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

1223RE: [newsml-g2] Relationship from concept to non G2 entities

Expand Messages
  • Michael Steidl (IPTC)
    Sep 12, 2011
    • 0 Attachment
      Hi Philippe

      I understand your concerns about the persistence of identifiers but this requirement cannot be cancelled that easily:

      - a news item of 12 September 2011 is categorized by a subject with the concept URI http://cvx.anewsprovider.com/locationsystemA/654321 which stands for the region Alsace in France.

      - on 1 October 2011 the LocationSystemA drops - for whatever reason - Alsace and assigns the code to Lorraine.

      - if a customer wants to search for news items about Alsace on 20 October what should this poor guy do? Using the code 654321 and constrain the search to items before 1 October? We can hope he knows about the date of the change.

      That's the reason for the G2 approach - and actually it is a Semantic Web approach - that a concept URI must never be reused. It may be "retired" = recommended not to use it any longer, e.g. for a dissolved company, but nothing else.

      What I don't understand about this location system is that locations usually do not disappear - except Atlantis, I know - so why are terms disappearing, the id is cancelled and re-applied to another location - but this only reflecting my pondering.

      A workaround could be to change the scheme URI when a code is re-assigned:

      URI from 1 January to 30 September (see example above): http://cvx.anewsprovider.com/locationsystemA20110101/
      http://cvx.anewsprovider.com/locationsystemA20110101/654321 = Alsace

      URI from 1 October: http://cvx.anewsprovider.com/locationsystemA20111001/
      http://cvx.anewsprovider.com/locationsystemA20111010/654321 = Lorraine

      ... but this would require that the sameAs relationships have to be extended for each of these changes.

