Re: [Talk-ca] Building Import update

[OSM Volunteer stevea]

On Feb 1, 2019, at 1:13 PM, john whelan  wrote:
> So how would you tackle it?
> 
> Adding buildings with JOSM and the buildings_tool is possible, I think Julia 
> tried to whip up some interest with the 2020 project.  Unfortunately 
> mapathons using iD and new mappers for some reason don't work too well for 
> buildings. They do work fine for adding tags though.
> 
> I seem to recall March 2nd is some sort of student GIS day and we can expect 
> something to happen in GEO/GIS week whenever it is.  I'd prefer adding tags 
> to existing outlines rather than having to clean up buildings added with iD.

Adding tags to buildings by students at a "student GIS day" (whether with iD or 
not) is "one thing," and honestly shouldn't even be in this same thread 
("Building Import update." ) Conflating the two is either mistaken, 
disingenuous or both.

> If we go back in time to the Ottawa import and the licensing issues I seem to 
> recall a Toronto mapper submitting the Toronto Open Data License to the legal 
> working group which implies at least one Toronto OSM mapper was after the 
> Toronto Open Data.

While it can be valuable to "look in the rear view mirror" (to learn from past 
mistakes), I fail to see how this comment matters.  Doing my best to stay on 
point, my educated guess is there are MANY users who want to see "the five 
provincial building datasets" (what we TRULY attempt to discuss here) enter 
OSM.  However, there appear to be questions about the data quality, with some 
saying "skilled editors are able to do a decent job with these data, but not 
without substantial post-data publication (now) improvements before the data 
are uploaded."  (This would be downloading a "square" on the Task Manager and 
rather heavily improving the data, building by building.  Yaro and Danny, 
please chime in and agree or disagree.  Should that be true, only 
intermediate-to-advanced — i.e. rather skilled OSM editors with practice and 
experience should "do" the importation of these data).  Others say "these data 
need wholesale algorithmic changes before they are good enough to be uploaded." 
 (As I've said, maybe "squaring" or "simplification," yet I and this mail-list 
still do not know exactly where consensus lies there).  There are likely other 
opinions along and even outside of that spectrum, I simply do not know that.  
But GETTING to know the answers to those questions really must be "next."  We 
appear to be hashing that out here and now.  Much else (all else?) is, largely 
speaking, noise, distraction and what Nate said, "red herring."

> My feeling at the moment is there is a suggestion that "cleaning" the data up 
> then some sort of team approach in a particular area would be acceptable but 
> how you put it together I'm not sure.

No, you do not appear to be sure, John.  Yet, somehow, Canada will get there.  
Either in talk-ca, the Import wiki, the wiki's Discussion tab ("Talk page") the 
consensus of what to do with the data will EVENTUALLY emerge (fix 'em, scrap 
'em, leave 'em alone and import 'em with great care...), then they will 
(slowly, there are a LOT!) begin to be imported into OSM.  Or not.

It is a maxim of good project management which is often unstated, yet now is 
the time to say it:  "Lead, follow or get out of the way."

SteveA
___
Talk-ca mailing list
Talk-ca@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ca

Re: [Talk-ca] Building Import update

[john whelan]

So how would you tackle it?

Adding buildings with JOSM and the buildings_tool is possible, I think
Julia tried to whip up some interest with the 2020 project.  Unfortunately
mapathons using iD and new mappers for some reason don't work too well for
buildings. They do work fine for adding tags though.

I seem to recall March 2nd is some sort of student GIS day and we can
expect something to happen in GEO/GIS week whenever it is.  I'd prefer
adding tags to existing outlines rather than having to clean up buildings
added with iD.

If we go back in time to the Ottawa import and the licensing issues I seem
to recall a Toronto mapper submitting the Toronto Open Data License to the
legal working group which implies at least one Toronto OSM mapper was after
the Toronto Open Data.

My feeling at the moment is there is a suggestion that "cleaning" the data
up then some sort of team approach in a particular area would be acceptable
but how you put it together I'm not sure.

Suggestions please

Thanks

Cheerio John

On Fri, 1 Feb 2019 at 15:52, Begin Daniel  wrote:

