Something Geograph has needed for some time is a way to get at the entire archive without causing us huge bandwidth bills – it’s currently around 100GB in size.
We’re going to release the archive as a series of volumes, each containing around 50,000 images and weighing in at around 5GB. The first volume is available now, with a smaller sample volume which allows you to preview how the metadata is delivered.
All the metadata for each volume is delivered as a RDF formatted file called manifest.rdf, which should be self explanatory, here is a sample fragment:
<dc:title>Durdle Door from the east</dc:title>
<georss:point relationshiptag=”is-picture-of”>50.6205085825182 -2.27398572077417</georss:point>
<georss:point relationshiptag=”was-taken-from”>50.6211552007986 -2.26663784055893</georss:point>
How to download
If you are new to using bittorrent, you will need some software, such as Vuze (aka Azureus) or uTorrent. Once installed, go to http://torrents.geograph.org.uk/ and click a torrent link to set it downloading – easy!
Please seed these torrents if you are able – simply leave your torrent client software running after your download has completed. You’ll be helping others to complete their downloads faster!