      Michael

      > -----Original Message-----
      > From: newsml-g2@yahoogroups.com [mailto:newsml-g2@yahoogroups.com] On
      > Behalf Of Philippe Mougin
      > Sent: Monday, September 12, 2011 12:34 PM
      > To: newsml-g2@yahoogroups.com
      > Subject: Re: [newsml-g2] Relationship from concept to non G2 entities
      >
      > Thanks Misha and Michael.
      >
      > So, say I create a "locsysA" scheme as in Michael example, I would then
      > instruct receivers that if they get a concept URI that starts with
      > "http://cvx.anewsprovider.com/locationsystemA/" then the identifier in
      > system A can be computed by removing this prefix from the URI and
      > percent decoding what remains. Right?
      >
      > Still, I have a concern with this approach regarding G2 requirements
      > and I would welcome guidance on this:
      >
      > G2 makes several requirements governing schemes and concepts URI. In
      > particular, in 12.5.4.1 of the implementation guide (revision 3):
      > - Concepts MUST NOT be deleted from a Scheme [...]
      > - For the same reason, Concept IDs MUST NOT be re-cycled, i.e. the same
      > identifier used for a different concept.
      >
      > The issue is that my non-G2 systems A and B do not work like that:
      > sometimes they delete stuff and they even recycle identifiers. And
      > while they are internal systems in my company, they are not under my
      > control and I can't change their identifier management policies, even
      > if I was to ask politely.
      >
      > This means that the concept URI
      > http://cvx.anewsprovider.com/locationsystemA/654321 won't necessarily
      > identify the same concept (i.e., the same location) forever, which
      > seems to break the G2 requirements.
      >
      > This is probably an issue we don't have with ISO codes.
      >
      > So how would you approach the situation. Would you go ahead and
      > consider its ok in such instances to loosen those G2 rules?
      >
      > Philippe
      >
      > Le 9 sept. 2011 à 17:01, Michael Steidl (IPTC) a écrit :
      >
      > > Adding to what Misha said: look at the http://cvx.iptc.org/ page: it
      > provides scheme URIs for schemes which are a) not G2 compliant schemes
      > by their authority and b) outside the governance of the IPTC. Despite
      > of this we assign a scheme URI which points to a page which explains
      > how to resolve the code.
      > >
      > > Let's go through Philippe's example
      > >
      > > Add to a catalog.
      > > <scheme alias="locsysA"
      > uri="http://cvx.anewsprovider.com/locationsystemA/" />
      > > <scheme alias="locsysB"
      > uri="http://cvx.anewsprovider.com/locationsystemB/" />
      > >
      > > Create web pages for these URIs explaining what location repository
      > this is and that the code used for this scheme is the unique ID of a
      > location in this system.
      > >
      > > Add sameAs to your concept:
      > > <sameAs qcode="locsysA:654321" />
      > > <sameAs qcode="locsysB:569843" />
      > >
      > > I guess that's it.
      > >
      > > Michael
      > >
      > > > -----Original Message-----
      > > > From: newsml-g2@yahoogroups.com [mailto:newsml-g2@yahoogroups.com]
      > On
      > > > Behalf Of misha.wolf@...
      > > > Sent: Friday, September 09, 2011 12:46 PM
      > > > To: newsml-g2@yahoogroups.com
      > > > Subject: RE: [newsml-g2] Relationship from concept to non G2
      > entities
      > > >
      > > > Hi Philippe,
      > > >
      > > > It doesn't matter whether systems A and B are G2-compliant. We use
      > > > qcodes for everything, eg ISO country codes, ISO currency codes,
      > etc.
      > > > The relevant ISO standards say nothing about G2 but this doesn't
      > make
      > > > any difference.
      > > >
      > > > In "scheme:code", the "scheme" simply identifies the place that the
      > > > "code" is taken from.
      > > >
      > > > Regards,
      > > > Misha
      > > >
      > > >
      > > > -----Original Message-----
      > > > From: newsml-g2@yahoogroups.com [mailto:newsml-g2@yahoogroups.com]
      > On
      > > > Behalf Of Philippe Mougin
      > > > Sent: 09 September 2011 09:57
      > > > To: newsml-g2@yahoogroups.com
      > > > Subject: [newsml-g2] Relationship from concept to non G2 entities
      > > >
      > > > Hi,
      > > >
      > > > I have a G2-based system where I use a <concept> element to
      > represent a
      > > > particular location (the city of Paris). It happens that I also
      > have
      > > > two
      > > > other systems (A and B), non G2 based, where I manage locations.
      > These
      > > > systems have their own identifiers for the city of Paris. In system
      > A,
      > > > the identifier is "654321" and in system B one the identifier is
      > > > "569843".
      > > >
      > > > In my G2 <concept> I want to provide those identifiers as part of
      > the
      > > > concept description.
      > > >
      > > > At first glance, it seems that <sameAs> correctly express the
      > > > relationship I want to represent. I my <concept> I would have:
      > > >
      > > > <sameAs type="cpnat:geoArea" literal="654321">
      > > > <sameAs type="cpnat:geoArea" literal="569843">
      > > >
      > > > But of course this doesn't work because for any meaningful
      > processing I
      > > > would also need to know that the first one relates to system A and
      > the
      > > > second one relates to system B.
      > > >
      > > > I can't use qcode instead of literal as system A and system B
      > aren't
      > > > G2-based and consequently don't expose their identifiers through G2
      > > > QCodes. I could create such schemes myself but since my goal is to
      > > > communicate the identifiers of Paris in A and B (and not in newly
      > > > created ad hoc schemes) that would be self defeating.
      > > >
      > > > So how would you approach the problem ?
      > > >
      > > > Note : as a fallback I'm thinking about using <remoteInfo>
      > elements.
      > > > For
      > > > example :
      > > >
      > > > <remoteInfo rel="rels:equivalent-in-systemA" residref="654321">
      > > > <remoteInfo rel="rels:equivalent-in-systemB" residref="569843">
      > > >
      > > > Thanks,
      > > >
      > > > Philippe
      > > >
      > > >
      > > >
      > > > ------------------------------------
      > > >
      > > > 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.
      > > >
      > > >
      > > > ------------------------------------
      > > >
      > > > 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
      > > >
      > > >
      > > >
      > >
      > >
      >
      >
      >
      > ------------------------------------
      >
      > 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
      >
      >
      >
    • Show all 12 messages in this topic