Re: [Talk-GB] Adjacent nature reserves

2019-06-27 Thread Mateusz Konieczny



28 Jun 2019, 02:29 by for...@david-woolley.me.uk:

> On 28/06/2019 00:56, Warin wrote:
>
>> that are also holes in them (they usually omit making the hole, so an added 
>> car parking area will be covered by trees until I notice
>>
>
> I believe that is a renderer bug.  Generally smaller, fully nested, areas 
> should cut out holes in incompatible backgrounds without explicit relations.
>
It is not a renderer bug. At least in case of OSM Carto, where tree-covered 
area is deliberately
working in this way to indicate missing multipolygons. And to support case of 
tree covered areas of 
various types.

Missing explicit holes are not a biggest mapping quality issue. but  it is 
better to 
make it explicit, in almost all cases "is it incompatible area" is not actually 
obvious and the
same across world. Also processing such data with implicit holes is 
unreasonably hard.

Though I am still not a fan of using multipolygons where there are no holes in
areas.
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-27 Thread David Woolley

On 28/06/2019 00:56, Warin wrote:
that are also holes in them (they usually omit making the hole, so an 
added car parking area will be covered by trees until I notice


I believe that is a renderer bug.  Generally smaller, fully nested, 
areas should cut out holes in incompatible backgrounds without explicit 
relations.


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-27 Thread Warin

On 27/06/19 22:11, Martin Wynne wrote:



seen this done in various places, but I've never understood the point
it. The two representations are identical in terms of the data, but
the latter requires 2.5 times as many objects and is much more of a
pain to work with in the editors.


This happens a lot in my area. Huge areas of "farmland" have been 
created as massive multipolygons, which are too big to fit in the iD 
editor, and include ways shared with other areas such as equally large 
multipolygon woods. It's a pain to split them up without damaging them 
where they include areas which should be mapped as meadow, orchard, 
scrub, etc., which I much prefer to map as separate closed field 
areas, sometimes with their own name. area


I am reworking an tree area that I created some time ago .. it is some 
300 km long, 40 km wide (186 miles by 25 miles). There are several 
similar sized areas that exist in my locally to me. They have to be 
mulipolygons as they have holes in them. And people keep adding things 
inside them .. that are also holes in them (they usually omit making the 
hole, so an added car parking area will be covered by trees until I 
notice).


Map your closed fields .. and simply included them as an inner in the 
tree area?




Likewise several woods are mapped as a single large multipolygon wood 
where in fact they are several separate woods each with a *name*. How 
can I apply names to parts of a multipolygon?


You cannot. However for 'my' above tree area there are several National 
Parks .. I map those and name them .. they still have the tree area 
through them (in parts) but they are still rendered and named on the 
map. I separate out the forestry areas as these don't have trees all the 
time, so they get landuse=forest rather than natural=trees.



I have other places where there are ways that are in up to 6 relations, 
much easier to map them this way for me. They all share this common way, 
when they move off this way they then 'only' share the remaining ways 
with typically 2 relations, possibly 4.


It is all variable as to what is easiest and suits the local map and the 
local mappers.


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-27 Thread Martin Wynne



seen this done in various places, but I've never understood the point
it. The two representations are identical in terms of the data, but
the latter requires 2.5 times as many objects and is much more of a
pain to work with in the editors.


This happens a lot in my area. Huge areas of "farmland" have been 
created as massive multipolygons, which are too big to fit in the iD 
editor, and include ways shared with other areas such as equally large 
multipolygon woods. It's a pain to split them up without damaging them 
where they include areas which should be mapped as meadow, orchard, 
scrub, etc., which I much prefer to map as separate closed field areas, 
sometimes with their own name.


Likewise several woods are mapped as a single large multipolygon wood 
where in fact they are several separate woods each with a *name*. How 
can I apply names to parts of a multipolygon?


cheers,

Martin.

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-27 Thread David Woolley

On 27/06/2019 10:49, Robert Whittaker (OSM lists) wrote:

The two representations are identical in terms of the data, but
the latter requires 2.5 times as many objects and is much more of a
pain to work with in the editors. All to avoid having a common line
segment between two areas



I'd certainly say there should be a strong presumption against the use 
of relations.


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-27 Thread Robert Whittaker (OSM lists)
On Wed, 26 Jun 2019 at 21:08, Brian Prangle  wrote:
> The whole area needs simplification to replace multiple overlaid ways with 
> multipolygon relations.

I'm curious about what you mean here. Are you referring to replacing
(in a simple example) two square closed ways that share a common edge,
with three non-closed ways (two with three segments, and one with one)
and two relations to represent the original two squares? If so, I've
seen this done in various places, but I've never understood the point
it. The two representations are identical in terms of the data, but
the latter requires 2.5 times as many objects and is much more of a
pain to work with in the editors. All to avoid having a common line
segment between two areas -- which doesn't seem to be a particular
problem to me. Are there some advantages that I'm missing?

Thanks,

Robert.

