At the recent Local Business Summit in Amsterdam I spoke about my work at the BBC on location services and mapping and tried to make the point that, as far as our users/consumers are concerned, the place name is all-powerful when it comes to identifying location.

Other means of identifying location (grid references, postcodes, lat/long coordinates) can be very useful for certain applications, particularly when dealing with single sets of data which associate themselves closely to a certain type of identifier, but are not really that useful when displaying location to a user.

The majority of us talk about our local area using the names of the town, road, river, pub, street, or even colloquial terms that would be completely un-familiar to someone who isn’t from the area.

Photo by Flickr user: Eric Kilby

The pagoda roundabout in Birmingham is actually the Holloway Circus roundabout but I’ve only ever heard people refer to it as the former.

The problem with place names is that they are very rarely unique and, unlike more formally defined identifiers, they do not fit nicely with the world of mathematics and software engineering which expects each component to be defined very specifically.

Gary Gale (@vicchi) recently posted an article on the perils of disambiguation which gets to the heart of the problem where geo-tagging is concerned.

A comment from John Fagan on this article suggests that “20% of UK places share the same name”.

Photo by Flickr user: Tim Green aka atoach

Gary also touched on the requirement to identify the concept of ‘place’ in the first, er, place.

For example, within a piece of text, does the word ‘bath’ refer to the place in the West Country or the item of furniture usually found in the bathroom?

We should be able to assume that if content is manually created then human knowledge will differentiate between various types of information and tag accordingly (assuming the tagging/meta-data process is manual).

We should also be able to assume that a user looking for content is very much aware of whether they are looking for ‘Bath’ the place or ‘bath’ the household object.

(An aside – luckily for me ‘Nuneaton’ is unique enough in location terms and in general language to make my latest website, the Nuneaton Guide, fairly easy to pull together using automated methods.)

But in reality, and in a web2.0 world, how many services exist that have this human editorial oversight from production right through to delivery?

At this point most people working in the web world will start talking about the semantic web, a greater understanding of meaning, and how the future web will be able to filter, aggregate, define, and ultimately solve problems such as this.

Web 3.0 from Kate Ray on Vimeo.

But what if the semantic web isn’t the answer?

This is too big an opportunity to leave to chance and I believe we need to start thinking now about how we are going to bring some order and some meaning to the definition and concept of ‘place’.

Geonames have made a good start and, in the UK at least, the free-ing up of Ordnance Survey data surely presents some possibilities.

But what next?

I don’t have the answer but I know that place names are the most powerful tool we have to deliver a sense of localness to our audience and turning the world wide web into a location aware service would be quite incredible.

At the very least we need a world where I can easily find stuff from the BBC about Morecambe the seaside town and, quite separately, content about the late, great comedian Eric Morecambe, without getting this page.