Semantic Exchange links the News Industry
News is at first sight just as a bunch of data providing some information. To drill into the semantics of it the news industry started to categorize and classify already in the 1990-ies, starting with very simple taxonomies which have been improved and extended to e.g. the IPTC Subject Codes of 1400 terms. At this starting time the domain of professional news was a quite isolated island which changed substantially with publishing news on the web and other digital formats. There was not only the taxonomy of the news industry relevant anymore but many others and users of news started to ask: “if a news item is classified with the IPTC Subject Code for Labour does this relate to a story classified with the term Work from another taxonomy?” This was the kick-off for discussing how to link vocabularies which are used for classifying news.
For this reason the IPTC members have launched in 2013 a project called Semantic Exchange which should result in a system which links classification vocabularies of different news/media companies.
In the scope of this project these topics are discussed:
Some background will be added to the topics in the presentation and first findings of the discussion will be shared.
- What exactly are such vocabularies: a simple list of terms or well selected named concepts with a reasonable structure – what efforts are needed to create and maintain them.
What is the value of such vocabularies:
Facet 1: do such vocabularies improve the usability of the news increasing their value this way?
Facet 2: are they real assets as creating such a vocabulary needs a lot of human resources?
- Should vocabularies be linked across company boundaries at all – as this includes linking my vocabulary to the vocabulary of my competitor.
- How should the linking be designed: all news provider specific vocabularies link to a single super-taxonomy or should they be linked directly.
- Who is responsible for setting relationships between terms of vocabularies of different owners – could this cause drawbacks?
- How well can relationships between terms be applied in a world with quite different cultures?
- What should a technical system providing the linking deliver: production level access or referential access to fill local caches. (Consider how many news items are spit out and read each second worldwide)