The metrics TeleAtlas and NAVTEQ give you are all smokescreens and impossible to verify. Completeness and spatial accuracy are interesting but what will be your reference to measure against? What I think is interesting is something you could call crowd quality, where you measure things like how many users have been active in an area, what is their experience / reputation, and how does their mapping activity affect individual features: how many versions, growing attribute richness, spatial convergence. If you can correlate this to the 'objective' quality metric (completeness, accuracy) you could predict how "good" OSM is even in places where you don't have any reference data to measure against.
Martijn van Exel +++ m...@rtijn.org laziness – impatience – hubris http://schaaltreinen.nl | http://martijnvanexel.nl | http://oegeo.wordpress.com/ twitter / skype: mvexel flickr: rhodes On Fri, Nov 26, 2010 at 10:49 PM, SteveC <st...@asklater.com> wrote: > > On Nov 26, 2010, at 2:44 PM, Martijn van Exel wrote: > > > ...and some metric that tells you that the data covers 99.1273% of > reality. fwiw. But there's a point there, serious users want to know more > about quality than they can find out easily right now. How you define > quality, that's another discussion. > > And that's kind of the problem - what is it? > > Everyone wants a simple definition and metric but it just doesn't exist. > > Even when you compare to ground truth, commercial providers are almost as > wrong as they are right. That means if OSM has 100 turn restrictions and > they have 100 it doesn't tell you very much about which ones are right and > which are wrong. Which is counter-intuituve and hard to explain when > advocating OSM as a source. > > > > > > > Martijn van Exel +++ m...@rtijn.org > > laziness – impatience – hubris > > http://schaaltreinen.nl | http://martijnvanexel.nl | > http://oegeo.wordpress.com/ > > twitter / skype: mvexel > > flickr: rhodes > > > > > > On Fri, Nov 26, 2010 at 10:36 PM, SteveC <st...@asklater.com> wrote: > > Speaking personally about what large orgs and what they want, I think > it's pretty simple. Have a look at commercial data and OSM and do a diff, > what are the main things missing? Addressing for geocoding and turn > restrictions for routing. > > > > > > On Nov 26, 2010, at 1:27 AM, Ed Avis wrote: > > > > > I think everyone agrees that detailed legal discussion belongs on the > legal list. > > > > > > Questions such as how any licence transition should proceed, deletion > of existing > > > bits of map, and how to organize the voting process are not legal > arcana but > > > questions of project governance, and surely belong on this list. > > > > > > I am sorry I asked about what Microsoft and others would like to see > from OSM's > > > licensing terms. I hoped that some concrete answers would help > discussion to > > > move on from the mostly fixed positions and legal nitpicking we see on > the legal > > > mailing list (of which I am just as guilty as anyone else). But I > guess the > > > big mapping sites are not willing to make a public statement for fear > of being > > > seen to influence the project. That is a shame, since we are somewhat > in the > > > dark about what the rest of the world thinks. > > > > > > -- > > > Ed Avis <e...@waniasset.com> > > > > > > > > > _______________________________________________ > > > talk mailing list > > > talk@openstreetmap.org > > > http://lists.openstreetmap.org/listinfo/talk > > > > > > > Steve > > > > stevecoast.com > > > > > > _______________________________________________ > > talk mailing list > > talk@openstreetmap.org > > http://lists.openstreetmap.org/listinfo/talk > > > > Steve > > stevecoast.com > >
_______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
