Re: [Tagging] RFC - Level:ref=*

[John Willis]

Javbw

> On Nov 30, 2015, at 6:05 AM, Simon Poole  wrote:
> 
> as an extra tag on POis that already has a level tag. I would rather be 
> looking for something that fits in with SIT 
> (http://wiki.openstreetmap.org/wiki/Simple_Indoor_Tagging) better. Since SIT 
> works really well (see OpenLevelUp),

TL;DR:

We're talking about two different things doing two different goals. 




I was thinking about this a bit more. 

Perhaps we are trying to do two different jobs. 

It looks like the SIT stuff is for rendering and actually displaying, in a 3D 
spatial way, a shop or location. 

That is not what I am trying to do. 

Just like it's zip code or phone number is not (sub-building) spatial 
information, but it is important information about the POI.  

OpenStreetMap is many things to many people - and much of my mapping is for 
replicating location level maps - a map of a park, a resort, or a tourist 
attraction. A person going to a park or a mall or a big building has searched 
and found a location (shop, viewpoint, attraction) and is shown that it is in a 
building that is difficult to navigate. I have searched for coffee shops, 
clothing stores, and restaurants in Tokyo, and the location is so complicated 
that the normally helpful POI in is not very helpful with them. I need more 
information - so I imagine most everyone else does too. Onsite navigation of 
very complicated complexes - like Tokyo Dome City or KitaSenju station or Tokyo 
Sky Tree - are dependent on on-site physical signs, information maps with ref 
#s on locations, and elevators with floor information for traveling around 
inside the POI. Also, printed location guides are common in Japan (at every 
door of every large mall), which are not common to see in the US. All of these 
signs and maps and guides use references *not* in the address - only refs and 
names on location level signage. 

I want to map this (sometimes crucial) information. 

In a vast majority of mundane POI search instances - visiting a lawyer's office 
in a big building - knowing it is on level 12 unit 1207 is just fine. 

But in complicated places where floors have names/refs that are more important 
than a level number, or when a "building" is a collection of various platforms 
and levels, not presented as a single building with levels, then providing that 
ref information with a POI pin info display would help the person navigate the 
last 200m through the structure to the actual location. 

This works with POI reference pins on the local information map. If I go to a 
mall, and I am told it is #42 on the mall's map and guidebook, it makes finding 
my way through the mall straightforward. Telling me it is unit 412 (it's 
mailing address) is useless, possibly counterproductive - if there is unit 
information in the official address at all. 
In some places, this offical address information is enough - Malls in 
California routinely show the unit number on the front window as the unit 
number is also in the mailing address - but that is not the case everywhere on 
earth. It seems Very Different here in Japan - I rarely see unit numbers for 
shops in malls in their address, hence the reliance on guides and maps and 
other physical onsite spatial aides. Which change very rarely (due to the huge 
cost of changing all the signage) 

How many years will it be until this SIT stuff will be able to guide me on the 
myriad of overlapping platforms, levels, and shops in KitaSenju station? 
Through the maze of walkways and tunnels in Horton Plaza in San Diego? The 
nightmare of overlapping pedestrian levels, footbridges, train stations, subway 
entrances, amusement parks, and malls that make up Tokyo Dome City? I bet it 
will be a really, really, really long time from now that it can correctly show 
me all the escalators and footbridges and multiple levels I have to traverse to 
get from place to place, let alone render the buildings in a way that shows me 
the proper connections for each transportation method and let me infer the path 
from the displayed information - which is how I use maps. I am not trying to 
get involved with this 3D placement. That is totally outside my interest, and I 
think the level:ref proposal is as well. 

Appending level:ref & unit:ref to needed 
POIs will help people searching for locations in crazy places right now. All we 
have to do is show it with the displayed POI information when searched. 

Anyways - the building rendering will still have to show floor numbers and refs 
eventually - so when your building rendering shows the floor name (observation 
deck) and my POI search told me the cafe is on the observation deck - then that 
sounds like a really good thing - but both are different sets of information. 

Javbw. 


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

Re: [Tagging] RFC - Level:ref=*

[Martin Koppenhoefer]

2015-12-01 21:33 GMT+01:00 Simon Poole :

>
> could have the tags added:
> level:ref=EG
> level:name=Erdgeschoss
>
>
> In a typical mall (were the non-floor plan variant is most  useful) you
> can easily have 20, 30 shops per level.



yes, I was not suggesting to add this to all shops in all cases. It should
likely go onto a "building level" object if present (indoor=level in your
proposal), or could be on some POIs just to have the reference, or could
only be added in buildings /parts where the mapper thinks that it is
noteworthy (e.g. not what you would expect or could infer with common
knowledge, e.g. level:name=Wintergarten).



>
>> Since SIT works really well (see OpenLevelUp),
>>
>
>
> The min_level-idea doesn't work AFAIK in all cases, where storey heights
> of adjacent buildings are different and they are connected by a bridge, and
> more general, it doesn't work where levels aren't simply stacked and
> uniform for the whole part/buidling but are inclined or have varying
> absolute elevations in different rooms (e.g. connected by a ramp). The
> min_level of which building should apply? I still suggest to use
> building_levels for the amount of all levels of that building part /
> building, not the concept of "building_levels=min_level of a neighbouring
> building until first level+amount of building levels of the part that is
> tagged".
>
> Also buildings with varying floor heights are not depictable (AFAIK).
>
>
> SIT is not intended as a replacement for S3DB which you seem to be
> implying and I'm not quite sure why you believe you can't model a
> connection between two floors potentially with different numbering schemes
> in two different buildings given that the numbering -is- local to the
> building or building:part.



Yes, you are right, I was referring to
http://wiki.openstreetmap.org/wiki/Simple_3D_buildings#Height
and in particular this picture:
http://wiki.openstreetmap.org/wiki/File:Minlevel.svg
Imagine the configuration in this picture was different: 7 floors on the
left and 9 floors on the right (lower ceiling height). Further imagine the
bridge had 2 floors: B1 with 5 meters height and
above it B2 with 3 m height.

Actually in SIT this situation is indeed depictable (but it will not define
the vertical position of these 2 floors, you'll need S3DB additionally). In
S3DB as far as I understand it, this situation is not clearly depictable
(because it is not clear which minlevel should apply to this bridge part).
You cannot refer to "non-existing"/missing building levels to define a
vertical position, because you don't know what height for these
non-existent levels should be asumed.

Don't get me wrong, your proposal looks good and works fine for most cases,
but it should have an additional/optional possibility to store absolute
elevation of levels as well (you do have height, but there should be an
optional ele suggestion IMHO). This would also solve inclined floors
(different ele on different nodes).

Also it is not clear what "height" refers to in case of a building level,
because currently the structure and installations/finishings are not
contemplated (you can see this also in the sketch).. May I suggest to
define it: distance from the finished surface of the floor of the
object/level to the finished surface of the floor level above. (This is the
standard definition for building levels, alternatively you could define raw
distances (but that's hardly surveyable).

Cheers,
Martin
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] RFC - Level:ref=*

[Simon Poole]

Am 30.11.2015 um 10:26 schrieb Martin Koppenhoefer:
>
> 2015-11-29 22:05 GMT+01:00 Simon Poole  >:
>
> I would rather be looking for something that fits in with SIT
> (http://wiki.openstreetmap.org/wiki/Simple_Indoor_Tagging) better.
>
>
>
> I don't see why this doesn't integrate. E.g. your example
> shop=supermarket
> level=0
>
> could have the tags added:
> level:ref=EG
> level:name=Erdgeschoss
>

In a typical mall (were the non-floor plan variant is most  useful) you
can easily have 20, 30 shops per level.

> This might look unneccessary at first glance (and in many cases it
> is), but there are cases with particular floor numbering and naming
> schemes where it is helpful.
>
>  
>
> Since SIT works really well (see OpenLevelUp),
>
>
>
> The min_level-idea doesn't work AFAIK in all cases, where storey
> heights of adjacent buildings are different and they are connected by
> a bridge, and more general, it doesn't work where levels aren't simply
> stacked and uniform for the whole part/buidling but are inclined or
> have varying absolute elevations in different rooms (e.g. connected by
> a ramp). The min_level of which building should apply? I still suggest
> to use building_levels for the amount of all levels of that building
> part / building, not the concept of "building_levels=min_level of a
> neighbouring building until first level+amount of building levels of
> the part that is tagged".
>
> Also buildings with varying floor heights are not depictable (AFAIK).

SIT is not intended as a replacement for S3DB which you seem to be
implying and I'm not quite sure why you believe you can't model a
connection between two floors potentially with different numbering
schemes  in two different buildings given that the numbering -is- local
to the building or building:part.

Simon


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

Re: [Tagging] RFC - Level:ref=*

[Martin Koppenhoefer]

2015-11-29 22:05 GMT+01:00 Simon Poole :

> I would rather be looking for something that fits in with SIT (
> http://wiki.openstreetmap.org/wiki/Simple_Indoor_Tagging) better.
>


I don't see why this doesn't integrate. E.g. your example
shop=supermarket
level=0

could have the tags added:
level:ref=EG
level:name=Erdgeschoss

This might look unneccessary at first glance (and in many cases it is), but
there are cases with particular floor numbering and naming schemes where it
is helpful.



> Since SIT works really well (see OpenLevelUp),
>


The min_level-idea doesn't work AFAIK in all cases, where storey heights of
adjacent buildings are different and they are connected by a bridge, and
more general, it doesn't work where levels aren't simply stacked and
uniform for the whole part/buidling but are inclined or have varying
absolute elevations in different rooms (e.g. connected by a ramp). The
min_level of which building should apply? I still suggest to use
building_levels for the amount of all levels of that building part /
building, not the concept of "building_levels=min_level of a neighbouring
building until first level+amount of building levels of the part that is
tagged".

Also buildings with varying floor heights are not depictable (AFAIK).

Some examples:
http://www.fk3.tu-braunschweig.de/blog/wp-content/uploads/2011/07/CloudClub-AWARD-2011_Anna-Katharina-Mielck_Schnitt.jpg
http://apb2013.de/publikumspreis/controller/thumbnailer/?query=Projekte/unkategorisiert/13/files/ze511_zanderroth_Schnitt_2.jpg=250=354
https://www.akbw.de/uploads/tx_templavoila/pt_19029_Zach_SchnittB.jpg
http://cdn2.world-architects.com/img/frontend/pages/1893/900:w/podest-berlin-schnitt.jpg
http://www.rundzwei.de/sites/default/files/styles/1120x640/public/projects/rd2_pr_web_026_02_schnitt_0.jpg?itok=jK9sgBAY

Cheers,
Martin
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] RFC - Level:ref=*

[John Willis]

> On Nov 30, 2015, at 6:05 AM, Simon Poole  wrote:
> 
> as an extra tag on POis that already has a level tag.

When a level has a name or reference that is used for navigation, there should 
be some way to tag it's level name. If indoor mapping can't handle that, then 
it is deficient. 

Addr=* always seems, to me, to be for official address information. 

If a shop is on "[height in Meters] 450" or "mezzanine" or "lobby" or 
"observation deck" - it is neither a floor=# nor addr:floor=* tag -

***but could be vital information for people on foot looking for the 
location*** 

Javbw. 
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] RFC - Level:ref=*

[Simon Poole]

This seems to lead to tons of unnecessary redundancy as proposed: as an
extra tag on POis that already has a level tag. I would rather be
looking for something that fits in with SIT
(http://wiki.openstreetmap.org/wiki/Simple_Indoor_Tagging) better. Since
SIT works really well (see OpenLevelUp), even without complicated indoor
elements, something that could be added to a building outline, or
building:part would see to be far better.

The addr:floor issue should be kept separate given that applies to
postal addresses that have different use cases and may different in any
case.

Simon

Am 29.11.2015 um 13:53 schrieb John Willis:
> Still looking for feedback on the idea. 
>
> Javbw
>
> On Nov 14, 2015, at 3:17 PM, johnw  > wrote:
>
>> I created an RFC page for level:ref=*
>>
>> I look forward to your comments. here or on the discussion page. 
>>
>> http://wiki.openstreetmap.org/wiki/Proposed_features/level:ref
>
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging



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

Re: [Tagging] RFC - Level:ref=*

[Richard]

On Sun, Nov 29, 2015 at 02:45:26PM +0100, Martin Koppenhoefer wrote:
> 
> 
> sent from a phone
> 
> > Am 29.11.2015 um 13:53 schrieb John Willis :
> > 
> > Still looking for feedback on the idea. 
> 
> 
> +1, I think it's good as is its brother level:name (at least as long as there 
> aren't proper level objects).


I would be in favor to untangle level:ref and use it only where level:name is 
not 
appropriate.

Richard



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

Re: [Tagging] RFC - Level:ref=*

[John Willis]

Still looking for feedback on the idea. 

Javbw

> On Nov 14, 2015, at 3:17 PM, johnw  wrote:
> 
> I created an RFC page for level:ref=*
> 
> I look forward to your comments. here or on the discussion page. 
> 
> http://wiki.openstreetmap.org/wiki/Proposed_features/level:ref
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] RFC - Level:ref=*

[Martin Koppenhoefer]

sent from a phone

> Am 29.11.2015 um 13:53 schrieb John Willis :
> 
> Still looking for feedback on the idea. 


+1, I think it's good as is its brother level:name (at least as long as there 
aren't proper level objects).

cheers 
Martin 
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] RFC - Level:ref=*

[Tobias Knerr]

On 14.11.2015 07:17, johnw wrote:
> http://wiki.openstreetmap.org/wiki/Proposed_features/level:ref

Thanks for the proposal! One question though: What's the rationale for
calling it level:ref, rather than level:name?



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

Re: [Tagging] RFC - Level:ref=*

[johnw]

> On Nov 15, 2015, at 12:48 AM, Tobias Knerr  wrote:
> 
> On 14.11.2015 07:17, johnw wrote:
>> http://wiki.openstreetmap.org/wiki/Proposed_features/level:ref
> 
> Thanks for the proposal! One question though: What's the rationale for
> calling it level:ref, rather than level:name?

That is a good point. 

I thought that a majority of floors mapped will look like refs: B3, W7, etc. 

Ref is short for Reference, and my brain went “this is how the floor is 
referenced in the floor maps and brochures”,

But you are right - most of the examples are are names of the floors.  

is ref still viable?

Javbw
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

[Tagging] RFC - Level:ref=*

[johnw]

I created an RFC page for level:ref=*

I look forward to your comments. here or on the discussion page. 

http://wiki.openstreetmap.org/wiki/Proposed_features/level:ref 



PS: I will create it’s cousins, addr:unit:ref later. having the 
directory/information map label value (#7 on the guide map) is very useful, 
especially if the actual addr:unit info is very different (unit E512). 



Javbw___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] RFC level:ref (was: Adding floor location information)

[johnw]

I created an RFC page for level:ref. 

I look forward to your comments. 

http://wiki.openstreetmap.org/wiki/Proposed_features/level:ref 




PS: I will create it’s cousins, addr:unit:ref later. having the 
directory/information map label value (#7 on the guide map) is very useful, 
especially if the actual addr:unit info is very different (unit E512). 



Javbw


> On Oct 1, 2015, at 4:28 PM, Martin Koppenhoefer  
> wrote:
> 
> I was referring to housenumbers, but would extend it to PO  boxes as well.
> 
> FWIW, I agree with your proposed tagging, in particular level:ref

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