First Dirt Party a Success
June 5th, 2007
Dirt Party @ the Eyebeam Benefit
First of all, the idea changed quite a bit since I last wrote about it, so here’s a brief rundown. The project was installed for the Eyebeam 10th Anniversary Benefit (invitation on right), which was a star-studded event to raise money for me and my fellow ‘beamers. Because of certain affiliations and promises that were made, and because of the extremely short time-line, we had to go with a much less automated approach to digging up dirt about the guests at the party. Instead we recruited our own superstar team.
Team Dirt consisted of 3 writers (Joe Garden from The Onion, Josh David Stein from Gawker, and Peggy Wang from Buzzfeed), 3 researchers (Karyn Lu from CNN and Jason Amey and Jen Yao from the NY Public Library), a designer (Clint Hild), 2 reporters (Steve Lambert and Julia Schwadron) , 1 editor (David Jimison), and me. Each team member did their job with the aid of the Dirt Party Publishing System (DPPS), a piece of software designed by Dave and I to facilitate the sharing of research and to dynamically populate the Dirt Party Tabloid Templates (DPTT). The end result was a rotating set of tabloid-style magazine covers, populated with photos found on the web or taken at the event, using gossip collected from the web by or researchers or by our reporters to craft the headlines. We also had these amazing uniforms.
The Interface
When Dirt Party team members logged into the system, they would see an interface suited to their job. There are a few examples below. If they were a researcher, they would see a form where they could enter “Leads”, or story suggestions for the writers. Writers would see a big text area, along with a list of assignments and leads, and the editor would see a bunch of tabs, each corresponding to one DPTT. This might sound confusing, but it will all become clear when I finish editing the demo video.
The Experience
Upon entering Eyebeam, guests were given devices called nTags that kept a tally of who was talking to whom and sent it back to a central server. These tallies, along with an initial Q-score, were used to generate an “interestingness” score for each guest. As editors, Dave and I looked at these ever-evolving interestingness scores and assigned stories to writers based on who was most interesting in the room. The writers would then use the leads and images collected from the researchers to write witty headlines. As the panels in each of the templates filled up with headlines and photos, we would make them live, and they would appear on plasma screens positioned around the room. The picture below is Dave and I consulting with Josh Stein from Gawker. In the upper right you can see the laptop that was running the template switcher, and in the background you can see Marko, who ended up helping us with our networking problems for pretty much the entire benefit. Sorry Marko.
The weeks leading up to the benefit are pretty much a blur. Sometimes I turn into a robot and this was one of those times. For these weeks, I came into work in the morning, worked for 12 hours, went home, worked some more, and then wne tot sleep to rest to do it again the next day. I built an entire publishing system (albeit a buggy and user-unfriendly one) in about 2 weeks, and I had way more responsibility than I was ready to handle. Dave also worked his ass off while he was supposed to be on vacation, so kudos to him.
Post Mortem
During the event, emergencies were coming at me from all angles. I was dodgin’ and weavin’ to avoid the onslaught of urgent problems, but to no avail. I didn’t really calm down until after midnight when the 3 Jack & Cokes brawled their way into my brain, past whatever asshole imaginary bodyguards were keeping the fun out and the stress in. I apologize to everyone who had to interact with me that day, or really the two weeks before the event. I was no fun. Here’s what I learned from all that not-fun.
- The system was far too elaborate. We would have been more productive with a copy of Photoshop and a text editor. I’m not sure why I suffer from this tendency to over-complicate things technologically, but it’s definitely something that I need to look out for.
- If there is a publishing system, avoid hierarchy and division of labor. Everyone should be able to do everything. I thought that this was such a huge job that the tasks needed to be divided into tiny, manageable pieces, and that these pieces would be brought together with the help of the publishing system. As it turns out, the writers really wanted to do their own research, and the reporters were pretty much set to write their own stories after they returned from the field. So these arbitrary limitations were much really only a hindrance.
- Make sure you are credited properly. Due to a misleading invitation, when CNet picked up the story, they credited our project, which they didn’t even mention by name, to BuzzFeed and Gawker, who did take part, but it was not their project. At this point, I guess it doesn’t matter so much, but for Internets posterity, we’d still like the article corrected. Although I can’t blame the reporter, (Ms. Caroline McCarthy) for making that mistake based on the information available to her at the benefit, I can blame her for not getting her mistake corrected after being asked several times.
