Platonides schrieb: > Liam Wyatt wrote: >> Dear All, >> >> [cross post to commons-l and wikimediaaustralia-l] >> >> Picture Australia is interested in adding Commons photos to their >> service but has several technical issues they would like to resolve >> first. Can we help? >> >> Picture Australia is an archive aggregation service run by the National >> Library of Australia and aggregates searches across many Australian >> institutions (such as the various state libraries, universities, >> government departments) and also Flickr. You can see the project at >> http://www.pictureaustralia.org/index.html and you can see their other >> contributors at >> http://www.pictureaustralia.org/contribute/participants/index.html > > Great! > > >> This is a quote from the email written to me from PA: >> At the moment our main source of contemporary images is Flickr and we >> are interested in investigating other sources of contemporary images. >> There are a few issues with the Wikimedia Commons that we foresee: >> 1- the metadata quality is highly variable. With Flickr contributors >> are able to provide a fair bit of additional metadata about their >> images. Before pulling images in from Wikipedia we'd need the data to >> adhere to some basic standards. (see >> http://www.pictureaustralia.org/contribute/metadata.html). > > That would mean adding Dublin Core metadata. A good target by itself. > Perhaps a task for a toolserver app?
We should *really* push for decent metadata support in mediawiki. An appropriate extension already exists, see <http://www.mediawiki.org/wiki/Extension:RDF>. That extension would make it easy to generate RDF (with Dublin Core, Creative Commons and other voicabularies) for each image (and any other page). We should request that extension as a community. If it's not quite ready for prime time, I'm sure it can be made ready. >> 2- there are certainly a number of images that have been sourced from >> Picture Australia or our contributors. Pulling these in would create an >> issue with duplicate images and would likely confuse users if they were >> attempting to buy a copy. > > They shouldn't have problems in keeping a hash of each image they store > to avoid duplicates. We can't guess Right - hashes are important. And images sourced from PA should have that info in their metadata, and are thus easily identified. >> 3- Wikipedia doesn't have an OAI interface so we would need to look at >> how to ingest the data. > > Use the API? > http://commons.wikimedia.org/w/api.php?format=xml&action=query&generator=allimages&prop=imageinfo&iiprop=url|timestamp|user|comment|url|size|sha1|mime|metadata|bitdepth We DO have an OAI interface: <http://meta.wikimedia.org/wiki/Wikimedia_update_feed_service>. -- daniel _______________________________________________ Commons-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/commons-l
