Ground truth for non-physical objects

classic Classic list List threaded Threaded
48 messages Options
123
Reply | Threaded
Open this post in threaded view
|

Ground truth for non-physical objects

Tomas Straupis
Hello

  I think we should settle the question of how "ground truth" or
"verifiability" applies to NON-PHYSICAL objects (it is clear with
physical objects). Because currently I see at least two opinions:

  1. Non-physical objects are mapped by observing/verifying their
REFLECTION in physical world.

  2. Non-physical objects are mapped by observing/verifying them
DIRECTLY where they originate and live - in non-physical world -
~documents.

  It is very demotivating to hear the argument that "opinion X is your
personal opinion, but (my) opinion Y is how OpenStreetMap works"
without any evidence. Especially by people with not too much actual
mapping/usage experience (say < 100000 objects done, no
application/map created etc.). And without thinking about the impact
of it.

  Opinion 1 would mean that we should remove all(most?) non-physical
objects: country, state, county, city, suburb, national/regional park
boundaries (and a lot more) as most of that is unobservable on the
ground and sometimes reflection of small part of them on the ground is
misleading/outdated.

  Opinion 2 would mean that objects are mapped according to
originating documents. De facto situation is that almost all
non-physical objects are currently mapped according to documents.

  Which opinion is chosen has a huge impact on both participation and
usage of OpenStreetMap. Decision would be able to remove this burden
from OSMF which by definition should not be deciding on such matters.

P.S. Wiki while not being authoritative talks about PHYSICAL objects.
P.P.S. Let's skip non-physical attributes for the beginning.

--
Tomas

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

Re: Ground truth for non-physical objects

Frederik Ramm
Hi,

On 11.12.2018 10:37, Tomas Straupis wrote:
>   1. Non-physical objects are mapped by observing/verifying their
> REFLECTION in physical world.

...

>   Opinion 1 would mean that we should remove all(most?) non-physical
> objects: country, state, county, city, suburb, national/regional park
> boundaries (and a lot more) as most of that is unobservable on the
> ground and sometimes reflection of small part of them on the ground is
> misleading/outdated.

I think that we should not have a "fundamental" approach here but one of
pragmatism.

Non-physical (non-observable) things should definitely be the exception
in OSM, and it is my opinion that each class of non-physical things we
add needs a very good reason for adding them.

For example, certain historic facts are very well documented, sometimes
even by old maps or photographs, but we don't want them in OSM if they
are not visible on the ground any more. I think that this is the right
approach, and we normally don't want things that are not visible on the
ground.

We are making an exception, though, for some types of boundaries because
we think they are important enough to warrant this exception. Not only
important for map users, but also for the mapping process itself - for
example, boundaries could be important for our own statistics or for
knowing whether or not you are even allowed to go somewhere.

"Let's delete all boundaries" would certainly be an overreaction; "let's
require a very good reason for boundaries to have them in OSM" is
better. But "let's map things according to documents" is IMHO worse, and
you haven't even touched on the question of authority (whose documents
do you believe). You are right in saying that most current boundaries in
OSM are actually copied from documents, but we only do that where
everyone agrees that the documents actually depict the situation on the
ground. As soon as they are out of touch with the situation, we won't
consider documents a useful source any more.

Also, I think you are too fast in discounting the verifiability of
boundaries. Even in the absence of actual marked lines, fences, or
walls, you will often find the "reflections" that you speak of if you
look a bit closer: Which government do I pay my taxes to? Which police
department is responsible for my area? Which local authority do I get my
food stamps from, whatever.

Bye
Frederik

--
Frederik Ramm  ##  eMail [hidden email]  ##  N49°00'09" E008°23'33"

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

Re: Ground truth for non-physical objects

Christoph Hormann-2
On Tuesday 11 December 2018, Frederik Ramm wrote:
>
> Also, I think you are too fast in discounting the verifiability of
> boundaries. Even in the absence of actual marked lines, fences, or
> walls, you will often find the "reflections" that you speak of if you
> look a bit closer: Which government do I pay my taxes to? Which
> police department is responsible for my area? Which local authority
> do I get my food stamps from, whatever.

