[OSM-talk] Unmapped areas and map confidence

[Jon Burgess]

On Mon, 2007-02-19 at 14:44 +0000, Tom Chance wrote:
> Ahoy,
> 
> I was mulling over this stuff with Robert Scott in a pub last night, along with other worries...
> 
> On Mon, 19 Feb 2007 13:48:56 -0000, "David Earl" <david at frankieandshadow.com> wrote:
> > These address one of my major concerns when I first joined OSM, that you
> > had no idea of whether what you were looking at bore any relation to reality,
> > especially if there was some information there. Of course, the map may
> > lie, but usually it is simply just not finished, and very often it is known it
> > is not finished and is being worked on.
> > 
> > Obviously nothing is ever "complete", in the sense that detail can always
> > be added, but I think if the surfaced, public streets are there with names,
> > then it is "complete" for the purposes of this discussion.
> > 
> > I and presumably RW, both want to say to people looking at the maps we are
> > working on "don't trust this area, we know there is more to do here".
> 
> This relates to knowing who else is mapping in your area. I worry that somebody might start mapping a part of St Albans but carelessly mis certain roads, make spelling mistakes (I found one yesterday), not get all of the pubs/postboxes/etc. My fastiduous side hates the thought of my wee city being defiled ;-)
> 
> Without having a conversation with the person who mapped an area, or going over it again yourself, it's impossible to know.
> 
> Currently we have the wiki pages for localities where you can say "I'm mapping this area", but they don't seem very heavily used and I doubt a lot of mappers find them anyway. There used to be emails in RSS feeds, but so far as I understand it that was hammering the server. There is talk of a facility to get in touch with people in your area by querying the API/db, but it doesn't exist yet. At least being able to get in touch will work if there aren't too many people and everyone is helpful in discussing their activity. But with lots of people, or unhelpful people, it still won't be too much help.
> 
> Ultimately this is one way in which OSM is very different to other "wiki-like" projects. With a page on Picasso in Wikipedia you can tell what needs doing. That's just not true of maps, and any tags we put in to try and indicate completeness will always be flawed.
> 
> The only system I think works well is updating the wiki pages and talking to fellow local mappers, but I don't suppose we can coerce people into doing that.
> 
> Regards,
> Tom

I'm not convinced that a system based on knowing who is mapping your
area is a scalable or long term solution. Much better IMHO would be an
enhancement to the tools like JOSM to do things like highlighting data
updated when you re-download an area. 

I just noticed that doing a search for "timestamp:2007-02" works to
highlight items with a February timestamp. That gives some ability to
to review recent changes in your area.

Personally i'd like to be able to collect data on streets or points of
interest whenever I travel around without needing to worry about
pre-arranging with anyone else or committing to exhaustively mapping an
area. 

The future (as I see it) is that one person with camera phone with
inbuilt GPS might be able to stand in front of a cinema. Do a lookup to
see if it is already in OSM and if not, create and upload an appropriate
node to OSM over the mobile network. Whether the classification is done
by a menu selection on the phone or alternatively the OSM project
receives a geo-tagged JPEG in an inbox which someone else can look at
and add the right data into the DB. 


In cases like this there is no chance that we get exhaustive data, but
by lowering the barrier to entry the overall volume of data can increase
allowing us to reach a higher level of coverage more rapidly. 

Similarly, if we can develop more real time techniques to viewing and
updating map data while travelling around then this allows more
opportunity for noticing areas which are incomplete or incorrect.

	Jon