Re: [OSM-talk] addr:interpolation and data consumers

2016-11-21 Thread Mike N 

On 11/21/2016 2:41 AM, Frederik Ramm wrote:

Consumers will often try to simply dissolve the interpolation line into
points which of course has its downsides in schemes like this where 100
interpolated house numbers are assigned but only 10 houses might exist.
(I wonder if we need an interpolation type of "sporadic" for that


  The addr:inclusion=potential tag would seem to cover this : 
https://wiki.openstreetmap.org/wiki/Addresses#Using_Address_Interpolation_for_partial_surveys


___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] addr:interpolation and data consumers

2016-11-20 Thread Andreas Vilén 
It's not always that easy. "3D housenumbering" has no clear tagging rules and 
sometimes it's hard or illegal to get close enough to see which number is used 
for which building/entrance. In other situations the mailboxes are not located 
where the houses are and the houses have no addresses on them, making it hard 
to know which mailbox belongs to which building.

 Interpolating when you know start and finish is more honest than guessing.

/Andreas

Skickat från min iPhone

> 20 nov. 2016 kl. 17:47 skrev nebulon42 :
> 
> Hi,
> 
> I have written an address QA script for Austrian addresses. Now I'm
> asked to support addr:interpolation. While the script is specific to
> Austria the more general problem of addr:interpolation is not.
> 
> In my opinion addr:interpolation is of little value for data consumers.
> Personally, I prefer addresses on nodes or buildings where the location
> of the address is clear. addr:interpolation rather leaves this open. I
> know that addr:interpolation is an established tag, but the Wiki also says:
> 
> "As long as we don't have a node or building outline for each
> house(number) along a way, it's also possible to use automatic number
> interpolation."
> (https://wiki.openstreetmap.org/wiki/Addresses#Using_interpolation)
> 
> For me that sounds like: use it until there is something better/more
> accurate. I tend to replace addr:interpolation with addresses on nodes
> or buildings when I see them and more accurate data is available.
> 
> What is the opinion on addr:interpolation here?
> 
> For reference: https://github.com/gmgeo/at-address-compare/pull/1
> 
> (not on tagging to reach a more diverse audience)
> 
> nebulon42
> 
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] addr:interpolation and data consumers

2016-11-20 Thread Frederik Ramm 
Hi,

On 11/20/2016 05:47 PM, nebulon42 wrote:
> For me that sounds like: use it until there is something better/more
> accurate. I tend to replace addr:interpolation with addresses on nodes
> or buildings when I see them and more accurate data is available.

Yes, for mappers the speedup can be tremendous, in some cities you just
have to walk down one street and for every side street you'll
immediately see, on the street sign, that this block has house numbers
1200 to 1299 or so. Sometimes the numbering is so regular that this
enables you to map house numbers for the next couple blocks down the
side street.

Consumers will often try to simply dissolve the interpolation line into
points which of course has its downsides in schemes like this where 100
interpolated house numbers are assigned but only 10 houses might exist.
(I wonder if we need an interpolation type of "sporadic" for that ;)

Surprisingly, I have occasionally been asked by paying customers to
produce shape files where each street segment has the attributes "house
number left from", "house number left to", "house number right from",
"house number right to" - which is of course less precision than usually
offered by OSM's individual house numbering, but this seems to be a
format commonly used by providers of proprietary data and hence there's
software that expects that format.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] addr:interpolation and data consumers

2016-11-20 Thread Russ Nelson 
nebulon42 writes:
 > For me that sounds like: use it until there is something better/more
 > accurate. I tend to replace addr:interpolation with addresses on nodes
 > or buildings when I see them and more accurate data is available.

That is the right thing to do.

 > What is the opinion on addr:interpolation here?

I cannot speak for other areas, but twenty years ago every town in my
county renumbered their roads so that your house number is 1/200th of
the mile down your road. When a new house gets built, there is no
question what its house number is.

Now, for my area, I've numbered all the houses (well, mailboxes,
because that's where the number is) AND used those points to create
address interpolation ways AND the property centroid has a node with
the cadastral address on it, as well. Mostly I did that on bicycle
rides near my house; not everywhere. And mostly that was done to test
the accuracy of mailbox vs cadastral centroid vs interpolation. And
y'know? It's all good. Any one of them works.

