I spent some time over the weekend perusing the list of possible recurring events that my search robot found, and recording the useful/valid/appropriate ones in a calendar that syndicates into the Keene calendar hub.
It took me a half hour to go through the first 125 items in that list of 3300 search results. I found ten new recurring events for the Keene calendar. Three or four of those came from PDF newsletters that contained English paragraphs like:
Community Singers: Open singing group, no experience necessary, come for the joy of it. Thursdays from 10:45 to 11:45.
Using ordinary calendar software — in this case Live Calendar, but it could as easily have been Google Calendar, Outlook, Apple iCal, Eventful, or Upcoming — I turned these into iCalendar paragraphs like:
BEGIN:VEVENT RRULE:FREQ=WEEKLY;INTERVAL=1;BYDAY=TH;WKST=SU DTSTART:20090416T104500 DTEND:20090416T114500 SUMMARY:Community Singers DESCRIPTION:http://www.lifeartkeene.org LOCATION:LifeArt Community Resource Center END:VEVENT
The first thought that will occur to technically-inclined readers is: “Hmm. How might I fully automate that transformation?” I understand, and share, that impulse. But I’m trying to set it aside for now, and focus on a different kind of solution.
At a geekish dinner recently, the conversation turned to automation. The geek mind and personality, someone suggested, tends toward an all-or-none approach. It cherishes algorithms that drive fully-automated processes to 100% completion. It does not value methods that achieve partial results, or systems that engage with people to help them do that refinement.
I think that’s true. As I went through the list of candidate events, I reflected on what I was doing. A lot of it wasn’t mere translation from English to iCalendarese. For example, here’s search result #37:
Recreation Center, 312 Washington St., Keene, NH. Western Style Square Dance Apparel ….. “We have a dance every first and third Saturday, no matter what!!!”
Here’s the source of the location information:
And here’s the source of the time information:
These components are unrelated. Or rather they are related, just not in a way that machine intelligence is likely to be able to detect anytime soon. But human intelligence can easily figure out that:
There is an organization called Monadnock Squares
The dances happen at the Keene Recreation Center
These are the kinds of events that happen on regular recurring schedules
So I searched for Monadnock Squares, and wound up adding this event to the calendar:
At this point I realized what the tagline for this project should be. The one I’ve been using is accurate but uninspiring:
community calendar syndication
So I’m going to try this instead:
finding and connecting social capital
When Robert Putnam says that we are bowling alone he adds:
More Americans are bowling than ever before, but they are not bowling in leagues.
Yochai Benkler points out that the networked information economy enhances our ability to:
…do more in loose commonality with others, without being constrained to organize their relationship through a price system or in traditional hierarchical models of social and economic organization…
Maybe there’s plenty of social capital around, but it’s just harder to find, and connect with, because it’s no longer tightly coupled to traditional clubs, leagues, and organizations.
A lot of it is represented online, it turns out. It just isn’t published in a way that’s easy to find and connect with. I hope this project will help change that.
To that end, I’m wondering how to help curators process lists of many thousands of candidate events. Mechanical Turk comes to mind. It would be great to enable curators to carve their lists into batches of 100 and farm them out to volunteers. Is there a free Mechanical-Turk-like service for doing that?
5 thoughts on “Finding and connecting social capital”
I think the somewhere-in-the-messy-middle stuff (rather than the all-or-none automation) is the key innovation that you’re developing here. Facilitate the curation rather than replace it.
On the question of a free Mechanical-Turk-like service – it’s not the answer, but maybe the guys developing http://bookoven.com/gutenberg/ might be more open to sharing the chop-and-share stuff they’ve developed?
“I think the somewhere-in-the-messy-middle stuff (rather than the all-or-none automation) is the key innovation that you’re developing here.”
That and, arguably, the use of Delicious and Friendfeed for lightweight syndicated metadata management and collaboration.
I know them. Will ask. Good thought, thanks. And…hope to see you light up a Liverpool hub one of these days!