[Tagging] Basic philosophy of OSM tagging

[Warin]

This comes from the tap discussion but has implications elsewhere.

What is the basic philosophy of OSM tagging at the top level?

Are 'we' tagging for

What things are? eg highways

OR

What things are used for? eg amenity


Explanation? By example;

Highways are used for transport so would be better tagged as 
transport=motorway, sub tags for vehicles etc.

OR

amenity=drinking_water would be better tagged as water=blubber

--
Is there an FAQ on this? Or has this never been documented/though of?
Have fun with this  :)




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

Re: [Tagging] Basic philosophy of OSM tagging

[David Bannon]

On Wed, 2015-01-14 at 11:28 +1100, Warin wrote:
> What is the basic philosophy of OSM tagging at the top level?
> 
> Are 'we' tagging for
> What things are? eg highways OR What things are used for? eg amenity

I think its a very good question Warin. Perhaps, at the hart of much
angst amongst OSM'ers. To complicate, should we tag for what things are
used for OR what they are officially intended for ?

Many bush roads in Australia are good examples, perhaps initially cut by
oil and gas explorers, adopted by recreational 4WD'ers, become tourist
roads. In many cases, never maintained, not gazetted as official roads.

I know a road, put in to maintain a railway line, that does not appear
on any official map but is an excellent way to access an iconic
destination, Chambers Pillar. Last time I used it there were no signs to
deny access but tourists are kept unaware of it. Should I map it next
time I'm there ?

David




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

Re: [Tagging] Basic philosophy of OSM tagging

[Frederik Ramm]

Hi,

On 01/14/2015 01:28 AM, Warin wrote:
> What is the basic philosophy of OSM tagging at the top level?

There is no basic philosophy at the top level from which everything else
can be derived. It's like evolution - some things are a bit strange but
you can often understand them by looking at how they came to be.

There is a tendency however to tag for

> What things are? eg highways

simply because something can always have side effects that are not
related to the primary purpose, or the primary purpose is not
immediately obvious.

For example, a motorway is not only a transport feature, it is also an
insurmountable barrier for pedestrians or cyclists.

Tagging is very often based on what you see, not what you know. If you
see a body of water (and you might be doing that from aerial imagery,
sitting 1000s of miles away), you tag it as a body of water even if you
don't know whether this is an artificial reservoir that supplies
drinking water or a crater lake or anything else. Tagging

> What things are used for? eg amenity

might require more knowledge than the mapper has, especially in the case
of mapping from aerial imagery.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

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

Re: [Tagging] Basic philosophy of OSM tagging

[Warin]

On 14/01/2015 11:00 PM, tagging-requ...@openstreetmap.org wrote:

Date: Wed, 14 Jan 2015 07:59:49 +0100
From: Frederik Ramm
To:tagging@openstreetmap.org
Subject: Re: [Tagging] Basic philosophy of OSM tagging
Message-ID:<54b613e5.8020...@remote.org>
Content-Type: text/plain; charset=UTF-8

Hi,

On 01/14/2015 01:28 AM, Warin wrote:

>What is the basic philosophy of OSM tagging at the top level?

There is no basic philosophy at the top level from which everything else
can be derived. It's like evolution - some things are a bit strange but
you can often understand them by looking at how they came to be.

There is a tendency however to tag for


>What things are? eg highways

simply because something can always have side effects that are not
related to the primary purpose, or the primary purpose is not
immediately obvious.

For example, a motorway is not only a transport feature, it is also an
insurmountable barrier for pedestrians or cyclists.

Tagging is very often based on what you see, not what you know. If you
see a body of water (and you might be doing that from aerial imagery,
sitting 1000s of miles away), you tag it as a body of water even if you
don't know whether this is an artificial reservoir that supplies
drinking water or a crater lake or anything else. Tagging


>What things are used for? eg amenity

might require more knowledge than the mapper has, especially in the case
of mapping from aerial imagery.

Bye
Frederik


I like this. I'm not after changing what has happened in the past but 
adopting an approach that will help future new tags. It may also 
influence future 'tiding' efforts, like those for power infrastructure? 
Don't know but I'd like some guidance as to the suggestion by some on 
the tap tag that it should be somewhere else ..where? And what are the 
'rules'?
From my 'diary' entry .. amenity and man_made are not liked ... and I 
can see the thinking from your post Frederik. Thanks. I'll put a bit of 
it up on my Diary .. might get more comments. The rest of you can think 
about it .. I know it is an essential question with a basic answer that 
will help guide in the future.