-- 
--my blog is athttp://blog.russnelson.com
Crynwr supports open source software
521 Pleasant Valley Rd. | +1 315-600-8815
Potsdam, NY 13676-3213  | Sheepdog   

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] addr:interpolation and data consumers

2016-11-20 Thread Mike N 

On 11/20/2016 11:47 AM, nebulon42 wrote:

In my opinion addr:interpolation is of little value for data consumers.
Personally, I prefer addresses on nodes or buildings where the location
of the address is clear. addr:interpolation rather leaves this open. I
know that addr:interpolation is an established tag, but the Wiki also says:


  Addr:interpolation is a convenient method for mappers to add many 
addresses at once.  Without an addr:inclusion tag, it implies a complete 
collection of addresses along the line.   The calculated addresses may 
not fall in the center of each building, but they are certainly useful 
to OSM.


___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] addr:interpolation and data consumers

2016-11-20 Thread Martin Koppenhoefer 


sent from a phone

> Il giorno 20 nov 2016, alle ore 17:47, nebulon42  ha 
> scritto:
> 
> For me that sounds like: use it until there is something better/more
> accurate. I tend to replace addr:interpolation with addresses on nodes
> or buildings when I see them and more accurate data is available.
> 
> What is the opinion on addr:interpolation here?


+1
explicit single addresses are better than interpolation data/instructions 


cheers,
Martin 
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] addr:interpolation and data consumers

2016-11-20 Thread Jochen Topf 
On Sun, Nov 20, 2016 at 05:47:28PM +0100, nebulon42 wrote:
> I have written an address QA script for Austrian addresses. Now I'm
> asked to support addr:interpolation. While the script is specific to
> Austria the more general problem of addr:interpolation is not.
> 
> In my opinion addr:interpolation is of little value for data consumers.
> Personally, I prefer addresses on nodes or buildings where the location
> of the address is clear. addr:interpolation rather leaves this open. I
> know that addr:interpolation is an established tag, but the Wiki also says:
> 
> "As long as we don't have a node or building outline for each
> house(number) along a way, it's also possible to use automatic number
> interpolation."
> (https://wiki.openstreetmap.org/wiki/Addresses#Using_interpolation)
> 
> For me that sounds like: use it until there is something better/more
> accurate. I tend to replace addr:interpolation with addresses on nodes
> or buildings when I see them and more accurate data is available.

You have exactly the right interpretation here. Interpolation was only
intended as a way to get going quickly and easily with lots of
addresses. And it has another use: Some non-OSM data sources use a
format where a "street" has attributes giving the house number range on
the left and right side. This can be reasonable easily be transformed
into the addr:interpolation format for importing that data into OSM.

It is here like with everything in OSM: There is a trend towards more
detail. If you have more detailed information, great. If not, it is
better than nothing.

Jochen
-- 
Jochen Topf  joc...@remote.org  https://www.jochentopf.com/  +49-351-31778688

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

[OSM-talk] addr:interpolation and data consumers

2016-11-20 Thread nebulon42 
Hi,

I have written an address QA script for Austrian addresses. Now I'm
asked to support addr:interpolation. While the script is specific to
Austria the more general problem of addr:interpolation is not.

In my opinion addr:interpolation is of little value for data consumers.
Personally, I prefer addresses on nodes or buildings where the location
of the address is clear. addr:interpolation rather leaves this open. I
know that addr:interpolation is an established tag, but the Wiki also says:

"As long as we don't have a node or building outline for each
house(number) along a way, it's also possible to use automatic number
interpolation."
(https://wiki.openstreetmap.org/wiki/Addresses#Using_interpolation)

For me that sounds like: use it until there is something better/more
accurate. I tend to replace addr:interpolation with addresses on nodes
or buildings when I see them and more accurate data is available.

What is the opinion on addr:interpolation here?

For reference: https://github.com/gmgeo/at-address-compare/pull/1

(not on tagging to reach a more diverse audience)

nebulon42



signature.asc
Description: OpenPGP digital signature
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk