Interior ring with Potlatch?

classic Classic list List threaded Threaded
19 messages Options
Reply | Threaded
Open this post in threaded view
|

Interior ring with Potlatch?

David Fawcett
I just added a building outline using Potlatch and drew a second area
polygon to delineate an open area in the building.  I am not sure how
to tie this second polygon to the first one to create the interior
ring in the polygon.  When I drew them, I went clockwise with the
exterior ring and counter-clockwise with the interior ring. I am not
sure if this makes any difference.  Google didn't turn up any good
answers.

The building is at:
http://www.openstreetmap.org/?lat=44.994395&lon=-93.30004&zoom=18

Thanks,

David.

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Richard Weait
On Wed, Dec 9, 2009 at 10:25 PM, David Fawcett <[hidden email]> wrote:

> I just added a building outline using Potlatch and drew a second area
> polygon to delineate an open area in the building.  I am not sure how
> to tie this second polygon to the first one to create the interior
> ring in the polygon.  When I drew them, I went clockwise with the
> exterior ring and counter-clockwise with the interior ring. I am not
> sure if this makes any difference.  Google didn't turn up any good
> answers.
>
> The building is at:
> http://www.openstreetmap.org/?lat=44.994395&lon=-93.30004&zoom=18

You'll want to create a relation
type=multipolygon
Make your ways members of the relation,
Then define the role=outer and role=inner for your two existing ways.

http://wiki.openstreetmap.org/wiki/Relation:multipolygon

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Steve Bennett-3
In reply to this post by David Fawcett
On Thu, Dec 10, 2009 at 2:25 PM, David Fawcett <[hidden email]> wrote:
I just added a building outline using Potlatch and drew a second area
polygon to delineate an open area in the building.  I am not sure how
to tie this second polygon to the first one to create the interior
ring in the polygon.  When I drew them, I went clockwise with the
exterior ring and counter-clockwise with the interior ring. I am not
sure if this makes any difference.  Google didn't turn up any good
answers.

The building is at:
http://www.openstreetmap.org/?lat=44.994395&lon=-93.30004&zoom=18


I'm curious too. The workaround I've used is to make the whole thing one shape, by splitting through one bit. Hard to describe, but imagine a C shape. Now make the ends of the C closer and closer until they touch - it looks like an O, but it's actually a C.

Or, um, do what Richard said. :)

Steve

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Randy-4
Steve Bennett wrote:

>I'm curious too. The workaround I've used is to make the whole thing one
>shape, by splitting through one bit. Hard to describe, but imagine a C
>shape. Now make the ends of the C closer and closer until they touch - it
>looks like an O, but it's actually a C.
>
>Or, um, do what Richard said. :)
>
>Steve

JOSM Validator has flagged one of those in my area. I think the tips of
the "C" may be overlapping a little, but I haven't gotten around to fixing
it yet. A multipolygon is, I believe, the best practice in this case,
since your "pseudo wall" isn't really an exterior part of the building.

--
Randy


_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Steve Bennett-3
On Thu, Dec 10, 2009 at 4:45 PM, Randy <[hidden email]> wrote:
> JOSM Validator has flagged one of those in my area. I think the tips of
> the "C" may be overlapping a little, but I haven't gotten around to fixing
> it yet. A multipolygon is, I believe, the best practice in this case,

Undoubtedly.

> since your "pseudo wall" isn't really an exterior part of the building.

Oh, I didn't know they had to be. What do you do when there's a
complex of clearly distinct buildings that touch each other? I thought
tagging them separately (but sharing ways) was the right thing to
do...

Steve

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Tsvi Mostovicz
In reply to this post by David Fawcett
> Oh, I didn't know they had to be. What do you do when there's a
> complex of clearly distinct buildings that touch each other? I thought
> tagging them separately (but sharing ways) was the right thing to
> do...
>
> Steve

I had a similar case over here of multiple buildings touching each other. I
decided to have them share ways although JOSM's validator complains about it.
If anyone has a better practice please share it with me.

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Ævar Arnfjörð Bjarmason
On Thu, Dec 10, 2009 at 13:11, Tsvi Mostovicz <[hidden email]> wrote:

>> Oh, I didn't know they had to be. What do you do when there's a
>> complex of clearly distinct buildings that touch each other? I thought
>> tagging them separately (but sharing ways) was the right thing to
>> do...
>>
>> Steve
>
> I had a similar case over here of multiple buildings touching each other. I
> decided to have them share ways although JOSM's validator complains about it.
> If anyone has a better practice please share it with me.

That's the right way to do it. Just because the validator whines about
something that doesn't mean it's incorrect. It complains loudly about
e.g. boundary areas sharing ways even though that's the best practice.

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Steve Bennett-3
On Fri, Dec 11, 2009 at 12:51 AM, Ævar Arnfjörð Bjarmason
<[hidden email]> wrote:
> That's the right way to do it. Just because the validator whines about
> something that doesn't mean it's incorrect. It complains loudly about
> e.g. boundary areas sharing ways even though that's the best practice.

The validator complains about colinear ways? Wtf? Definitely best
practice, as far as I'm concerned.

(By sharing ways, I mean that several nodes are part of two distinct
ways - not two nodes almost on top of each other.)

Btw, I'm starting to question the fact we don't do it with roads
though. The logic goes that the line represents the middle of the
road, and having an area touch the line would mean that the area goes
up to the middle of the road. I would argue that the line represents
the middle of the road *plus an arbitrary width*. When a footway
touches a road node, it doesn't mean that the middle of the road is
the middle of the footway. It means that road plus arbitrary width =
footway plus arbitrary width.

Steve

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Pieren
On Thu, Dec 10, 2009 at 3:15 PM, Steve Bennett <[hidden email]> wrote:
> On Fri, Dec 11, 2009 at 12:51 AM, Ævar Arnfjörð Bjarmason
> <[hidden email]> wrote:
>
> The validator complains about colinear ways? Wtf? Definitely best
> practice, as far as I'm concerned.
>

hmm, what means "best practice" in OSM...
Anyway, it is reported in the "Warnings" and I never consider what is
below "Errors" from this plugin.

Pieren

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Ævar Arnfjörð Bjarmason
In reply to this post by Steve Bennett-3
On Thu, Dec 10, 2009 at 14:15, Steve Bennett <[hidden email]> wrote:

> On Fri, Dec 11, 2009 at 12:51 AM, Ævar Arnfjörð Bjarmason
> <[hidden email]> wrote:
>> That's the right way to do it. Just because the validator whines about
>> something that doesn't mean it's incorrect. It complains loudly about
>> e.g. boundary areas sharing ways even though that's the best practice.
>
> The validator complains about colinear ways? Wtf? Definitely best
> practice, as far as I'm concerned.
>
> (By sharing ways, I mean that several nodes are part of two distinct
> ways - not two nodes almost on top of each other.)
>
> Btw, I'm starting to question the fact we don't do it with roads
> though. The logic goes that the line represents the middle of the
> road, and having an area touch the line would mean that the area goes
> up to the middle of the road. I would argue that the line represents
> the middle of the road *plus an arbitrary width*. When a footway
> touches a road node, it doesn't mean that the middle of the road is
> the middle of the footway. It means that road plus arbitrary width =
> footway plus arbitrary width.

Who is "we"? Some of us map just as you suggest, when I'm mapping e.g.
landuse areas that are marked by some road I'll make the landuse area
share nodes with the way.

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Steve Bennett-3
In reply to this post by Pieren
On Fri, Dec 11, 2009 at 1:30 AM, Pieren <[hidden email]> wrote:
> hmm, what means "best practice" in OSM...
> Anyway, it is reported in the "Warnings" and I never consider what is
> below "Errors" from this plugin.

Hmm, just had a go of keepright now, and I'm told that "landuse=wood"
is deprecated. Just when I finally thought I had understood
vegetation...heh.

Steve

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Steve Bennett-3
In reply to this post by Ævar Arnfjörð Bjarmason
On Fri, Dec 11, 2009 at 1:45 AM, Ævar Arnfjörð Bjarmason
<[hidden email]> wrote:
> Who is "we"? Some of us map just as you suggest, when I'm mapping e.g.

I came across this rule in the wiki somewhere.

Steve

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Ævar Arnfjörð Bjarmason
On Thu, Dec 10, 2009 at 14:59, Steve Bennett <[hidden email]> wrote:
> On Fri, Dec 11, 2009 at 1:45 AM, Ævar Arnfjörð Bjarmason
> <[hidden email]> wrote:
>> Who is "we"? Some of us map just as you suggest, when I'm mapping e.g.
>
> I came across this rule in the wiki somewhere.

Just because someone says something on the wiki doesn't make it true.
In actuality the wiki is very under maintained and the typical page
describing a given tag combination or mapping practice has seen input
from maybe 5-10 users while an order of magnitude more users use those
tags/mapping practices, and often in ways not covered by the wiki
page.

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Steve Bennett-3
On Fri, Dec 11, 2009 at 2:08 AM, Ævar Arnfjörð Bjarmason
<[hidden email]> wrote:
> Just because someone says something on the wiki doesn't make it true.
> In actuality the wiki is very under maintained and the typical page
> describing a given tag combination or mapping practice has seen input
> from maybe 5-10 users while an order of magnitude more users use those
> tags/mapping practices, and often in ways not covered by the wiki
> page.

Yeah. IMHO, this is a good reason to improve wiki practice, but I can
understand if others prefer to simply disregard the wiki.

Steve

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Mike N.
In reply to this post by Tsvi Mostovicz
> I had a similar case over here of multiple buildings touching each other.
> I
> decided to have them share ways although JOSM's validator complains about
> it.
> If anyone has a better practice please share it with me.

