- From: Dan Brickley <danbri@danbri.org>
- Date: Wed, 30 Mar 2011 08:15:08 +0200
- To: public-poiwg@w3.org
Hi folks Some thoughts after yesterday's F2F discussion. http://www.w3.org/2010/POI/wiki/Use_Cases has a number of use cases, dominated for now by 'guide' scenarios oriented towards "Person: end-user", but with a few others tagged "Persona: interconnected systems". I suggest it is worth being more explicit about the latter, and sketching some human-sounding potential users of the POI W3C spec(s) such that we can think about them when creating and closing issues. Here are some that came to mind from personal experience. They are developer centric for now; obviously we have other stakeholders and constituencies but for now: "P3 is a backend developer working on a consumer-facing site that aggregates POI descriptions from around the Web, and attaches social information such as recommendations, endorsements and reviews. T has been working on Web APIs (which have both JSON and XML versions, and which use OAuth for app authorization), on data de-duplication and is currently trying to choose between MySQL, Postgres/PostGIS and a variety of 'NoSQL' options for their main database software. P3 has been meaning to get around to making KML and RSS/Atom feeds or maybe iCalendar feeds, but isn't sure which to do first or how to QA/test/evaluate the results". "P4 is a front-end Web developer working alongside P3. P4 is trying to figure out how many Web frontends to build (main one, a mobile one, 'or do we want one for low-end phones and another for fancy recent smartphones?'). P4 understands that page filesize can be a practical issue for some of their (potential) users in developing countries, and also even for wealthier users when they're travelling and hence roaming. S/he wonders whether the smartphone users should be using an app that caches everything, anyhow? P4 knows Javascript pretty well, is excited to play more with HTML5, has heard a couple of talks about Microformats and has heard a bit lately about Google/Facebook adoption of RDFa. But in the short-term is more concerned to figure out what markup etc to put into pages to add social recommendations (FB likes, digg etc.) or Bing/Google/Yahoo maps, and is wondering if there's a way to get stuff from their pages into users' navigation systems or personal mapping environments like Google's My Maps, or a route-finder. Doesn't KML have something to do with that?" "P5 is an iPhone developer working alongside P4 and P3. S/he works mainly writing Objective-C code for a social/location consumer Web app that calls P3's APIs, although there has been a request to port everything to Android so is wondering how much of the logic can be pushed into Javascript for portability, or up into the server. The app encourages the user to login using Facebook and Twitter credentials so it can share info with friends, access 'social graph' and recommendations/links, and share links. It uses the smartphone's various geo-related APIs and tries to minimise the amount of traffic pulled down from the 'net by using smart (contextual, personalising) filters - or that's the plan anyway. P5 is current debating with P3 how exactly to do this, and what kind of category system would best support this given the hetrogenous nature of their data." "P6 works alongside P3, P4 and P5 at a location services / social Web / mobile geo company, and juggles a variety of roles including platform evangelism and new business development. This includes trying to negotiate access to interesting new geo-tagged datasets, whether this is via adoption of their platform APIs and formats, or giving them Excel files on a data stick. P6 is interested in the idea that there might be a new W3C standard for POI info, so that s/he could ask for that instead, or so that these organizations would already have made it available. P6 works on the assumption that much this data will eventually be public anyhow and that they will distinguish themselves by adding social/business value through filtering, aggregating and linking it, and that to do so will involve connecting to a variety of other similar platforms since different users will make recommendations/checkings/annotations using a variety of services." Others? - developer within a platform ecosystem; eg. creating layar layers, or KML data, or plugin/addon/calling of someone else's geo APIs - someone navigating their own city - someone navigating a city elsewhere when they are 'on roaming' - someone creating a data mashup site that links info about organizations with open public-sector datasets eg. schools, restaurant health inspections, crime data, ... - blind and partially sighted navigators (of their own city, or travelling) - ... Once we've a few more of these, I'd like to make sure we also have at least 10 (ideally 100) POI spots described in human-friendly textual form, that we can use to map into different concrete data formats. For example we might finish describing http://www.pathe.nl/bioscoop/tuschinski as a business, as a tourist spot, historically interesting building or venue for watching movies. We might look at iCalendar feeds, RSS/Atom, KML and HTML home pages as 'carriers' for W3C POI data. All of this would I think better be stored in a filetree than in wiki, ... perhaps we can discuss the mechanics of CSV or distributed versioning options during the F2F? cheers, Dan
Received on Wednesday, 30 March 2011 06:15:40 UTC