Back at my desk I looked at the Telegraph website. Then I went to the Guardian, Sky News and as if I hadn’t seen enough of the same stuff, the BBC news website. We thirst for information these days.
The BBC website, like all of them, had the item as its main news. Colonel Gadaffi “killed”: Latest updates. Below this were links to “Gadaffi’s Quixotic and brutal rule, The Muammar Gadaffi story and His Life in pictures.
It was only then that I realised I was looking at “semantic metadata” in action. This time last week I had never heard of semantic metadata or at least would have had no idea what it meant. That was before I saw a talk by John O’Donovan, Director, Technical Architecture and Development of the Press Association.
If you ever get the chance go and see one of his talks you should. John is a quietly spoken guy but he knows his stuff – he was one of the architects of the BBC iPlayer > enter semantic metadata stage left.
One of the problems that an organisation like the BBC has is the sheer volume of information being generated. 6,000 journalists, 1,200 of whom are writing for tinterweb. Semantic metadata provides information about content that allows it to be associated with other relevant content.
In the case of the BBC News website the application of semantic metadata allows synergistic articles to be placed close to the one being read – in the case of el coronel, the three links that provided additional information of interest to the average punter.
The beauty of the BBC machine is that this information is all gathered in the absence of any human involvement other than the filing of the initial story. If you think about it, and coming back to what got me going on this subject in the first place, we all want that information now, in real time. Involving people is going to slow things down.
This all demands a high degree of discipline of the BBC journos writing the stuff in the first place. O’Donovan mentioned that the BBC has millions of tapes with very little descriptive data. Just “rugby” for example with no hint of which rugby match. You might laugh but apparently the sport department is the worst offender and previously the descriptor might just have read “sport”. Tracing partiular incidents often meant someone watching a whole programme to find it.
So that’s semantic metadata. As we drop our needles in the giant haystack that is the world wide web we would do well to remember what it stands for – sanity and the saving of time in our increasingly busy and information hungry world.