[OSM-talk] Building Equals Yes
straup
straup at gmail.com
Wed Jun 8 16:24:11 BST 2011
Hi all,
[I made buildingsequalyes.spum.org]
First, thanks to everyone for their kind words about the project! It's
been a couple of weeks since it was pushed out the door and I've mostly
been letting it "bake" to see if anything breaks and to solicit comments
and suggestions.
The "=yes" part will be corrected in the next version/release. I was not
aware that there were people tagging buildings with more specific values
so the next dump will filter on "building=*" although the project name
will stay the same since "buildingequalssplat" doesn't have quite the
same ring.
Second, I've seen a few people express concern over a comment I made on
the site's about page to the effect that I wanted to import all the WOE
IDs [1] (for neighbourhoods, localities, etc.) that I had associated
with each of the building way as new tags. For example:
# k="woe:locality" v="3534"
# k="woe:region" v="2344924"
# k="woe:country" v="23424775"
http://buildingequalsyes.spum.org/woeid/3534/
So, I'd like to toss the idea out there and ask: Will this make people
upset and if so why? It's obviously something I think would be
interesting but it's also not something I want to pick a fight with
people over.
I know there have been some very bad experiences in the past with any
kind of bulk import but, on that level, this seems different.
Specifically it's only adding new (and easy identifiable by the "woe:"
prefix) tags to existing ways and not touching any other data.
Personally, I've long wanted to have a way to query OSM by "place"
rather than geography and this is one way to do that. I was reading the
wiki page for XAPI, this morning, and seeing the proposed "search by
region" extension seems like much the same thing so maybe I'm not alone?
There's nothing inherently special about WOE identifiers over others
except that a) I am familiar with them (I used to work at Flickr) b)
it's an open dataset which means whatever thrashing around Y! does going
forward won't have any effect on them c) the ~200M geotagged photos at
Flickr all of which have WOE IDs. [2]
The WOE IDs themselves were derived by using Flickr's reverse geocoder
[3] which I can guarantee you makes some weird/just plain wrong
suggestions. I know this because I wrote the reverse geocoder but it
also means I have more confidence in it than the results that come back
from other services.
To the extent that there are errors in the buildingsequalsyes collection
most of the results are correct and one benefit of getting them in to
OSM would be to have the WOE IDs vetted by the many eyes of the community.
And if we can do that then there is also the opportunity to start to
create shapes for those places, using all of OSM, in the same way that
Flickr generated their shapefiles using geotagged photos. [4]
This is not something I want to do if it's going to upset or freak
people out. If they need to those mappings between WOE IDs and OSM way
IDs can always live outside OSM but I think it would be useful if they
were part of the larger whole.
Crazy talk?
--
[1] http://developer.yahoo.com/geo/geoplanet/guide/concepts.html
[2] For example: http://www.flickr.com/places/3534/
[3] http://www.flickr.com/services/api/flickr.places.findByLatLon.html
[4] http://code.flickr.com/blog/2008/10/30/the-shape-of-alpha/
On 5/23/11 2:50 PM, Robin Paulson wrote:
> On 17 May 2011 17:29, Michal Migurski<mike at stamen.com> wrote:
>> By my colleague Aaron Cope:
>>
>> "building=yes is a searchable and linkable index of every singleway
>> tagged building=yes in OpenStreetMap (OSM).
>>
>> A web page for every building in OpenStreetMap!
>
> i love it. excellent, and very imaginative.
>
> great title too
>
More information about the talk
mailing list