[Tagging] Housenumber interpolation with regularlyskippednumbers

Anthony osm at inbox.org
Wed Oct 14 03:59:21 BST 2009


On Tue, Oct 13, 2009 at 10:28 PM, Mike N. <niceman at att.net> wrote:
>
>>>  A relation would work, but would certainly hide addressing details from
>>> any untrained community wishing to submit corrections.
>>
>> We shouldn't jump through convoluted hoops avoiding relations simply
>> because the editors don't yet make relations easy to edit.
>
>   It's not just the editors - it's the case of getting the help of more
> people who can glance at a common map rendering and see that a correction
> needs to be made.

Renderers can render things that are in relations.  There's nothing
stopping them from doing that.

> Secondly, there are the consumers of OSM data.  There is already a defined
> sequence of steps that must be followed to determine a physical location
> from a street address.  Now add to that a completely separate search and
> calculation algorithm that operates on different objects and must be coded
> and tested.   It's only code and someone can crank out the code and unit
> tests, but completely unnecessary since we already have a construct that is
> well suited when properly qualified.

Determining a physical location from a street address is trivial using
either schema.  Geocoding software would simply need to check for the
address in the Karlsruhe Schema, then, if and only if it could not
find the address there, check for it in the Tiger Schema.  It would
then return a latitude and longitude plus a flag to indicate which
method it used.

Putting both forms of data in the same schema would not make this any
easier.  There still would need to be a check first using the data
marked as complete, and then, if and only if that failed, using the
data marked as incomplete.  The process would be pretty much
identical.

In fact, it's trivial to convert from the Tiger Schema to the
Karlsruhe Schema.  Just copy the ways from the relation, combine them
into one way, and move them by some arbitrary distance away from the
original way.  The reverse conversion would be far from trivial.
Merely finding the original way(s), which may have moved, been
deleted, been renamed, been split, been combined, etc. will be a
nightmare, or even impossible.



But look, I don't really care that much any more.  Do whatever you
want.  If you want to create millions of ways offset from the highways
by some small arbitrary distance, be my guest.  As long as it doesn't
lose information (and it doesn't seem like it's going to, at least not
that much), I'll learn to live with it.  And if it does lose
information, I'll just maintain my own private database separate from
OSM, with the extra information.

I've done some tagging of some different methods of tagging strip
malls under the current (and slightly extended) Karlsruhe Schema.
Unfortunately, Mapnik hasn't gotten around to rendering all the tiles
yet, so I'll wait until later to reveal them.




More information about the Tagging mailing list