Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-30 Thread Mateusz Konieczny via talk



Oct 30, 2020, 16:33 by talk@openstreetmap.org:

> A split polygon with only an outer MP is not an "area".
>
It is a valid multipolygon representing an area.

A bit pointless multipolygon and maybe something that
should be converted to version not
using relation, but it is a valid tagging. 

> There's a clue in the name 'MultiPolygon' there has to be more than one.
>
Strictly speaking it means that it CAN have more than one part.

> Splitting into two serves no purpose, adds no quality.
>
It does not make it invalid

> Incomplete MP relations are not beneficial to OSM quality.
>
MP with single outer  and 0 inners is not invalid.


(this does not make it desired and preferable, but in this
case accusing iD of producing invalid relations is
baseless and invalid)

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


Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-30 Thread Simon Poole


Am 30.10.2020 um 16:33 schrieb Dave F:



But anyway... Point slit stands: Why did iD take this authoritarian 
position.
Already pointed this out n-times now: because it synthesizes an area 
object type.



As has been noted other, editors don't make this assumption.


Other editors don't try to synthesize an area type.



A split polygon with only an outer MP is not an "area".


It is, you are really totally mistaken on this. Particularly if you are 
reusing ways that are parts of other MPs it is quite common to have an 
MP with a sole outer ring composed of multiple ways (aka a "split 
polygon"). That it is typically unnecessary in the case of a building 
doesn't make it invalid.


Simon





The correct solution to split polygons with tags on the ways is to 
rejoin those ways, not create a MP.



As I pointed out, the question is -when- to rejoin those ways.


As I pointed out, that's for the contributor to decide, not the editor.


 A MP with only one* outer is invalid.


Nope.


There's a clue in the name 'MultiPolygon' there has to be more than one.
Splitting into two serves no purpose, adds no quality. Entropy isn't 
beneficial for the OSM database.


Incomplete MP relations are not beneficial to OSM quality.

DaveF


OpenPGP_0x4721711092E282EA.asc
Description: application/pgp-keys


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


Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-30 Thread Dave F via talk

On 29/10/2020 06:41, Simon Poole wrote:

Am 29.10.2020 um 00:17 schrieb Dave F:
iD editor attracts a hell of a lot of "WTFs", doesn't it? I mean, 
even its most ardent fan must occasionally raise a Roger Moore eyebrow.


bhuousel has taken the presumptive decision that the contributor's 
desired end result will always be a MP relation. This is wrong, plain 
& simple (& quite arrogant). iD editor should provide tools to allow 
contributors to make their own decisions as easily as possible & not 
take them on their behalf.


I'm not sure why you believe Bryan has or had anything to do with that 
specific design decision, but he didn't, that happened a substantial 
time before he had any formal involvement.




Because he was the only one to reply to github queries (2018) on this 
subject. He closed the query & he talks in the first person: "I'm OK 
with this being hard to do in iD."


But anyway... Point slit stands: Why did iD take this authoritarian 
position.



As has been noted other, editors don't make this assumption.


Other editors don't try to synthesize an area type.



A split polygon with only an outer MP is not an "area".



The correct solution to split polygons with tags on the ways is to 
rejoin those ways, not create a MP.



As I pointed out, the question is -when- to rejoin those ways.


As I pointed out, that's for the contributor to decide, not the editor.


 A MP with only one* outer is invalid.


Nope.


There's a clue in the name 'MultiPolygon' there has to be more than one.
Splitting into two serves no purpose, adds no quality. Entropy isn't 
beneficial for the OSM database.


Incomplete MP relations are not beneficial to OSM quality.

DaveF

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


Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-29 Thread Simon Poole


Am 29.10.2020 um 00:17 schrieb Dave F:
iD editor attracts a hell of a lot of "WTFs", doesn't it? I mean, even 
its most ardent fan must occasionally raise a Roger Moore eyebrow.


bhuousel has taken the presumptive decision that the contributor's 
desired end result will always be a MP relation. This is wrong, plain 
& simple (& quite arrogant). iD editor should provide tools to allow 
contributors to make their own decisions as easily as possible & not 
take them on their behalf.


I'm not sure why you believe Bryan has or had anything to do with that 
specific design decision, but he didn't, that happened a substantial 
time before he had any formal involvement.



As has been noted other, editors don't make this assumption.