Indeed.

Note i have explained to Tomas in length the meaning of the concept of
verifiability for not directly physically manifested statements in

http://blog.imagico.de/verifiability-and-the-wikipediarization-of-openstreetmap/#comments

Using the example of a bus stop without signs or shelter i wrote:

> A bus stop, even one without a sign or shelter, can be verified by
> observing that a bus regularly stops at the location. There is
> nothing in the concept of independent verifiability that limits its
> application to physical objects.
>
> Ultimately most verifiable cultural geography features are related to
> human activities and can be verified by either observing these human
> activities themselves or physical effects of these activities.

--
Christoph Hormann
http://www.imagico.de/

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

Re: Ground truth for non-physical objects

Tomas Straupis
In reply to this post by Frederik Ramm
2018-12-11, an, 12:06 Frederik Ramm rašė:
> Non-physical (non-observable) things should definitely be the exception
> in OSM, and it is my opinion that each class of non-physical things we
> add needs a very good reason for adding them.

  I agree, but that is a different question. My suggestion is to
discuss this later as a separate topic so that initial question of
"what ground truth means" would not be buried. We could later have
either (preferably) a criteria, or (if criteria is not possible)
simply a table listing acceptable or not acceptable non-physical
objects in the database.

> Also, I think you are too fast in discounting the verifiability of
> boundaries. Even in the absence of actual marked lines, fences, or
> walls, you will often find the "reflections" that you speak of if you
> look a bit closer: Which government do I pay my taxes to? Which police
> department is responsible for my area? Which local authority do I get my
> food stamps from, whatever.

  Well, the first thing is to decide if boundaries as non-physical
objects originate in documents, or physical observation and which one
we use. Mixing those is what is introducing subjectivity and thus
different interpretation and problems.

  Then we can decide on priorities (if required at all). For example
for all boundaries (except country boundaries) there is a clear
candidate - local authority (government for administration division to
states, counties, cities, suburbs etc.), same local authority or some
national park administration whoever is deciding on official
boundaries of national/regional parks, protected areas etc.
  I cannot think of an example, where some important object worth
being in OpenStreetMap database would not have a single authority
deciding on its geometry.
  And this could work with country border only if we accept the
possibility of overlapping borders (which sometimes do exist even
without conflicts between countries).

  Tax, police does not look like a firm criteria because:
  1. You would need some documents to verify that anyway?
  2. Tax/police regions do not necessarily correspond to
administrative divisions and they could differ/overlap.

  Note that while it is relatively easy to spot a missing non-physical
object and then add it, it is much harder to notice a change of it. If
we would agree on using official documents it would allow to do such
checking by local community regularly (which does not necessarily mean
updating the data automatically by import, this could simply raise a
flag "please check here"). This is what is done in "some" countries
currently with ALL sides getting benefit and thus being a very good
selling point for OSM and now it is very disturbing to find it is
"against the old standing rules" :-)

--
Tomas

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

Re: Ground truth for non-physical objects

Tomas Straupis
In reply to this post by Christoph Hormann-2
> Note i have explained to Tomas in length the meaning of the concept of
> verifiability for not directly physically manifested statements in
>
> http://blog.imagico.de/verifiability-and-the-wikipediarization-of-openstreetmap/#comments
>
> Using the example of a bus stop without signs or shelter i wrote:
>
> > A bus stop, even one without a sign or shelter, can be verified by
> > observing that a bus regularly stops at the location. There is
> > nothing in the concept of independent verifiability that limits its
> > application to physical objects.

  This is a very good example of possibly misleading reflection.
  What if a driver is stopping in unofficial position somewhere
outside of large city to let local people he knows out/in even when
there is no official stop?
  What if a national park had a small sign in the forest track and the
sign was not moved when national park boundaries have moved?
  I had an actual situation 5 or so years ago when an address was
mapped in Vilnius. Address does not exist in official records. The
user sent me a picture of this house number. I contacted municipality
ant they explained that the sign is not an official one, it means
nothing, there is no such address.
  You can think of a gazillion of such examples and analysing them (in
my personal opinion) would lead to pointless endless discussions.
  The simpler the rules - the better?

