As many online publishers, both large and small, now move to implement rNews 1.0, the IPTC’s new data model for embedding publishing metadata in web pages, some questions are coming up.
Most individual IPTC members represent their respective employers and I’m no exception: Working for the world’s largest photo agency, the question how to best mark up images is often smilingly passed along to me. Most questions are easy to answer and just require a look at the mapping table between the Photo Metadata standards IPTC Core and IPTC Extension to rNews.
But in particular the rNews properties about and mentions seem to warrant a bit of elaboration.
Let’s start with our official definitions:
rnews:about – “Indicates that the NewsItem is specifically about a concept.”
rnews:mentions – “Indicates that the NewsItem contains a reference to, but is not necessarily about a concept.”
So what does that mean?
If the NewsItem is a text/article, this usually makes sense to everyone. After all, a text is (ideally) clearly focused on a certain subject, so it’s about that subject. And in the process of writing about said subject, other facts, people, organizations etc. are mentioned (but they’re not the focus of that article).
Applying this paradigm to a picture doesn’t require too much of a mental leap. Of course, there are edge cases. (We could probably fill another five blog posts just discussing those and who knows, maybe I’ll do that some day down the road.)
But aside from edge cases, what is the focus of a given image? Who is the person or what is the thing, action etc. clearly visible in the image? That’s what you should mark up using the rnews:about property – it’s essentially saying that this is the core focus of the image. What is also visible but not the main focus of attention? That’s what the image mentions, of you will. Use rnews:mentions to mark up those concepts.
As I said above, it’s not a perfect analogy and there certainly are edge cases. But the usage I lay out here is the most straightforward application of the rNews data model to images and their metadata and therefore the preferred one. With this, I’m confident that publishers and rNews consumers alike can be expected to understand the semantics of about and mentions in the context of an image.
Thoughts? Objections? Ideas?
Leave a comment below or respond to @agebhard on Twitter.