[mkgmap-dev] [Patch v2] addr:interpolation

Mon Dec 22 13:34:05 GMT 2014

Hi WanMil,

the reason is that we have multiple OSM ways with the name Rosewood,
and RoadMerger doesn't merge them all to a single way.
So, a list of roads with the name Rosewood is processed
against a list of housenumbers.
Now, e.g. the node 427882156 with number 170
is closer to the OSM way 29272180 than to 
way 29272175, which would be the better match
when you notice the addr:interpolation way.

Same happens with other nodes, so that some 
road segments get only one number instead of two.
As a result, the map doesn't contain enough 
information for a good interpolation.

The patched version is a bit better although it
doesn't solve the initial problem, it just adds 
more data so that each road segement has at least
a few correct numbers.
Maybe I find a way to code a plausibilty test
for the numbers, I need something like that also 
for the addr:place evaluation.

Gerd

> Date: Sat, 20 Dec 2014 13:45:44 +0100
> From: wmgcnfg at web.de
> To: mkgmap-dev at lists.mkgmap.org.uk
> Subject: Re: [mkgmap-dev] [Patch v2] addr:interpolation
> 
> Hi Gerd,
> 
> I didn't expect a big improvement by supporting addr:interpolation. 
> There are only few things I can imagine:
> * addr:interpolation is tagged with addr:street but the nodes with 
> addr:housenumber are not tagged with addr:street
> * addr:interpolation span over more than one street segment
> 
> I am curious about your results!
> 
> WanMil
> 
> > Hi Carlos,
> >
> > thanks for the feedback.
> > I must confess that I was a bit surprised about the improvement.
> > In other words, I thought that I might have found an error
> > in the existing code, so I started to look closer at the
> > HousenumberGenerator,
> > also because of the addr:place problem.
> >
> > @WanMil: Do you have a good explanation for the improvement?
> >
> > Gerd
> >
> >  > Date: Fri, 19 Dec 2014 10:37:07 +0100
> >  > From: cdavilam at orangecorreo.es
> >  > To: mkgmap-dev at lists.mkgmap.org.uk
> >  > Subject: Re: [mkgmap-dev] [Patch v2] addr:interpolation
> >  >
> >  > El 09/12/14 a las 16:38, Gerd Petermann escribió:
> >  > > Hi all,
> >  > >
> >  > > attached is v2.
> >  > > It turned out to be quite complicated to interpret the data
> >  > > provided with this tag, and I found only few places were this
> >  > > helps to provide better data. On the other hand, I see no negative
> >  > > effects besides the additional code complexity.
> >  > >
> >  > > A good place to see how address search is improved is here:
> >  > > http://www.openstreetmap.org/way/36787892#map=17/51.88859/-8.57449
> >  > >
> >  > > With the trunk version, a search for 178 Rosewood lists several
> > numbers,
> >  > > e.g. 184, 170, 199 and so on. None is really correct.
> >  > > The patched version shows exactly one entry, which is probably very
> > close.
> >  > > Same result with other numbers which do not appear as housenumbers.
> >  > >
> >  > > A patched binary based on r3366 is here:
> >  > > |http://files.mkgmap.org.uk/download/236/mkgmap.jar
> >  > >
> >  > > |Gerd
> >  > >
> >  > I tested your patch on a different area than the one you suggest and it
> >  > really improves house number search.
> >  > _______________________________________________
> >  > mkgmap-dev mailing list
> >  > mkgmap-dev at lists.mkgmap.org.uk
> >  > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> >
> >
> > _______________________________________________
> > mkgmap-dev mailing list
> > mkgmap-dev at lists.mkgmap.org.uk
> > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> >
> 
> _______________________________________________
> mkgmap-dev mailing list
> mkgmap-dev at lists.mkgmap.org.uk
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mkgmap.org.uk/pipermail/mkgmap-dev/attachments/20141222/08cf3c59/attachment.html>