-- 
Robert Whittaker

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-26 Thread Brian Prangle
I've changed this back to 3 reserves based on the unanimous opinion against
Martin's proposal. The whole area needs simplification to replace multiple
overlaid ways with multipolygon relations . I do have some knowledge of the
area having done 2 walking surveys there. There is also an SSSI  for part
of the site  which I'll need to work on further to see if and how it should
be mapped.

regards

Brian

On Mon, 10 Jun 2019 at 01:20, Warin <61sundow...@gmail.com> wrote:

> On 09/06/19 23:58, Martin Wynne wrote:
> >> we now have 2 natural=heaths named  as nature reserves and with
> >> operator tags but
> >> without nature reserve tags.
> >
> > Hi Adam,
> >
> > But they are now nested within a larger area which does have a nature
> > reserve tag. Much of the publicity material for this area treats it as
> > a single nature reserve.
>
> But the nature reserveS have different names and ownership and should be
> tagged as such.
>
> The heath is the common aspect between the two yet this is where the
> names and ownership are applied? This is clearly tagging for the render.
>
> And it fails
>
> https://www.openstreetmap.org/query?lat=52.3600=-2.2836
> Results in the wood and the combined nature reserve ... but no idea of
> which nature reserve applies here.
>
>
> The two nature reserves should be separate entries as nature reserves -
> which is what they are. Truth in tagging should be applied.
>
>
> --- Quibble
> There is also the aspect that the heath has a wood in it .. yet the
> heath covers the wood. The heath should be a multipolygon relation with
> an inner for the wood
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-09 Thread Warin

On 09/06/19 23:58, Martin Wynne wrote:
we now have 2 natural=heaths named  as nature reserves and with 
operator tags but

without nature reserve tags.


Hi Adam,

But they are now nested within a larger area which does have a nature 
reserve tag. Much of the publicity material for this area treats it as 
a single nature reserve.


But the nature reserveS have different names and ownership and should be 
tagged as such.


The heath is the common aspect between the two yet this is where the 
names and ownership are applied? This is clearly tagging for the render.


And it fails

https://www.openstreetmap.org/query?lat=52.3600=-2.2836
Results in the wood and the combined nature reserve ... but no idea of 
which nature reserve applies here.



The two nature reserves should be separate entries as nature reserves - 
which is what they are. Truth in tagging should be applied.



--- Quibble
There is also the aspect that the heath has a wood in it .. yet the 
heath covers the wood. The heath should be a multipolygon relation with 
an inner for the wood


___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-08 Thread Mateusz Konieczny



7 Jun 2019, 20:11 by mar...@templot.com:

> It's possible to see the property boundaries on old maps, but after visiting 
> the site again yesterday I can find little remaining physical evidence of the 
> boundaries, with many footpaths crossing between them.
>
> The web sites refer to them being seamlessly linked together with connecting 
> footpaths.
>
> So I have changed the OSM mapping to show a single nature reserve, with the 
> individual reserve names applied to the land parcels within it
>
So you would also delete it altogether in case where there is no border visible 
at all for outside way?

This is not how nature reserve mapping is typically done.

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-07 Thread Rob Nickerson
Hi Martin,

I echo Gareth, Mateusz and Warin's view that this should have stayed mapped
as separate nature reserves. The previous areas would have (hopefully)
marked the legal boundary of the individual nature reserves - something
which we have now lost. I see for example that the area you have now marked
as Burlish Top Nature Reserve is much smaller than it used to be.

Old:
https://overpass-api.de/achavi/?changeset=71023754

New:
https://www.openstreetmap.org/way/506211512

I'm also finding it very difficult to see what you are changing as you have
created so many changesets and they all have the same description (
https://www.openstreetmap.org/user/Martin%20Wynne/history#map=13/52.3652/-2.2799).
Would it please be possible to create fewer changesets and/or provide more
detailed changeset comments.

You may wish to ask Andy to help with reverting this as he will have plenty
of experience from his role in the Data Working Group. The fact that it is
over many changesets makes it too challenging for me.

Best regards,
*Rob*
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-07 Thread Martin Wynne

Thanks for the comments.

There are in fact 3 adjacent nature reserves with different names and 
ownerships.


It's possible to see the property boundaries on old maps, but after 
visiting the site again yesterday I can find little remaining physical 
evidence of the boundaries, with many footpaths crossing between them.


The web sites refer to them being seamlessly linked together with 
connecting footpaths.


So I have changed the OSM mapping to show a single nature reserve, with 
the individual reserve names applied to the land parcels within it:


 https://www.openstreetmap.org/#map=15/52.3647/-2.2802

This seems a better way of showing what is actually on the ground for 
visitors.



http://www.worcswildlifetrust.co.uk/reserves/the-devils-spittleful-rifle-range-and-blackstone-farm-fields


https://www.wyreforestdc.gov.uk/things-to-see-do-and-visit/countryside-and-nature/nature-reserves/burlish-top.aspx

cheers,

Martin.

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-06 Thread Warin
I am reminded of at least one single way I have edited (there could be 
more, it was some time ago)... it is a single way used for;


boundary of 2 states of Australia
boundary of 2 councils
boundary of 2 National Parks - note that these 'National Parks' are 
administered by the individual states and have different rules...


All of these are separate relations... with quite a few shared ways. 
Messy, but done.