> +1
>
>
>
> *From:* Nate Wessel [mailto:bike...@gmail.com]
> *Sent:* Friday, February 01, 2019 08:54
> *To:* talk-ca@openstreetmap.org
> *Subject:* Re: [Talk-ca] Building Import update
>
>
>
> John,
>
> IMO, this is a red herring and I think you must recognize that to at least
> some degree. Just like no one suggested we do 3700 import plans, no on has
> suggested that we not add buildings to OSM. The question is how, and if
> that "how" in part is an import, then what data, at what speed, by who, etc?
>
> We're not debating between "import" and "nothing" here. There were tens of
> thousands of carefully hand-mapped buildings in Toronto before you and a
> couple others rode in and quietly changed everything in the course of a
> week.
>
> I'd like to point out to you the interesting case of Kenton County
> Kentucky:
> https://www.openstreetmap.org/relation/361564
>
> Go ahead and zoom in and take a good look at that data. Poke around the
> rest of Northern Kentucky too while you're at it. Notice how good not only
> the building data is, but landuses, named places, etc. The only substantial
> import this area has ever seen is the TIGER road import of about a decade
> ago. By the time we started our Hamilton County building import (just north
> of the river), there were more than 150,000 buildings added by hand in the
> region already.
>
> I'm not saying this is the way Toronto/Canada needs to develop, but don't
> imply that it's impossible - it isn't.
>
> Cheers,
>
> Nate Wessel
> Jack of all trades, Master of Geography, PhD candidate in Urban Planning
> NateWessel.com 
>
> On 2/1/19 7:35 AM, john whelan wrote:
>
>
> https://www.openstreetmap.org/search?query=arthur%20mark%20drive%20port%20hope%20ontario#map=17/43.96262/-78.27069
>
>
>
> https://www.google.ca/maps/@43.9631101,-78.2732195,17.25z
>
>
>
> https://www.bing.com/maps?FORM=Z9LH3
>
>
>
> Port Hope Ontario is relatively obscure yet both Bing and google have
> buildings and neither company would spend the money dropping them in unless
> they saw a demand.
>
>
>
> A small sample but I'm sure that others are quite capable of looking
> locally for themselves.
>
>
>
>
>
> I'm a shades of grey person so to me there is no absolute need to have
> buildings in OpenStreetMap and I think different end users have different
> expectations.  I seem to recall osmand has a street only map which takes up
> less room on the device.  It's perfectly adequate for some users.
>
>
>
> I can make a case for both having them and not having any.  On the not
> having any way up there would be the buildings added by inexperienced
> mappers using iD often in HOT projects.  There are duplicates, strange
> shapes that bare no relation to any imagery, and city blocks marked as a
> single building.  On the having them side would be where can I get a coffee
> and wifi?
>
>
>
> There are many users of the map who would like to see buildings or more
> importantly have building information available in an electronic form.
>
>
>
> For Ottawa I think I can safely say the local mappers are happy with the
> imported buildings.  In OpenStreetMap there will always be a range of
> points of view.
>
>
>
> As you say it is for the local mappers to decide what they would like to
> do.  In this case is it difficult to define the local mappers.
>
>
>
> Cheerio John
>
>
>
>
>
>
>
>
>
> ___
>
> Talk-ca mailing list
>
> Talk-ca@openstreetmap.org
>
> https://lists.openstreetmap.org/listinfo/talk-ca
>
> ___
> Talk-ca mailing list
> Talk-ca@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-ca
>
___
Talk-ca mailing list
Talk-ca@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ca

Re: [Talk-ca] Building Import update

[Begin Daniel]

+1

From: Nate Wessel [mailto:bike...@gmail.com]
Sent: Friday, February 01, 2019 08:54
To: talk-ca@openstreetmap.org
Subject: Re: [Talk-ca] Building Import update


John,

IMO, this is a red herring and I think you must recognize that to at least some 
degree. Just like no one suggested we do 3700 import plans, no on has suggested 
that we not add buildings to OSM. The question is how, and if that "how" in 
part is an import, then what data, at what speed, by who, etc?

We're not debating between "import" and "nothing" here. There were tens of 
thousands of carefully hand-mapped buildings in Toronto before you and a couple 
others rode in and quietly changed everything in the course of a week.

I'd like to point out to you the interesting case of Kenton County Kentucky:
https://www.openstreetmap.org/relation/361564

Go ahead and zoom in and take a good look at that data. Poke around the rest of 
Northern Kentucky too while you're at it. Notice how good not only the building 
data is, but landuses, named places, etc. The only substantial import this area 
has ever seen is the TIGER road import of about a decade ago. By the time we 
started our Hamilton County building import (just north of the river), there 
were more than 150,000 buildings added by hand in the region already.

I'm not saying this is the way Toronto/Canada needs to develop, but don't imply 
that it's impossible - it isn't.

Cheers,
Nate Wessel
Jack of all trades, Master of Geography, PhD candidate in Urban Planning
NateWessel.com
On 2/1/19 7:35 AM, john whelan wrote:
https://www.openstreetmap.org/search?query=arthur%20mark%20drive%20port%20hope%20ontario#map=17/43.96262/-78.27069

https://www.google.ca/maps/@43.9631101,-78.2732195,17.25z

https://www.bing.com/maps?FORM=Z9LH3

Port Hope Ontario is relatively obscure yet both Bing and google have buildings 
and neither company would spend the money dropping them in unless they saw a 
demand.

A small sample but I'm sure that others are quite capable of looking locally 
for themselves.


I'm a shades of grey person so to me there is no absolute need to have 
buildings in OpenStreetMap and I think different end users have different 
expectations.  I seem to recall osmand has a street only map which takes up 
less room on the device.  It's perfectly adequate for some users.

