Re: Creating a complete LOD example

+1 to Kingsley encouraging you to do something you're interested in - 
that's always more fun and more likely to succeed.

I think I'd add a couple of extra criteria to your list. First of all, 
choose something that it's easy to build upon (i.e. where links can 
readily be made).

The chemical elements tick these boxes and most of your original list. 
DBpedia has identifiers for all the elements and Wikipedia is one place 
where you'll find a periodic table with links to related articles 
(http://en.wikipedia.org/wiki/Periodic_table). The Royal Society of 
Chemistry has something similar (and perhaps more authoritative) at 
http://www.rsc.org/periodic-table and you can follow links to chemspider 
which has loads more info (and loads more identifiers). You may need to 
mint terms to link some of these things.

But... given that all those things exist and can be navigated easily by 
humans here's my second suggested criterion: something that would be 
harder to do without LOD.

In other words, show the value of linked data over grabbing a CSV and 
visualising it. Maybe something that would allow people to assert that a 
product, a technology or a process relies on the properties of a 
specific element. To take well known examples, maybe you/someone else 
can link ancient Egyptian cosmetics to antimony and street lamps to 
sodium. A geospatially minded person might then want to link places like 
Ytterby in Sweden to the various rare earth elements discovered there, 
notably by Dr Gadolin ... another link - to the people who first 
isolated them and so on.

HTH

Phil.

On 21/10/2012 14:08, Nicholas Humfrey wrote:
> Hello,
>
> I have created quite number of examples using EasyRdf:
> https://github.com/njh/easyrdf/tree/master/examples
>
> But I would like to create a more complete real-world example, demonstrating how to publish Linked Data, as part of a PHP website. However I am not sure what the subject of the example should be!
>
> Requirements:
> - A subject that people will easily understand
> - A finite number of entities in the domain (in the range of 50-500)
> - An existing vocabulary/ontology for that domain
> - Preferable to already have some kind of identifier
>
> Ideas so far:
> - Chemical Elements
> - Countries of the World (already covered by World Factbook)
> - British Monarchy
> - Stations on the London Underground
> - Airports
>
> Technologies that I plan to use:
> - PHP
> - EasyRdf
> - Slim Framework
> - SPARQL Querying
> - Turtle
> - SPARQL Graph Store HTTP Protocol
> - Twitter Bootstrap
> - PHPUnit tests
> - Some form of visualisation (Graphviz/Maps)
>
>
> Any suggestions?
>
>
> Thanks,
>
> nick.
>
>
>

-- 


Phil Archer
W3C eGovernment
http://www.w3.org/egov/

http://philarcher.org
+44 (0)7887 767755
@philarcher1

Received on Monday, 22 October 2012 07:47:49 UTC