  And in general. While it could be interesting to become some kind of
detectives and follow the leads, use deduction to calculate the
properties of non-physical object. Does it have to be
mandatory/primary way when there is a simpler and more correct way?
Isn't there enough of physical objects (or non-physical without
open/accessible/official documents) to observe, verify and map?

--
Tomas

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

Re: Ground truth for non-physical objects

Jochen123
On Tue, Dec 11, 2018 at 01:08:35PM +0200, Tomas Straupis wrote:
>   I had an actual situation 5 or so years ago when an address was
> mapped in Vilnius. Address does not exist in official records. The
> user sent me a picture of this house number. I contacted municipality
> ant they explained that the sign is not an official one, it means
> nothing, there is no such address.

It seems you haven't understood the on-the-ground rule 5 years ago and
you still haven't. For all intents and purposes there is such an
address. Mail will arrive there, people can find the house when looking
for it. It doesn't matter what the official record says. It doesn't
matter whether the address should be there or not according to some
authority. The address is there and it should be mapped that way. That
is what on-the-ground rule means. It works in practice. It works well.
And, yes, there are always corner cases. But that's no reason to
discredit the rule.

Jochen
--
Jochen Topf  [hidden email]  https://www.jochentopf.com/  +49-351-31778688

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

Re: Ground truth for non-physical objects

Tomas Straupis
2018-12-11, an, 13:27 Jochen Topf rašė:
> It seems you haven't understood the on-the-ground rule 5 years ago and
> you still haven't. For all intents and purposes there is such an
> address. Mail will arrive there, people can find the house when looking
> for it.

  Mail will not arrive there as mail will be stopped in post-office
because of incorrect (not existing) address.
  People will not find such address in any IT solution.
  Such "address" is the same as "the red house on the corner with
small pool in front of it".
  If someone puts a label "Military academy" on their house, would we
map it as an actual military academy?

  This whole topic is to clear up what "ground rule" actually means
for non-physical objects. But so far I'm not successful in avoiding
getting drowned in questionable micro-examples with subjective
explanations.

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

Re: Ground truth for non-physical objects

Victor Shcherb-2
In reply to this post by Jochen123
I think, the problem that rule says "on-the-ground" and if it doesn't mean on-the-ground and people cannot find it,  for example there is no sign at all like houses missing the number plate or abandonned houses or forest / national park divisions.

Indeed, mail address is one of the possibility to check but they are not consistent. Sending 2 emails to correct-like address could give contradicting results easily because they are human processed. 

There is no need to discredit the rule, especially where it couldn't applied, there is a need to enhance the rule for a non-physical objects which are mostly driven by documents. And OSM was always fine to accept these imports driven by municipality documents.


On Tue, 11 Dec 2018 at 12:28, Jochen Topf <[hidden email]> wrote:
On Tue, Dec 11, 2018 at 01:08:35PM +0200, Tomas Straupis wrote:
>   I had an actual situation 5 or so years ago when an address was
> mapped in Vilnius. Address does not exist in official records. The
> user sent me a picture of this house number. I contacted municipality
> ant they explained that the sign is not an official one, it means
> nothing, there is no such address.

It seems you haven't understood the on-the-ground rule 5 years ago and
you still haven't. For all intents and purposes there is such an
address. Mail will arrive there, people can find the house when looking
for it. It doesn't matter what the official record says. It doesn't
matter whether the address should be there or not according to some
authority. The address is there and it should be mapped that way. That
is what on-the-ground rule means. It works in practice. It works well.
And, yes, there are always corner cases. But that's no reason to
discredit the rule.

Jochen
--
Jochen Topf  [hidden email]  https://www.jochentopf.com/  +49-351-31778688

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk

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

Re: Ground truth for non-physical objects

ebel
In reply to this post by Tomas Straupis
On 11/12/2018 12:38, Tomas Straupis wrote:
>    If someone puts a label "Military academy" on their house, would we
> map it as an actual military academy?

No, but you would put "addr:housename=Military academy".

Sometimes governments won't put actual military installations on
"official maps", but you're free to map it as such if it's there
on-the-ground.


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

Re: Ground truth for non-physical objects

Tomas Straupis
2018-12-11, an, 16:41 Rory McCann rašė:
> On 11/12/2018 12:38, Tomas Straupis wrote:
>>    If someone puts a label "Military academy" on their house, would we
>> map it as an actual military academy?
>
> No, but you would put "addr:housename=Military academy".

  Well IF you know it is not actually a Military academy. But if
you're not allowed to enter it would be difficult to say.

  Take any different example, say banner says "Ministry of silly
walks", how would it be possible to decide if it is a real ministry or
not (if you're not allowed to enter) and decide if it is housename or
office=government without looking into official documents?

  Also following the same logic, number "3A" in my example could go to
addr:housename, but not addr:housenumber

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

Re: Ground truth for non-physical objects

Jmapb
In reply to this post by ebel
On 12/11/2018 9:41 AM, Rory McCann wrote:
> On 11/12/2018 12:38, Tomas Straupis wrote:
>>    If someone puts a label "Military academy" on their house, would we
>> map it as an actual military academy?
>
> No, but you would put "addr:housename=Military academy".

Sidebar, according to my reading of the address tagging standards, one
should only tag addr:housename=* when it's an official (or at least de
facto) part of the postal address. It's not for people who just decide
their house has a name and write that name on a sign -- though you can
use name=* for that. J


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

Re: Ground truth for non-physical objects

ebel
On 11/12/2018 16:49, Jmapb wrote:

> On 12/11/2018 9:41 AM, Rory McCann wrote:
>> On 11/12/2018 12:38, Tomas Straupis wrote:
>>>    If someone puts a label "Military academy" on their house, would we
>>> map it as an actual military academy?
>>
>> No, but you would put "addr:housename=Military academy".
>
> Sidebar, according to my reading of the address tagging standards, one
> should only tag addr:housename=* when it's an official (or at least de
> facto) part of the postal address. It's not for people who just decide
> their house has a name and write that name on a sign -- though you can
> use name=* for that. J

If you put a sign on your house and use that to direct people, and the
postal worker delivers you post, then it is de facto the name. In
Ireland house names, as opposed to numbers, are very common in rural
areas. There is no official registry.

On 11/12/2018 16:03, Tomas Straupis wrote:> 2018-12-11, an, 16:41 Rory
McCann rašė:
 >> No, but you would put "addr:housename=Military academy".
 >
 >    Well IF you know it is not actually a Military academy. But if
 > you're not allowed to enter it would be difficult to say.

Use your head. Does it look like a military academy? Or someone's house.
Do you see military students coming in and out.


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

Re: Ground truth for non-physical objects

Florian Lohoff-2
In reply to this post by Tomas Straupis
On Tue, Dec 11, 2018 at 01:08:35PM +0200, Tomas Straupis wrote:

>   This is a very good example of possibly misleading reflection.
>   What if a driver is stopping in unofficial position somewhere
> outside of large city to let local people he knows out/in even when
> there is no official stop?
>   What if a national park had a small sign in the forest track and the
> sign was not moved when national park boundaries have moved?
>   I had an actual situation 5 or so years ago when an address was
> mapped in Vilnius. Address does not exist in official records. The
> user sent me a picture of this house number. I contacted municipality
> ant they explained that the sign is not an official one, it means
> nothing, there is no such address.
There is no such address in municipal records. Does not mean a lot
in 95% of the World. In the other 5% of the World it means someone
is using an inofficial address. Still it works. So we document its
existance by mapping it. We are not a replica of some other dataset
which is as error prone as ours.