Other editors don't try to synthesize an area type.



The correct solution to split polygons with tags on the ways is to 
rejoin those ways, not create a MP.



As I pointed out, the question is -when- to rejoin those ways.

 A MP with only one* outer is invalid.


Nope.

* splitting it still means there's only one.

Relations were created to allow mapping of entities, not possible with 
just ways. They aren't meant to be the default for all objects.



See above.

Simon



DaveF

On 27/10/2020 08:11, Simon Poole wrote:
Its done that essentially since day one. As Bryce points out doing so 
keeps the object a valid "area" (and iD makes a valiant effort to 
stop you from breaking that).


It is also one of my favourite examples in talks why trying to keep 
things simple for the user is very difficult and some times 
counterproductive.


Lots of people have had the wtf moment when they come along a 
multi-polgon consisting of just one ring built from two ways. The 
problem is that once the user has split the polygon, there is no 
obvious point in time were you can be sure that the user is finished 
with it and you could simplify, particularly when you are trying to 
get the user to save often and early. So the simplification for the 
iD user comes at the expense of wtf's of everybody else.


Simon

Am 27.10.2020 um 02:05 schrieb Dave F via talk:

Hi

I don't use iD editor much, but I've just discovered it 
auto-converts closed polygons which are split (Shortcut Key = X) 
into MP relations.


I'm struggling to comprehend a logical reason for this. Is there 
one? If there's been a previous discussion which I've missed please 
post a link.


There's a couple of threads on iD's github issues, bhousel closed 
them with "wontfix - I Saw A Thing I Didn't Like (but is valid in OSM).


It may be valid, but is it desirable or helpful? I split closed 
ways, in P2, for various reasons without wanting them to be 
converted. How many newbies would even know what a MP relation is?


Having them as as split tagged ways is just as "valid". More so, 
considering splitting long ways is desirable.


DaveF




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


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




OpenPGP_0x4721711092E282EA.asc
Description: application/pgp-keys


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


Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-28 Thread Dave F via talk

On 27/10/2020 03:56, Bryce Cogswell wrote:

I agree it’s a little counterintuitive for experienced users but I understand 
the rationale: If you’re splitting a building (closed way) how is the result 
valid unless it’s converted to a multipolygon?


The editor is presuming the splitting of the way is the final & desired 
result of the contributor. This is wrong.


DaveF

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


Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-28 Thread Dave F via talk
iD editor attracts a hell of a lot of "WTFs", doesn't it? I mean, even 
its most ardent fan must occasionally raise a Roger Moore eyebrow.


bhuousel has taken the presumptive decision that the contributor's 
desired end result will always be a MP relation. This is wrong, plain & 
simple (& quite arrogant). iD editor should provide tools to allow 
contributors to make their own decisions as easily as possible & not 
take them on their behalf.


As has been noted other, editors don't make this assumption.

The correct solution to split polygons with tags on the ways is to 
rejoin those ways, not create a MP.


 A MP with only one* outer is invalid.

* splitting it still means there's only one.

Relations were created to allow mapping of entities, not possible with 
just ways. They aren't meant to be the default for all objects.


DaveF

On 27/10/2020 08:11, Simon Poole wrote:
Its done that essentially since day one. As Bryce points out doing so 
keeps the object a valid "area" (and iD makes a valiant effort to stop 
you from breaking that).


It is also one of my favourite examples in talks why trying to keep 
things simple for the user is very difficult and some times 
counterproductive.


Lots of people have had the wtf moment when they come along a 
multi-polgon consisting of just one ring built from two ways. The 
problem is that once the user has split the polygon, there is no 
obvious point in time were you can be sure that the user is finished 
with it and you could simplify, particularly when you are trying to 
get the user to save often and early. So the simplification for the iD 
user comes at the expense of wtf's of everybody else.


Simon

Am 27.10.2020 um 02:05 schrieb Dave F via talk:

Hi

I don't use iD editor much, but I've just discovered it auto-converts 
closed polygons which are split (Shortcut Key = X) into MP relations.


I'm struggling to comprehend a logical reason for this. Is there one? 
If there's been a previous discussion which I've missed please post a 
link.


There's a couple of threads on iD's github issues, bhousel closed 
them with "wontfix - I Saw A Thing I Didn't Like (but is valid in OSM).