The rendering looks good to me.

While you may have 2 nature reserves adjacent they both need to be on 
the map, so they can be individually found. So they should not be 
combined in the data base.


Rendering of boundaries of the same type .. but with less prominence?
Would not be high on my priority list... but doable. The render could 
should them with the same prominence as one single boundary.
Councils and countries usually share boundaries so they would have some 
thought to combining there boundary rendering.

There must be similar things between England/Scotland/Wales...


On 06/06/19 18:12, Mateusz Konieczny wrote:

5 Jun 2019, 19:55 by mar...@templot.com:

But on the OSM standard map, the common boundary is shown as a
bold green line, which bears no relation to anything on the ground
and could be misleading for visitors.

Note that maps are not aerial images - there is often significant 
level of abstraction and

especially for borders there is often nothing visible on the ground.

This rendering was used as compromise between several different problem.

Note also that the same styling applies to all nature reserves across 
the world.


Is there a better way to map this?

There are two nature reserves there, right?

If I combine them as a single nature reserve

sounds like tagging for the renderer - 
https://wiki.openstreetmap.org/wiki/Tagging_for_the_renderer


Is there a way to show the common boundary less prominently?

This is on side of renderers. This one has repository at
https://github.com/gravitystorm/openstreetmap-carto/
where proposals to improve it or pull requests with code improving it 
may be submitted.


Though again, the same rendering rules are applied globally, and for 
every single one
there are cases where it fails horribly. Improving one specific place 
may have really bad

results elsewhere.

On the data side - I would consider tagging borders on the shared way 
(mapping boundaries

as multipolygons), currently each nature reserve is a separate way.
https://www.openstreetmap.org/way/694760748#map=17/52.37217/-2.28169

(it would not change rendering, at least on default OSM map)



___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb



___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-06 Thread Mateusz Konieczny
5 Jun 2019, 19:55 by mar...@templot.com:

> But on the OSM standard map, the common boundary is shown as a bold green 
> line, which bears no relation to anything on the ground and could be 
> misleading for visitors.
>
Note that maps are not aerial images - there is often significant level of 
abstraction and
especially for borders there is often nothing visible on the ground.

This rendering was used as compromise between several different problem.

Note also that the same styling applies to all nature reserves across the world.

> Is there a better way to map this?
>
There are two nature reserves there, right?

>  If I combine them as a single nature reserve
>
sounds like tagging for the renderer - 
https://wiki.openstreetmap.org/wiki/Tagging_for_the_renderer 


> Is there a way to show the common boundary less prominently?
>
This is on side of renderers. This one has repository at
https://github.com/gravitystorm/openstreetmap-carto/ 

where proposals to improve it or pull requests with code improving it may be 
submitted.

Though again, the same rendering rules are applied globally, and for every 
single one
there are cases where it fails horribly. Improving one specific place may have 
really bad 
results elsewhere.

On the data side - I would consider tagging borders on the shared way (mapping 
boundaries
as multipolygons), currently each nature reserve is a separate way.
https://www.openstreetmap.org/way/694760748#map=17/52.37217/-2.28169

(it would not change rendering, at least on default OSM map)

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb


Re: [Talk-GB] Adjacent nature reserves

2019-06-05 Thread Gareth L
The bounds of an area don’t mean there’s a barrier there. But a nature reserve 
does render on that map in a similar way to a tree line or hedgerow would be 
rendered.

I’d leave it as it is. The problem is appears to how it renders, rather than 
how it is mapped. It could be totally fine with a different tile set. It’d be 
better to try and get the standard Osm map rendering scheme tweaked. They do it 
fairly frequently.

Gareth

> On 5 Jun 2019, at 18:56, Martin Wynne  wrote:
> 
> At this location there is a large area of open sandy heath, forming a nature 
> reserve:
> 
> https://www.openstreetmap.org/#map=16/52.3716/-2.2816
> 
> In fact it is two nature reserves side by side with different names and 
> ownership. One is charity-owned and managed by the county Wildlife Trust, the 
> other is owned and managed by the local District Council.
> 
> On the ground the boundary between them is barely visible, just odd bits of 
> old fencing in places, and footpaths criss-cross between them. The visitor 
> material tends to combine them as a single nature reserve, and that is how 
> most folks think of them:
> 
> http://www.worcswildlifetrust.co.uk/reserves/the-devils-spittleful-rifle-range-and-blackstone-farm-fields
> 
> The council's web site refers to them linking "seamlessly":
> 
> https://www.wyreforestdc.gov.uk/things-to-see-do-and-visit/countryside-and-nature/nature-reserves/rifle-range-sssi.aspx
> 
> But on the OSM standard map, the common boundary is shown as a bold green 
> line, which bears no relation to anything on the ground and could be 
> misleading for visitors.
> 
> Here's a picture of the boundary, running approx from 8 o'clock to 2 o'clock:
> 
> http://85a.uk/rifle_range_boundary_960x448.jpg
> 
> Is there a better way to map this? If I combine them as a single nature 
> reserve, is there a way to name the two parts of it separately? Is there a 
> way to show the common boundary less prominently?
> 
> Thanks,
> 
> Martin.
> 
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb