Developers are People, Too Mor Naaman Yahoo! Research Berkeley *Any views or opinions expressed during this panel are those of the author only and do not represent those of Yahoo, Yahoo! Research, or Flickr -- sometimes they do not even represent the author himself.
People are Strange People are Strange. http://flickr.com/photos/genial23/334807527
Why do People Tag? Well, mostly, they don’t But if they do (say, on Flickr): “ Why We Tag”, CHI2007
Explanation, please Simple Source: Rashmi Sinha Powerful Tied with their natural motivations
Our Mission Automatically generate tag semantics Few places to start Word-based models  E.g., Flickr Clustering E.g., Subsumption E.g., LSI Metadata can help, too
tagmaps.research.yahoo.com “ World Explorer”, JCDL2007
Extracting Event/Place Semantics “ Extracting Tag Semantics”, SIGIR2007
Developers are people, too! http://flickr.com/photos/fotex/217896072/
Developer Needs Requirements for developing applications that make use or manipulate metadata: Simple Powerful Tied with their natural motivations Just like people!
I too am a Web2.0 Developer My team will support a format: If it’s simple If it’s useful for us If it “plays nice” with key applications So far: RSS and GEO-RSS KML and GPX EXIF (read only) Flickr Machine Tags
Machine Tags?? “ Machine tags are not RDF but they could play RDF on TV.” - Aaron Cope, Hackr (Flickr) Book:author=“Tim B.L.” celltower:lac=205 Geo:latitude=23.12 ylocal:venue=224323 camel:humpcount=2 upcoming:event=12413 flower:color=red summit:altitude=2321
What’s the Point? Format is emerging Reflects needs of developers  Reflects properties of data Simple Powerful
What’s the Problem? Namespace/semantics not defined Multiple formats for same semantics Multiple semantics for same subjects/predicates/objects… It’s up to us to make sense of it Us == researchers Schema mapping, pattern extraction and recognition, semantic analysis…
Discuss: The Semantic Web is Dead Welcome to the  Emerging-Semantics Web!
Send me your wrath More fun and games: www.yahooresearchberkeley.com Mor Naaman http://infolab.stanford.edu/~mor

Developers Are People, Too

  • 1.
    Developers are People,Too Mor Naaman Yahoo! Research Berkeley *Any views or opinions expressed during this panel are those of the author only and do not represent those of Yahoo, Yahoo! Research, or Flickr -- sometimes they do not even represent the author himself.
  • 2.
    People are StrangePeople are Strange. http://flickr.com/photos/genial23/334807527
  • 3.
    Why do PeopleTag? Well, mostly, they don’t But if they do (say, on Flickr): “ Why We Tag”, CHI2007
  • 4.
    Explanation, please SimpleSource: Rashmi Sinha Powerful Tied with their natural motivations
  • 5.
    Our Mission Automaticallygenerate tag semantics Few places to start Word-based models E.g., Flickr Clustering E.g., Subsumption E.g., LSI Metadata can help, too
  • 6.
  • 7.
    Extracting Event/Place Semantics“ Extracting Tag Semantics”, SIGIR2007
  • 8.
    Developers are people,too! http://flickr.com/photos/fotex/217896072/
  • 9.
    Developer Needs Requirementsfor developing applications that make use or manipulate metadata: Simple Powerful Tied with their natural motivations Just like people!
  • 10.
    I too ama Web2.0 Developer My team will support a format: If it’s simple If it’s useful for us If it “plays nice” with key applications So far: RSS and GEO-RSS KML and GPX EXIF (read only) Flickr Machine Tags
  • 11.
    Machine Tags?? “Machine tags are not RDF but they could play RDF on TV.” - Aaron Cope, Hackr (Flickr) Book:author=“Tim B.L.” celltower:lac=205 Geo:latitude=23.12 ylocal:venue=224323 camel:humpcount=2 upcoming:event=12413 flower:color=red summit:altitude=2321
  • 12.
    What’s the Point?Format is emerging Reflects needs of developers Reflects properties of data Simple Powerful
  • 13.
    What’s the Problem?Namespace/semantics not defined Multiple formats for same semantics Multiple semantics for same subjects/predicates/objects… It’s up to us to make sense of it Us == researchers Schema mapping, pattern extraction and recognition, semantic analysis…
  • 14.
    Discuss: The SemanticWeb is Dead Welcome to the Emerging-Semantics Web!
  • 15.
    Send me yourwrath More fun and games: www.yahooresearchberkeley.com Mor Naaman http://infolab.stanford.edu/~mor

Editor's Notes

  • #2 http://en.wikipedia.org/wiki/Semantic_web Subject predicate object