It may be valid, but is it desirable or helpful? I split closed ways, 
in P2, for various reasons without wanting them to be converted. How 
many newbies would even know what a MP relation is?


Having them as as split tagged ways is just as "valid". More so, 
considering splitting long ways is desirable.


DaveF




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


___
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] id Editor auto-converts split polygons into MP relation

2020-10-27 Thread Simon Poole
Its done that essentially since day one. As Bryce points out doing so 
keeps the object a valid "area" (and iD makes a valiant effort to stop 
you from breaking that).


It is also one of my favourite examples in talks why trying to keep 
things simple for the user is very difficult and some times 
counterproductive.


Lots of people have had the wtf moment when they come along a 
multi-polgon consisting of just one ring built from two ways. The 
problem is that once the user has split the polygon, there is no obvious 
point in time were you can be sure that the user is finished with it and 
you could simplify, particularly when you are trying to get the user to 
save often and early. So the simplification for the iD user comes at the 
expense of wtf's of everybody else.


Simon

Am 27.10.2020 um 02:05 schrieb Dave F via talk:

Hi

I don't use iD editor much, but I've just discovered it auto-converts 
closed polygons which are split (Shortcut Key = X) into MP relations.


I'm struggling to comprehend a logical reason for this. Is there one? 
If there's been a previous discussion which I've missed please post a 
link.


There's a couple of threads on iD's github issues, bhousel closed them 
with "wontfix - I Saw A Thing I Didn't Like (but is valid in OSM).


It may be valid, but is it desirable or helpful? I split closed ways, 
in P2, for various reasons without wanting them to be converted. How 
many newbies would even know what a MP relation is?


Having them as as split tagged ways is just as "valid". More so, 
considering splitting long ways is desirable.


DaveF




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


OpenPGP_0x4721711092E282EA.asc
Description: application/pgp-keys


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


Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-27 Thread Shawn K. Quinn
On 10/26/20 22:56, Bryce Cogswell via talk wrote:
> I agree it’s a little counterintuitive for experienced users but I
> understand the rationale: If you’re splitting a building (closed way)
> how is the result valid unless it’s converted to a multipolygon?

JOSM has no issue splitting a closed way into two separate closed ways.
It's entirely possible the user wants to have two separate buildings
(happens frequently when MapWithAI/RapID mistakenly decides that two or
three (or more) closely packed buildings are one big building).

-- 
Shawn K. Quinn 
http://www.rantroulette.com
http://www.skqrecordquest.com

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


Re: [OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-26 Thread Bryce Cogswell via talk
I agree it’s a little counterintuitive for experienced users but I understand 
the rationale: If you’re splitting a building (closed way) how is the result 
valid unless it’s converted to a multipolygon?

> On Oct 26, 2020, at 6:09 PM, Dave F via talk  wrote:
> 
> Hi
> 
> I don't use iD editor much, but I've just discovered it auto-converts closed 
> polygons which are split (Shortcut Key = X) into MP relations.
> 
> I'm struggling to comprehend a logical reason for this. Is there one? If 
> there's been a previous discussion which I've missed please post a link.
> 
> There's a couple of threads on iD's github issues, bhousel closed them with 
> "wontfix - I Saw A Thing I Didn't Like (but is valid in OSM).
> 
> It may be valid, but is it desirable or helpful? I split closed ways, in P2, 
> for various reasons without wanting them to be converted. How many newbies 
> would even know what a MP relation is?
> 
> Having them as as split tagged ways is just as "valid". More so, considering 
> splitting long ways is desirable.
> 
> DaveF
> 
> 
> 
> 
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk


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


[OSM-talk] id Editor auto-converts split polygons into MP relation

2020-10-26 Thread Dave F via talk

Hi

I don't use iD editor much, but I've just discovered it auto-converts 
closed polygons which are split (Shortcut Key = X) into MP relations.


I'm struggling to comprehend a logical reason for this. Is there one? If 
there's been a previous discussion which I've missed please post a link.


There's a couple of threads on iD's github issues, bhousel closed them 
with "wontfix - I Saw A Thing I Didn't Like (but is valid in OSM).


It may be valid, but is it desirable or helpful? I split closed ways, in 
P2, for various reasons without wanting them to be converted. How many 
newbies would even know what a MP relation is?


Having them as as split tagged ways is just as "valid". More so, 
considering splitting long ways is desirable.


DaveF




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