[Talk-us] [Imports] [Talk-ca] TIGER considered harmful

[Peter Batty]

When I said "messy", I guess I was thinking of two things - one is doing the
import, as you mention here (which is sort of where the discussion started).
This seems quite a bit more complex if you have to split ways and insert
nodes.

The other is in writing a geocoding engine based on the data which is
produced. If you have the data all on the way, it is a simple query to find
one record, and you interpolate along the geometry. I'm not sure how you
would write an effective geocoding engine directly based on the model with
nodes - I think you would need to write some additional that traced the
network and created a new data structure similar to what you would have in
the case with attributes on the road. So it seems to me simpler to just
create and maintain that data structure directly.

In terms of how to decide what number you use when you split a way, you have
the same problem in either case (whether you have nodes at the beginning and
end of the way, or an attribute range). The most obvious approach is to
interpolate based on distance, which is what geocoding engines using address
ranges do - this would give you the same geocoding results before and after.
If you specifically know the street numbers either side of the split then
you can enter those instead.

Cheers,
    Peter.

On Sun, Nov 15, 2009 at 6:43 PM, Anthony <osm at inbox.org> wrote:

> On Sun, Nov 15, 2009 at 8:32 PM, Peter Batty <peter.batty at gmail.com>
> wrote:
> > If you have two streets intersecting and put a number on that node, it
> isn't
> > clear which street that applies to. You could add an artificial node
> close
> > to the end of the street, but that seems a bit more messy to me.
>
> If you're adding the nodes manually, it's reasonable - you'd want the
> numbers to start at the place where the house is anyway, not at the
> intersection.
>
> If you're adding things automatically, I guess I have to admit it's a
> little messy - much less than adding two ways, but yeah, it's a bit
> artificial (you could always add a second node in the same exact
> location as the intersection, but only connected to one way, but let's
> not go there).
>
> Alternatively, you could use a relation, to specify which way you're
> talking about.  From a technical standpoint I guess that's better, but
> people don't like relations.
>
> > So my gut feel is that the simplest approach is still attributes on the
> street.
>
> How do you split a way?  Do you just guess at the address at the point
> of the split?  Isn't that even more "messy"?
>



-- 
Peter Batty - President, Spatial Networking
W: +1 303 339 0957  M: +1 720 346 3954
Blog: http://geothought.blogspot.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20091115/5b28ac1c/attachment.html>