I can make a case for both having them and not having any.  On the not having 
any way up there would be the buildings added by inexperienced mappers using iD 
often in HOT projects.  There are duplicates, strange shapes that bare no 
relation to any imagery, and city blocks marked as a single building.  On the 
having them side would be where can I get a coffee and wifi?

There are many users of the map who would like to see buildings or more 
importantly have building information available in an electronic form.

For Ottawa I think I can safely say the local mappers are happy with the 
imported buildings.  In OpenStreetMap there will always be a range of points of 
view.

As you say it is for the local mappers to decide what they would like to do.  
In this case is it difficult to define the local mappers.

Cheerio John






___

Talk-ca mailing list

Talk-ca@openstreetmap.org

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

Re: [Talk-ca] Building Import update

[Nate Wessel]

John,

IMO, this is a red herring and I think you must recognize that to at 
least some degree. Just like no one suggested we do 3700 import plans, 
no on has suggested that we not add buildings to OSM. The question is 
how, and if that "how" in part is an import, then what data, at what 
speed, by who, etc?


We're not debating between "import" and "nothing" here. There were tens 
of thousands of carefully hand-mapped buildings in Toronto before you 
and a couple others rode in and quietly changed everything in the course 
of a week.


I'd like to point out to you the interesting case of Kenton County Kentucky:
https://www.openstreetmap.org/relation/361564

Go ahead and zoom in and take a good look at that data. Poke around the 
rest of Northern Kentucky too while you're at it. Notice how good not 
only the building data is, but landuses, named places, etc. The only 
substantial import this area has ever seen is the TIGER road import of 
about a decade ago. By the time we started our Hamilton County building 
import (just north of the river), there were more than 150,000 buildings 
added by hand in the region already.


I'm not saying this is the way Toronto/Canada needs to develop, but 
don't imply that it's impossible - it isn't.


Cheers,

Nate Wessel
Jack of all trades, Master of Geography, PhD candidate in Urban Planning
NateWessel.com 

On 2/1/19 7:35 AM, john whelan wrote:

https://www.openstreetmap.org/search?query=arthur%20mark%20drive%20port%20hope%20ontario#map=17/43.96262/-78.27069

https://www.google.ca/maps/@43.9631101,-78.2732195,17.25z

https://www.bing.com/maps?FORM=Z9LH3

Port Hope Ontario is relatively obscure yet both Bing and google have 
buildings and neither company would spend the money dropping them in 
unless they saw a demand.


A small sample but I'm sure that others are quite capable of looking 
locally for themselves.



I'm a shades of grey person so to me there is no absolute need to have 
buildings in OpenStreetMap and I think different end users have 
different expectations.  I seem to recall osmand has a street only map 
which takes up less room on the device.  It's perfectly adequate for 
some users.


I can make a case for both having them and not having any.  On the not 
having any way up there would be the buildings added by inexperienced 
mappers using iD often in HOT projects.  There are duplicates, strange 
shapes that bare no relation to any imagery, and city blocks marked as 
a single building.  On the having them side would be where can I get a 
coffee and wifi?


There are many users of the map who would like to see buildings or 
more importantly have building information available in an electronic 
form.


For Ottawa I think I can safely say the local mappers are happy with 
the imported buildings.  In OpenStreetMap there will always be a range 
of points of view.


As you say it is for the local mappers to decide what they would like 
to do.  In this case is it difficult to define the local mappers.


Cheerio John




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

Re: [Talk-ca] Building Import update

[john whelan]

https://www.openstreetmap.org/search?query=arthur%20mark%20drive%20port%20hope%20ontario#map=17/43.96262/-78.27069

https://www.google.ca/maps/@43.9631101,-78.2732195,17.25z

https://www.bing.com/maps?FORM=Z9LH3

Port Hope Ontario is relatively obscure yet both Bing and google have
buildings and neither company would spend the money dropping them in unless
they saw a demand.

A small sample but I'm sure that others are quite capable of looking
locally for themselves.


I'm a shades of grey person so to me there is no absolute need to have
buildings in OpenStreetMap and I think different end users have different
expectations.  I seem to recall osmand has a street only map which takes up
less room on the device.  It's perfectly adequate for some users.

I can make a case for both having them and not having any.  On the not
having any way up there would be the buildings added by inexperienced
mappers using iD often in HOT projects.  There are duplicates, strange
shapes that bare no relation to any imagery, and city blocks marked as a
single building.  On the having them side would be where can I get a coffee
and wifi?

There are many users of the map who would like to see buildings or more
importantly have building information available in an electronic form.

For Ottawa I think I can safely say the local mappers are happy with the
imported buildings.  In OpenStreetMap there will always be a range of
points of view.

As you say it is for the local mappers to decide what they would like to
do.  In this case is it difficult to define the local mappers.

Cheerio John
___
Talk-ca mailing list
Talk-ca@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ca