Step 1: Converting the schedule

I'm following Edd Dumbill's call to remix the XTech schedule. I wrote a little xml parser for the input file and converted the data to RDF/XML. The remixed schedule is based on RDFical (the version without datatyped timezones), Dublin Core, FOAF and Jen Golbeck's conference ontology.

Step 2: RDF/XML import

So, let's see what the machine-readable data is good for:
confoto.org is a semantic photo service for Web conference pictures. It can't compete with flickr (by no means), but it allows you to add "rich" annotations to photos. It's also not required to upload your files, you can simply post a list of URLs to confoto or, if you are a SemWebber, you can tell confoto where to find your already RDFized photo descriptions.

As we know, metadata doesn't come for free. Flat, tag-based approaches taken by del.icio.us or flickr are currently a huge success, and any tool adding complexity beyond simple tagging could possibly shy away potential users. However, folksonomies are limited, they only define a basic relation between the tag and the annotated resource. With confoto, I'm trying to combine simple tagging with more advanced annotations. For tags I'm following Libby Miller's pigsty approach and use dc:subject literals. For the more sophisticated stuff (creator, event, rating, license) I tried to hide complexity in the UI, keeping the annotation process simple (and metadata cheap). And with an AJAX-like* extension it was possible to add several interactive form elements to the annotator. This brings us back to the XTech data. Using the imported information, we can accelerate annotation tasks as outlined below.
* My stuff is actually AJAI - Asynchronous JavaScript and Iframe ;)

Step 3: Using XTech data in confoto

Let's say you take a picture of Dave Beckett giving his "Bootstrapping RDF applications with Redland" talk. The photo is already uploaded to your own server, so you use confoto's "Link remote images" form and start annotating.


The upload/link-forms allow you to group-annotate photos with basic information such as the depicted conference, the license for photo re-use, or tags.

Once you've uploaded or linked your photos, you can switch to the photo browser and check, if the thumbnail and the basic annotations were created correctly (Don't be surprised if you encounter bugs, I started working on the site less than two weeks ago). Clicking on the thumbnail will display a photo summary from where you can access the annotator, which is split into 5 sections.


The "Essentials" tab is for flickr-like quick annotations such as title, tags, and description. If you don't want to add more information, fine. Del.icio.us & Co have demonstrated how a lot can be achieved with just basic metadata. Likewise, you can use confoto's photo browser to generate galleries based on simple tags defined by you or other users.

For a linked XTech image it would be interesting to add more precise stuff, though, and it's time to finally benefit from the remixed schedule data. So you select the "Persons" tab which will bring up another form where you can define people depicted by the photo. In order to say that your photo doesn't just depict "a" Dave Beckett but the Dave Beckett, the one presenting Redland at XTech, you need at least one identifying property. Luckily, Edd has added encoded mailbox URIs to the XTech schedule, so after entering the first letters we can simply pick an entry from the list of matching persons and are done.


The same type-ahead feature can be used to find conference-related events. confoto allows you to define a main event (e.g. XTech) and a sub-event (e.g. a talk)

Step 4: Instant gratification

The updated photo summary will show available annotations and generate links to related resource descriptions. You can for example open the "depicts David Beckett" link in the [wl http://www.confoto.org/resource_browser resource browser], where the "Incoming relations" tab will bring you to a detailed description of the XTech talk.

You can also generate a gallery of XTech pictures or a gallery of photos tagged with "redland". Or you may subscribe to custom galleries with RSS 1.0 readers, too, as photos are ordered by date.

What else?

Are you a SemWebber? At the bottom of each page, you'll find an "rdf/xml" link which points to a machine-readable version of the currently generated gallery or photo summary. And if the gallery filters are still too restrictive for you, you may try the (experimental) SPARQL interface (some SPARQL endpoint details). Oh, and MGETs on photo URIs should work as well.

That's it for the moment. Now have fun and kill the low-level server this is all running on ;)
Credits to the gargonza crowd for illustrating various SPARQL endpoint/protocol details. That helped me a lot. And thanks in advance for adding some photos and annotations to confoto.org.