- Consider using the Amazon Mechanical Turk to farm out research tasks. I kind of fell in love with the idea of crowdsourcing after hearing about Evan and Ben’s White Glove Tracking project, and it got me thinking that perhaps there was a way to crowdsource the research that it takes to drive a dirt party.
- Do traffic tests! In my infinite wisdom, I assumed that my Alienware machine would be able to handle serving the publishing software to the 10 members on the team. I was sadly mistaken. This could have been avoided with a traffic test, but I was never able to get the whole team together before the actual night of the event, which was a problem in itself. I probably could have ironed out many problems if I was able to have a tech rehearsal with all of the team members, but getting 10 busy New Yorkers into a room at the same time is a Herculean task if there ever was one.
- Make sure all team members are into the project. Most people on Team Dirt were extremely devoted to the project and I was very happy with the work that they did. By no fault of their own, some of the people didn’t realize what they were getting into and so instead of being excited about being part of the project, they were more just disappointed that they weren’t out enjoying the party. There are many reasons for this that were out of my control, but in the future I need to make sure that everyone is devoted and understands that the project won’t work if everyone doesn’t participate.
- Scheduled breaks. People were getting burnt out. Lack of food and drink and socialization was rampant. If we had scheduled and timed breaks, we possibly could have avoided this.
- New Job role: Controller. With live video feeds and multiple templates to rotate through, there should be one person whose job it is to “mix” the show. This will be especially important if we go with a more low-tech publishing solution.
Video documentation will be coming as soon as we can get all of the video digitized and edited.
Interns!
My faithful intern, Jung-Eun has been making some progress on Fartsy, but it is still slow going. I am considering a wholly different approach where I have a database of phrases and then search for sentences that contain those phrases using a web search API, and then piece them together. The phrases will be chosen based on the words that the user enters.
The new intern, Andrew Mahon, is working on the Invisible Threads project that Stephanie Rothenberg and I proposed to some conference-type-things. He is picking up on the Second Life stuff remarkably fast, so I have high hopes.
If he finishes that, I want him to build me a system to make YouTube tryptics like the ones that I made at The Advanced Audio Listeners League. Check it out. It is some seriously advanced listening.
FeedBlast Finalist in Rhizome Commission
Congratulations to FeedBlast! It is a finalist in the Rhizome 2008 Commission cycle. Now that the benefit is over, this is my number one priority, so why haven’t I really done any work on it yet? Oh, right, it’s the 7 other projects that I have been distracted by. But right after I publish this entry, I am going back to writing the design document, which will be sent off to my collaborator, Bjoern Wilmsmann of Topicalizer, and we will start discussing serious development. If I do get this grant (and/or the NYSCA grant), it will go directly towards hiring a programmer with whom I can start planning the video editing component, which will be my part of the project. If the grants don’t come through, I will start conferring with myself on this matter. Either way, some conferrin’ is gonna happen.
The one thing that has been giving me trouble is the effort to get some kind of discount on text to speech software. IBM and AT&T are the target organizations. Both of them have charitable giving programs and both have amazing TTS packages. Since I will be releasing FeedBlast as open source, my sense of cosmic justice tells me that I shouldn’t have to pay the $10,000 that they usually charge. Five digit numbers with a ‘$’ in front are for huge corporations and mustachioed kingpins, not for dudes who have trouble keeping their checking account above the minimum balance and give their work away for free. Plus, Eyebeam has offered to put their logo on the Eyebeam website and Eyebeam is a proper NPO, and I hear that corporations like that shit. So far, my efforts have been met with indifference, both within Eyebeam and from IBM/AT&T. I think this project appeals mostly to the nerdy types - those who live life in the fast lane on the Information Superhighway (italicized to emphasize the fastness of their lifestyles), and I just haven’t found anyone like that yet to champion FeedBlast.
Hit! or Sh!t
Things with Hit! or Sh!t couldn’t be going better, sort of. Last week, Dave and I went up to the MTV Death Star on Broadway to talk with the Vader types up there (or, went to Endor to talk with the Cheif Ewoks, if you prefer). We’re excited, their excited, everyone is excited. Whether all this oozing excitement translates into any actual progress is yet to be seen. Even more unlikely is Hit! or Sh!t making its way to where they want it to go and Dave and I having anything to do with it afterwords. Oh well. When the steamroller does finally come and roll over everything we’ve worked so hard on, hopefully we will have gotten out of the way first. That’s all I’m saying.
Related Stuff
- June 1st, 2008 - Here I am
No Comments
Leave a comment