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

Re: [Tagging] Basic philosophy of OSM tagging

[johnw]

I’m a newcomer, and somewhat of a noob, But I’ll take a crack at it.:

 **   We are drawing existence, and tagging purpose, usage, and metadata - with 
a varying balance of importance between those 3 things. **

There are some caveats - it needs to stay put for a long time, and it needs to 
be such a size that a point, way, area, or relation can be used to accurately 
enough describe it’s existence (tagging poodles is out of the question, 
unfortunately, and a point cannot accurately define a road). 

A road exists, and we then tag what purpose the road has. It’s usage and 
construction informs it’s classification and descriptor tags, and it’s name, 
ref and other metadata are tagged, along with it’s meta-meta data of what 
system the road belongs to in a larger context (relations). 

usually for man-made features, what exists is there because it’s purpose is why 
it exists. Things get fuzzy in edge cases when something is reused (eg: a 
business in an old church building), but so many things are fragile enough that 
disuse means destruction and repurposing of the land - or its current disuse 
and decay are the tags to use when showing that it exists (for the time being). 

Trouble arises since we all should use the same tag values and definitions, 
which is somewhat impossible because taggers are tagging the world as they feel 
it exists (and what purpose it has), and at the detail level they are 
comfortable or familiar with, which gives rise to issues between taggers, and 
between regions, as people from different regions see the world and define 
their world a bit differently. 

an overly long example: 

For example, here in Japan, they have a relatively rigid definition in OSM of 
what a primary and secondary road are - which in many cases has absolutely 
nothing to do with construction or usage. for the most part, they have no 
bearing to if they are truly primary or secondary roads, but if they are 
legally a certain type of road for political reasons (who maintains them, who 
owns them) - in some cases they are old, narrow, and meandering roads that used 
to be a major route (100 years ago) - but are now bypassed by newer and newer 
“bypass" roads meant exclusively for cars, with modern standards (like I would 
find in California).  Most bypass roads are considered “tertiary” - only 
because they are not in the same legal classification as the older, more 
“important” roads - though they are better in almost every single measurable 
way than the road they are bypassing. 

There is a “primary” road near my house that is thinner than many alleys (less 
than 2.5m in one spot) , has an awkward level crossing impassible by trucks, 
bridged over by the trunk road (it doesn’t connect), and I wouldn’t recommend 
using it for any reason. The nearby “tertiary” and secondary roads are a 
superior choice - and connect to the trunk roads - and even have painted center 
line(!) -  the OSM:JA definition of a tertiary road - which doesn’t apply to 
the secondary or primary roads.

Similarly, Large primary roads legally take turns at intersections (they are 
almost never straight inside a city), and even though the road itself continues 
on straight in an identical manner (lanes, width, traffic, standards, etc), it 
becomes a tertiary road - and then intersects with several narrow, underused 
secondary roads! The larger, 4 lane ‘tertiary" road that handles 5 times the 
vehicle traffic, traveling on to connect with 2 major trunk roads -  intersects 
the narrow two lane “secondary road”  that is one of the small roads coming 
down from the “suburbs” into the city. 

But this is the way Japanese people *expect* their maps to be portrayed, so I 
have to accept that this balance of metadata (the ref # on the sign) is more 
important to them than usage when it comes to road classification above 
unclassified. 

In other places, the primary road I mentioned would be classified as an 
“unclassified” road, and the tertiary as a primary - but it is inconsistent 
with what Japanese mappers expect, so “shoganai” - it can’t be helped. 


Figuring out this *balance* between purpose, usage, and metadata is a 
difficult, almost impossible task - not to mention how to organize the tags in 
a human usable and machine parseable manner (go-go tagging mailing list!), but 
the sentence seems to encapsulate OSM pretty well. :

 **   We are drawing existence, and tagging purpose, usage, and metadata - with 
a varying balance of importance between those 3 things. **

Javbw.


> On Jan 14, 2015, at 9:28 AM, Warin <61sundow...@gmail.com> wrote:
> 
> This comes from the tap discussion but has implications elsewhere.
> 
> What is the basic philosophy of OSM tagging at the top level?
> 
> Are 'we' tagging for
> 
> What things are? eg highways
> 
> OR
> 
> What things are used for? eg amenity
> 
> 
> Explanation? By example;
> 
> Highways are used for transport so would be better tagged as 
> transport=motorway, sub t

Re: [Tagging] Basic philosophy of OSM tagging

[Mateusz Konieczny]

> 4 lane ‘tertiary" road that handles 5 times the vehicle traffic,
traveling on to connect with 2 major trunk roads -
> intersects the narrow two lane “secondary road”  that is one of the small
roads coming down from the “suburbs” into the city

Interesting. I was unaware about so drastic changes in meaning of tag by
editors. This one cripples Japanese road data -
and I am unsure what could be done about it.

