From W3C Wiki
Jump to: navigation, search

What might be the components of a user interface for annotating images? What makes sense for ImageDescription and capturing/editing ImageDescriptionRdfExamples, especially for the [WWW2004] Photo History project?


General Description of Interface

Based on looking at the various prototype implementations and cogitating on this for a while, I think it makes much sense to have generally think of the default interface as being as simple as possible and labeling fields specifically for the context of the W3 conferences. This simple interface could be expanded to an advance interface which allows greater flexibility of setting semantic web information. This is only the default interface. Other interfaces could also be created. --GregElin

Necessary Features

The UI really needs the following:

  • Has a easy newbie interface; doesn't require semantic web skill to use
  • Upload multiple images easily, including whole directories
  • Gracefully handles weird image names
  • Minimal number of structured fields
  • Free text field
  • Shows permalink
  • Supported by IE and Mozilla (and as many platforms as possible)
  • Conforms to web standards
  • Makes browsing larger collections feasible if not easeful
  • Identifies owner of image
  • Human and machine readable (photos, license, etc.)

List of "things we want to talk about" (p: for the image itself:

  • dc description
  • creator/maker
  • date
  • license
  • the event depicted
  • perhaps the location

for the areas of the picture:

  • depicts people, maybe objects
  • dc:description perhaps

for the metadata itself:

  • creator/maker
  • license
  • date annotated
  • URL for more information on creator, annotation?

Desirable Features

  • Supports an expert interface
  • Can spider an indicated directory
  • shows copy & paste code (e.g., view source)
  • notifies 'owner' of change
  • supports unique identifiers (mbox, etc.)

Challenging Features and Issues

  • How do we represent LOTS of annotations on an image?
  • How do we show/navigate through layers of annotations? What if 10 people annotate the same person in the same image? Should they be able to? Or are those "annotations of annotations" (e.g., attached comments?)


See also ImageDescription