Flo
--
Florian Lohoff                                                 [hidden email]
        UTF-8 Test: The 🐈 ran after a 🐁, but the 🐁 ran away

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Ground truth for non-physical objects

Florian Lohoff-2
In reply to this post by Tomas Straupis
On Tue, Dec 11, 2018 at 01:38:37PM +0200, Tomas Straupis wrote:
> 2018-12-11, an, 13:27 Jochen Topf rašė:
> > It seems you haven't understood the on-the-ground rule 5 years ago and
> > you still haven't. For all intents and purposes there is such an
> > address. Mail will arrive there, people can find the house when looking
> > for it.
>
>   Mail will not arrive there as mail will be stopped in post-office
> because of incorrect (not existing) address.

Nope - Try it yourself. If you live in Examplestreet 10 and you start
sending yourself postcards with Examplestreet 10a they'll reach you.

Deutsche Post has its complete own address record set than the State
and if Examplestreet 1-100 go into a certain post delivery district
any of those addresses, existant or not will be on the Desk of the
right guy to take your postcard.

Otherwise it'll take 2 years for new construction areas to be able to
receive post - It works from day 1. It may take a day longer on the
first occurance of an address - but it works.

I used this for a couple years to taint address databases so i could
track who actually leaked the address to mail spam companies.

Flo
--
Florian Lohoff                                                 [hidden email]
        UTF-8 Test: The 🐈 ran after a 🐁, but the 🐁 ran away

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Ground truth for non-physical objects

Jóhannes Birgir Jensson
>>   Mail will not arrive there as mail will be stopped in post-office
>> because of incorrect (not existing) address.
>
> Nope - Try it yourself. If you live in Examplestreet 10 and you start
> sending yourself postcards with Examplestreet 10a they'll reach you.

Post offices often have a good time trying to decipher incorrect
addresses. Sometimes they even decipher a map [1] - other times they get
a name, town and country and go from there - playing detective can be
fun.


[1]
http://skessuhorn.is/2016/05/20/frumleg-aritun-sendibrefs-en-dugdi-tho/

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

Re: Ground truth for non-physical objects

Tomas Straupis
Discussions about mapping invented addresses shows exactly what I
wanted to say: we get drowned in endless pointless
counter-counter-examples of counter-examples. Rules would have to be
invented for addresses separately, and then separately for each
country or even more detailed. We once again get to the same old
example of reflections/shadows in the end of the cave.

Vilnius is not a large city, with 0,5M population it has only ~60K
addresses. Still EACH week ~50-100 addresses change (changes,
additions, deletions). I do not imagine how would it be possible to
capture all that "on the ground" without an army of mappers devoted
specifically to this very boring and uninteresting but useful class -
addresses.

Correct me if I'm wrong, but regions (larger than 1 square km) with
best (accurate and up to date) address coverage are the ones which use
official address registries.

P.S. I agree that when there is no open official source, physical
observation is the only thing we have.

--
Tomas

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

What is OSM for? (was: Re: Ground truth for non-physical objects)

Andy Townsend
On 12/12/2018 13:05, Tomas Straupis wrote:
> ... I do not imagine how would it be possible to
> capture all that "on the ground" without an army of mappers devoted
> specifically to this very boring and uninteresting but useful class -
> addresses.

If you're looking for a project that essentially mirrors "official" data
without actually checking that its valid then OpenStreetMap might not be
the project for you.

What makes OSM unque and better than the alternatives is that the data
in it is, where possible, verified by people on the ground.  In a sense
it's the "anti-wikipedia" - original research is not just allowed it's
positively encouraged.  Only this original research will catch corner
cases like the house that has a name (but that name isn't in any way
"official") that still gets mail delivered to it using that name (like
the house that I'm sat in right now, actually).

Obviously different OSM communities in different regions differ over how
much they want to rely on "official" data* - indeed some different
regions within the same country have argued about this in the past, but
the general view, which I think we can see from the balance of the posts
in this thread, is that most people back the "on the ground" principle -
if there's a housename that looks like looks like a house name, it's a
house name, even if it's not in an "official" list.

Best Regards,

Andy

* for the avoidance of doubt here I'm talking about "official data"
outside of any conflict or dispute.