  I didn't remember that the validator complains ... you do need to be
careful to be sure that all common nodes are shared by both buildings.
Otherwise it's easy to have them cross by a barely visible amount, and the
validator notes the 'Crossing Ways'.
 


_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Randy-4
In reply to this post by Steve Bennett-3
Steve Bennett wrote:

>On Thu, Dec 10, 2009 at 4:45 PM, Randy
><[hidden email]> wrote:
>>JOSM Validator has flagged one of those in my area. I think the tips of
>>the "C" may be overlapping a little, but I haven't gotten around to fixing
>>it yet. A multipolygon is, I believe, the best practice in this case,
>
>Undoubtedly.
>
>>since your "pseudo wall" isn't really an exterior part of the building.
>
>Oh, I didn't know they had to be. What do you do when there's a
>complex of clearly distinct buildings that touch each other? I thought
>tagging them separately (but sharing ways) was the right thing to
>do...
>
>Steve

Concede. "Exterior" was not a good word to use here, I was being lazy. How
about "building boundary demarcation"? (Whether you agree with the intent
of the statement or not.) Even though the building will be rendered as a
whole (with a hole), the underlying data will indicate a discontinuity.

Regarding using a single way for multiple purposes, this is done quiet
often, and I think very appropriate. However, I think the mapper should
consider the potentially common boundary when doing this. Are either side
of the boundary mutually exclusive? That's definitely a case for a common
boundary. If the purposes of the potentially common way are unrelated,
e.g. a highway and an area boundary, then is the area boundary likely to
move if the highway moves? If yes, use a common boundary. If the two uses
are not exclusive, and the relocation of one will not necessarily require
the relocation of the other, then I think separate ways should be used.
That will certainly ease the effort of the mapper who may eventually have
to move one and not the other.

--
Randy


_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Randy-4
In reply to this post by Mike N.
Mike N. wrote:

>>I had a similar case over here of multiple buildings touching each other.
>>I
>>decided to have them share ways although JOSM's validator complains about
>>it.
>>If anyone has a better practice please share it with me.
>
>   I didn't remember that the validator complains ... you do need to be
>careful to be sure that all common nodes are shared by both buildings.
>Otherwise it's easy to have them cross by a barely visible amount, and the
>validator notes the 'Crossing Ways'.
>

I believe that's precisely the problem, with the local area I found,
although I haven't gotten back to it yet. The "glue joints" overlap,
rather than being common ways.

I still contend that even though the renderer will make this look like a
contiguous area, the underlying data will contain a discontinuity which
does not reflect reality,and which, for some future purpose, could cause a
problem. For example, how do I get from point A to point B without leaving
the area? (Oops, here's the boundary, is it a real boundary or just a
pseudo boundary?) So, I'll use multipolygons for interior courtyards, etc.
because, from my personal view, it is the "safer" approach, and is a
better way to "map reality" rather than map for the renderers.

As always, in OSM, YMMV. (Your mapping may vary.)

--
Randy


_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Steve Bennett-3
On Fri, Dec 11, 2009 at 5:33 AM, Randy <[hidden email]> wrote:
> I still contend that even though the renderer will make this look like a
> contiguous area, the underlying data will contain a discontinuity which
> does not reflect reality,and which, for some future purpose, could cause a
> problem. For example, how do I get from point A to point B without leaving
> the area? (Oops, here's the boundary, is it a real boundary or just a
> pseudo boundary?) So, I'll use multipolygons for interior courtyards, etc.
> because, from my personal view, it is the "safer" approach, and is a
> better way to "map reality" rather than map for the renderers.

Err, I don't think anyone (least of all myself) was advocating using
the hack I described. I simply didn't know about multipolygons.

Steve

_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies
Reply | Threaded
Open this post in threaded view
|

Re: Interior ring with Potlatch?

Randy-4
Steve Bennett wrote:

>On Fri, Dec 11, 2009 at 5:33 AM, Randy
><[hidden email]> wrote:
>>I still contend that even though the renderer will make this look like a
>>contiguous area, the underlying data will contain a discontinuity which
>>does not reflect reality,and which, for some future purpose, could cause a
>>problem. For example, how do I get from point A to point B without leaving
>>the area? (Oops, here's the boundary, is it a real boundary or just a
>>pseudo boundary?) So, I'll use multipolygons for interior courtyards, etc.
>>because, from my personal view, it is the "safer" approach, and is a
>>better way to "map reality" rather than map for the renderers.
>
>Err, I don't think anyone (least of all myself) was advocating using
>the hack I described. I simply didn't know about multipolygons.
>
>Steve

I thought I read advocation into some of the thread, but I wasn't pointing
a finger at you. Sorry if it came across that way. I try not to be
accusative, as it's rarely productive. Still, I do slip up on occasion, as
my wife is quick to point out:-)

--
Randy


_______________________________________________
newbies mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/newbies