create an account | login

A Flickr Approach to Making Sense of the World

Review this session

Overall:
  • Rate this talk
  • 1
  • 2
  • 3
  • 4
  • 5

4.25 (12 votes)
Dan Catt:
  • Rate this speaker
  • 1
  • 2
  • 3
  • 4
  • 5

3.91 (11 votes)
leave a written review
Speaker(s): Dan Catt
Date: Thursday, April 24
Time: 1:30 - 2:20PM
Location: 2006

Track: Development
Tags Intermediate, Development

Presented by Dan Catt (Yahoo!, Inc. ). Follow along as Flickr takes something simple, geotagging, makes it nearly unfathomably complex by having an awful lot of it, sifts through it while trying not to make the databases cry (that's the trick), and presents it beautifully and elegantly. A behind-the-curtains peek at the fiery hoops we jump through.

Comments

As a dev guy I found the concepts here interesting, but hard to follow. Dan is kind of like a mad scientest that jumps around from topic to topic. He clearly knows his stuff, but doesn't have a well-paced presentation style.

 

vespa, hilbert space, morton curves all very interesting to listen to but I couldn't follow enough to be able to share with my peers who couldn't come. Maybe a link to something published in this area would be a help.

 

This guy knows his stuff, but it was definitely not scratching where I was itching. It was interesting to see the tech issues of organizing the data, but it wasn't clear that this was what I was getting into. Also, the speaker spoke *really* fast and interjected about a million parenthetical jokes that made it very difficult to even understand what the heck he was saying.

 

Dan is probably a brilliant engineer, but until quite late in the talk I had no idea what he was trying to get across. It would have helped to spend two minutes describing the problem they were trying to solve on a level that normal users would understand.

 

As a counterpoint, I'd say that this was probably the best session that I went to at Web 2.0 Expo - certainly hard to follow if you're not familiar with the problem-space (which perhaps could have also been made clearer), but if you are, this stuff is pure gold. Also, true that some of the jokes fell flat, although the "divisive" agglomerative clustering joke was hilarious.

Also, the "space-filling curve 101" bit on the hilbert and morton curves I thought was quite good and probably as clear an explanation as I've heard.

This stuff, while not rocket science, isn't your run of the mill "web development" so I wouldn't expect anyone to get it (or care), just wanted to add a different perspective to the conversation.