > >
        > > > >
        > > > >
        > > > > Any member of this IPTC moderated Yahoo group must comply with the
        > > Intellectual Property Policy of the IPTC, available at
        > > http://www.iptc.org/goto/ipp. Any posting is assumed to be submitted under
        > > the conditions of this IPTC IP Policy.
        > > > >
        > > > > Yahoo! Groups Links
        > > > >
        > > > >
        > > > >
        > > > > http://groups.yahoo.com/group/newsml-g2/
        > > > >
        > > > >
        > > > >
        > > > > Individual Email | Traditional
        > > > >
        > > > >
        > > > >
        > > > > http://groups.yahoo.com/group/newsml-g2/join
        > > > >
        > > > > (Yahoo! ID required)
        > > > >
        > > > >
        > > > >
        > > > > newsml-g2-digest@yahoogroups.com
        > <mailto:newsml-g2-digest%40yahoogroups.com>
        > > <mailto:newsml-g2-digest%40yahoogroups.com>
        > > > >
        > > > > newsml-g2-fullfeatured@yahoogroups.com
        > <mailto:newsml-g2-fullfeatured%40yahoogroups.com>
        > > <mailto:newsml-g2-fullfeatured%40yahoogroups.com>
        > > > >
        > > > >
        > > > >
        > > > > newsml-g2-unsubscribe@yahoogroups.com
        > <mailto:newsml-g2-unsubscribe%40yahoogroups.com>
        > > <mailto:newsml-g2-unsubscribe%40yahoogroups.com>
        > > > >
        > > > >
        > > > >
        > > > > http://docs.yahoo.com/info/terms/
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > > 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.
        > > > >
        > > >
        > > >
        > > >
        > > >
        > > > ------------------------------------
        > > >
        > > > Any member of this IPTC moderated Yahoo group must comply with the
        > > Intellectual Property Policy of the IPTC, available at
        > > http://www.iptc.org/goto/ipp. Any posting is assumed to be submitted under
        > > the conditions of this IPTC IP Policy.
        > > > 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.
        > > >
        > >
        >
      Your message has been successfully submitted and would be delivered to recipients shortly.