[OSM-talk] Tagging Multiple Services to a Company/Building

Ian Mc Shane ianmcshane at live.co.za
Thu Feb 4 06:41:10 GMT 2010

Hello Roy,

>> 1)  Do you want all that detail?
> Yes please :)

I understand why you want all that data, but very little of it is being 
visualized through www.openstreetmap.org, leading on to your next comment:

> But I don't have my own database - I want to use the OSM database...
If you are not post processing the data then you will not likely be able to 
get any value from the extra data.  And if people can not see the data, they 
won't know it is there, and then it will go out of date very quickly as 
products that a shop sells will change more often than the road network... 
even with road maintenance taken into account.

With that said, OSM allows users to add as much detail as they want to 
add... and I am quite happy to add the detail, I just want to find out what 
is the best way to do that taking into account that this data will likely 
only be used by people who post-process OSM Data as the renders can not 
handle all the different options right now.

>From what I understand the openstreetmap.org map is purely for show and not 
meant to be THE place to view all the data... its up to other groups to show 
what they want e.g. cloudmade, cyclemap, etc...

Maybe one day we will have "pipemap.org" showing us the world of pipes that 
live beneath our world but that is up to "pipemap.org" and not necessarily 

> An alternative is to use something like a "namespace" concept.
I do like the namespace concept, as well as what Steve was saying ("Are you 
sure it wouldn't be better to push for "amenity=cafe;bakery;atm" style 
multi-tagging instead").  I found it documented here: 

So going with our Spa example:




Multiple values separated by semicolons by far seems to me the simplest way 
to denote multiple values for a single key with less worry that keys will be 

I think that maybe namespace option when things get horribly complicated and 
then the semicolon separated on a shorter/simpler list?

Either mechanism is easy enough for people to do post processing on and 
still be able to group data nicely.

Thanks for the feedback so far, it has been a learning experience.


More information about the talk mailing list