Wednesday, October 04, 2006

deCarta Conference drilling down

Jeepx will be covering the deCarta Developers Conference being realized at Mission Bay UCSF Life Services Conference Center at China Basin. Along the way you can see that the huge SBC Park parking lot is now being converted into housing, streets and new development.

A dog friendly park now follows the channel flowing below the old metal bridge at 4th St. You can get there through Muni or try a ride on the UCSF color coded shuttles. Or walk from the Caltrain Station at 4th Street.

Last Conference deCarta (then still using the Telcontar name) took everyone to SBC Park on a baseball theme. This year we had the Exploratorium to play with and an opportunity to hear the MythBusters live. They even shared a non exhibited part episode on "Do Girls Pass Gas?". Well, maybe not totally correct, but well.

Killer LBS App

Anyway, on a more serious take the 250 attendees heard more than once that LBS is here but with a twist: nobody really wants an LBS app, but some app with LBS integrated to it. More on this in future posts.

Today on a panel driven by John Giudice from TeleAtlas you could hear so far the best direction to where LBS apps can go to help everyone and take off. Among the panelists was Craig Lauer, VP of Engineering at Qualcomm's fleet tracking unit.

In a straightforward approach he suggests to follow the collaboration sharing space and from POI data to life patterns have a way to store historic data about recommended venues.

In practice and in Craig's words: "We need to get rid of the clutter".

Craig continues somehow like this: "Let's say you are at a bus stop and by pushing a button ( in whatever generic device you carry/own/evangelize) you start a geospatial search where based on the historical patterns from the last 12 people doing the same route you get back a result something like this:

"The next bus will arrive in 20 minutes but if you walk down the block you will find a cool movie theater showing good stuff."

Casting, Counting Votes

Based on the same concept inaugurated by Google in using links of popularity a contextual analysis of the real-world can be performed.

Today local searches involve selecting among layers of categories which is just too much. As Craig puts it so plainly: "Let's keep it simple, people already voted about what they do and even wrote about it."

So why not continue casting votes and counting them doing something which is already working like Wikepedia and write, review, edit, put this data out there in an documented and shared format and through an open data access api, let everyone collect, organize and share this material?

How does anyone makes money? Well, you package it in whatever color someone is willing to pay for.

Remember HereCast? Their point is about using position to drop a virtual note on a real building. Maybe it is time to leave the virtual note around here, and figure out a way to find it when you get real at the building.

No comments: