Re: Facebook's definition of a Place

+1!

We are never going to change the POI format of major players like Facebook, Google, NAVTEQ, OpenStreetMap, etc. Our job is how to make these all work together in a way that creates an information community. 

This is why use cases and prototyping is so important. *Maybe* we need a single data model for people to implement, and if so, it's probably good enough to stop at a normative abstract model with informative encodings. But what we definitely need is to define a suite of messages/commands that organizations can make information requests to each other. 

For example, what does Facebook (or a Facebook developer) want to ask of NAVTEQ? Or what requests does a disaster relief worker need to make of the various POI services to aid in search and rescue?

---
Raj

On Nov 4, at 10:45 AM, Dan Brickley wrote:

> Well, i guess what I mean is that gazetteers, gml, kml, georss, geojson, facebook stuff, various AR things, plus essential associated info about places ... these are all very much deployed and won't go away in a hurry. So I see our work in large part, as inventing and deploying a convincing story for how poi info in any of these environments can flow around, without too much loss of meaning, damage to extension info, etc. 
> 
> In other words, multiple formats are a fact of life and hopefully we'll do something more useful than just adding yet another format to support. At some point I'd like to see a test cases repository with descriptions of 1000 POI items, mapped to various formats and to use cases. If we get 1000 and handle say 12 diverse use cases, from shopping guide, tourism and education to emergency response and grassroots journalism, games and accessibility; and if we then show how those items look in kml, gml, geojson, facebook rdfa etc, including URIs for linking, and extensibility fields, ... then I'd say we were done for a v1.0. Is that a realistic hope? :)
> 
> Cheers,
> 
> Dan

Received on Thursday, 4 November 2010 18:14:54 UTC