If you can’t link to it… does it exist ?

“If a tree falls in a forest and no one is around to hear it, does it make a sound?”

So goes the well-known philosophical thought experiment,  however rather than a discourse on observation and perception I’d like to hijack the experiment for an argument I have been making on and off for the last couple of years and which was  well summarised in a tweet summarising my point last month..

Does information published on the web which is not easily linkable actually exist ?

Well of course if I chose to publish my large spatial database of whatever using a Web Feature Service or some other application server the data actually exists, but as far as users of the web does it exist if I cannot find it using web search or more importantly as far as the way the web works cannot link to it?

This issue of making the so-called Deep Web more discoverable is still challenging , efforts such as the sitemap protocol have had only limited impact.

I would argue for the geospatial community in particular we need to take a more fundamental look at how we make information accessible and linkable on the web.  We need to start from the basic use case, common if you think about it but radical it would appear in the GIS world..

I need to let people link to each record in my spatial database and to share that link..

This actually requires perhaps a much more granular approach to making spatial data available, something that nearly got started with OS Mastermap but which for many issues was never fully implemented.

Rather than publishing online a database of railway station locations in the Netherlands and expecting a user to then query the database for  “Amsterdam Centraal Station”,  publish the database giving each record a URI so for example Amsterdam Centraal Station becomes; 

Now this is something I can paste into an email, tweet or even share on Facebook !

Kudos to the Dutch Kadaster for taking this approach and providing this example, Ordnance Survey you could do the same ?

This approach also results in such data becoming part of the “mainstream” web indexable and searchable, but I argue the key benefit is the “linkability”

The Spatial Data on the Web best practice document, something of course I recommend you taking a longer look at provides many excellent practical pointers to taking this type of approach.

Maybe really this is just an issue of semantics rather than publishing spatial data should we be talking about sharing spatial data ?


4 replies on “If you can’t link to it… does it exist ?”

agree wholeheartedly – the GIS layer paradigm is essentially about data management, not user needs – though of course typically the user needs to know about the data in its context – which includes its “layer” or other collection. so the challenge is two ways – search engines need to be able to handle zillions of individual items – but perhaps report back results aggregated by a higher order collection concept – i.e. if i search for “Amsterdam weather” and there are individual records for every weather observation then i dont want a list of all of them – i want a link to the historical archive service, the last recorded value and forecast service(s). How do you envisage this happening – can you point to an example from the Dutch Kadaster – maybe a search on a suburb name…?

I might be able to point to an example of the Dutch Kadaster where you can also find the history:

This link points to the Base Registry Addresses and Building and I selected the same building as Ed did (Amsterdam Central station). If you follow the link you will find two ‘voorkomens’ of the same building, the first one being valid from 2010-09-09 until 2013-03-05, the second one being valid from 2013-03-05 until now. In this case we provide you with a link to the building, but also give you the history. Is this what you are looking for?

I’m aiming for this as well.

There are a lot of geographical assets that public bodies require to track and process. Many have reference numbers. It would be great to have the reference number as a parameter in the url allowing for navigation direct to a map focused on the individual record.

The scottish assessors does this with UPRNs

By providing this facility it allows for relations between databases between unrelated organisations and makes their systems an extension of the system of other organisations. If you actually want your data used that’s got to be a big win. Clearly they could do with making the UI of the map display better but the linking is fine.

Thanks Mark, I agree using things like UPRNs help make url’s more readable. I know this is a point of some contention but I think human readable url’s are helpful. I would also say that it’s important for urls to be persistent and reliable, your example relies on the dynamic generation of a page which does not exist beyond this session ? Is that right ?

Add a comment?

This site uses Akismet to reduce spam. Learn how your comment data is processed.