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

Re: Ground truth for non-physical objects

Florian Lohoff-2
In reply to this post by Tomas Straupis
On Wed, Dec 12, 2018 at 03:05:06PM +0200, Tomas Straupis wrote:
> Discussions about mapping invented addresses shows exactly what I
> wanted to say: we get drowned in endless pointless
> counter-counter-examples of counter-examples. Rules would have to be
> invented for addresses separately, and then separately for each
> country or even more detailed. We once again get to the same old
> example of reflections/shadows in the end of the cave.

I think you are down the wrong path. There is no such thing as
an invented address.

Either it is an address somebody with enough knowledge of the addressing
scheme is able to find or not.

In the western hemisphere we have had addressing for like ages and we
have regulations which make the state and their bodys the source
of the official addresses. Nevertheless it is pretty common even in
Germany to put up a 10a 10b 10c on your shed, barn or garage. These
addresses while not been issued by an official body end up in all the
other address sets. You will even be able to get them into your national
id card because the "Meldewesen" (State body for registering residents)
is not linked to the "Katasteramt" (Geo body on the county level
issueing addresses). I know that because i have caused ~100 residents to
get new id cards because they all had a wrong street name in their ID.

I have been working for some years with addresses from Deutsche Telekom
and they differ from the state issued addresses by approx. ~5%. Telekom
itself "invents" addresses for difficult to describe locations etc,
lists your barn as 10b etc. They do so because there might be no
official address, residents describe the location with that address, or
they simply need to describe the location they put up a connection.

So there is no such thing as an invented address. An address is
something people will be able to find with knowledge about the
addresses scheme.

At least in Germany we might have 95% of the addresses beeing officially
issued but the other 5% of addresses in use are unknown to the
"Katasteramt" because people use the addressing scheme to put up new
Housenumbers whenever they see fit.

You will not find 2 address datasets without a significant difference.

The question now is - What is Openstreetmap? Are we a copy of one of the
datasets? Of which one? Are we trying to merge datasets? Are we having
rules what to add?

At least in some German "Bundesländer" we have had access to state
issued address information and we used it to add a signifikant amount of
addresses into OSM. Then we added stuff we observed on the ground which
we did not initially get from the "Landesvermessungsamt".

Do we now have a better or worse dataset?

IMHO we have a much better dataset because we are able to geocode stuff
people expect to be addressable which official bodys cant
address/geocode.

Official address datasets are as incomplete, broken, buggy as all other
datasets. There is no such thing as the one and only truth.

Flo
--
Florian Lohoff                                                 [hidden email]
        UTF-8 Test: The 🐈 ran after a 🐁, but the 🐁 ran away

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Ground truth for non-physical objects

Tomas Straupis
Germany is not the "whole world". If you have multiple datasets for addresses then you have to decide, and physical check could be the solution for your country because of registry collision, whatever German community decides.

In Lithuania there is one and only one official source for ANY official dataset. Process, owner and access is approved by law.

With all due respect to Germany and ordnung, why a country with strict and not conflicting data should be bound to vague solution because of some other country which does not have such a solution?

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

Re: Ground truth for non-physical objects

Florian Lohoff-2
On Wed, Dec 12, 2018 at 05:56:51PM +0200, Tomas Straupis wrote:
> Germany is not the "whole world". If you have multiple datasets for
> addresses then you have to decide, and physical check could be the solution
> for your country because of registry collision, whatever German community
> decides.
>
> In Lithuania there is one and only one official source for ANY official
> dataset. Process, owner and access is approved by law.

Exactly - there is one OFFICIAL source - There are hundrets of others.


Where is the body removing numbers just beeing put up by residents
because they see fit?

Does the police or other state body remove the physical address
from the public space? I dont know of any country where this happens
which means there is an OFFICIAL dataset which is typically to sparse
to be usable by people.

Flo
--
Florian Lohoff                                                 [hidden email]
        UTF-8 Test: The 🐈 ran after a 🐁, but the 🐁 ran away

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk

signature.asc (849 bytes) Download Attachment
123