2015-01-15 9:13 GMT+01:00 johnw :

> I’m a newcomer, and somewhat of a noob, But I’ll take a crack at it.:
>
>  **   We are drawing existence, and tagging purpose, usage, and metadata -
> with a varying balance of importance between those 3 things. **
>
> There are some caveats - it needs to stay put for a long time, and it
> needs to be such a size that a point, way, area, or relation can be used to
> accurately enough describe it’s existence (tagging poodles is out of the
> question, unfortunately, and a point cannot accurately define a road).
>
> A road exists, and we then tag what purpose the road has. It’s usage and
> construction informs it’s classification and descriptor tags, and it’s
> name, ref and other metadata are tagged, along with it’s meta-meta data of
> what system the road belongs to in a larger context (relations).
>
> usually for man-made features, what exists is there because it’s purpose
> is why it exists. Things get fuzzy in edge cases when something is reused
> (eg: a business in an old church building), but so many things are fragile
> enough that disuse means destruction and repurposing of the land - or its
> current disuse and decay are the tags to use when showing that it exists
> (for the time being).
>
> Trouble arises since we all should use the same tag values and
> definitions, which is somewhat impossible because taggers are tagging the
> world as they feel it exists (and what purpose it has), and at the detail
> level they are comfortable or familiar with, which gives rise to issues
> between taggers, and between regions, as people from different regions see
> the world and define their world a bit differently.
>
> an overly long example:
>
> For example, here in Japan, they have a relatively rigid definition in OSM
> of what a primary and secondary road are - which in many cases has
> absolutely nothing to do with construction or usage. for the most part,
> they have no bearing to if they are truly primary or secondary roads, but
> if they are legally a certain type of road for political reasons (who
> maintains them, who owns them) - in some cases they are old, narrow, and
> meandering roads that used to be a major route (100 years ago) - but are
> now bypassed by newer and newer “bypass" roads meant exclusively for cars,
> with modern standards (like I would find in California).  Most bypass roads
> are considered “tertiary” - only because they are not in the same legal
> classification as the older, more “important” roads - though they are
> better in almost every single measurable way than the road they are
> bypassing.
>
> There is a “primary” road near my house that is thinner than many alleys
> (less than 2.5m in one spot) , has an awkward level crossing impassible by
> trucks, bridged over by the trunk road (it doesn’t connect), and I wouldn’t
> recommend using it for any reason. The nearby “tertiary” and secondary
> roads are a superior choice - and connect to the trunk roads - and even
> have painted center line(!) -  the OSM:JA definition of a tertiary road -
> which doesn’t apply to the secondary or primary roads.
>
> Similarly, Large primary roads legally take turns at intersections (they
> are almost never straight inside a city), and even though the road itself
> continues on straight in an identical manner (lanes, width, traffic,
> standards, etc), it becomes a tertiary road - and then intersects with
> several narrow, underused secondary roads! The larger, 4 lane ‘tertiary"
> road that handles 5 times the vehicle traffic, traveling on to connect with
> 2 major trunk roads -  intersects the narrow two lane “secondary road”
> that is one of the small roads coming down from the “suburbs” into the city.
>
> But this is the way Japanese people *expect* their maps to be portrayed,
> so I have to accept that this balance of metadata (the ref # on the sign)
> is more important to them than usage when it comes to road classification
> above unclassified.
>
> In other places, the primary road I mentioned would be classified as an
> “unclassified” road, and the tertiary as a primary - but it is inconsistent
> with what Japanese mappers expect, so “shoganai” - it can’t be helped.
>
>
> Figuring out this *balance* between purpose, usage, and metadata is a
> difficult, almost impossible task - not to mention how to organize the tags
> in a human usable and machine parseable manner (go-go tagging mailing
> list!), but the sentence seems to encapsulate OSM pretty well. :
>
>  **   We are drawing existence, and tagging purpose, usage, and metadata -
> with a varyin

Re: [Tagging] Basic philosophy of OSM tagging

[Kotya Karapetyan]

Now that the water_tap proposal discussion is over, I'd like to join
this important discussion.

My opinion: Since OSM is a *map*, we should *map* things. That means,
we should tag what actually exists on the planet, not what is implied.
Sometimes things are tagged in real life. For example, motorways are
marked with special traffic signs, therefore we can tag a road as
motorway. In other cases, we should use common sense to call the
things by their names. I am usually asking myself: How would I explain
to a tourist how to find his way? I will use something like "Pass by
..." or "You will see ...". This "..." is then the name (hence tag) to
be used.

A good example of the contrary is amenity=drinking_water. Though the
original intention is clear, I believe the solution is suboptimal: a
tourist wouldn't know what to search for (since it is not drinking
water but rather its source that is actually visible in a given
place). A mapper may also have hard times identifying whether a
specific water source provides potable water or not.

So, my answer would be we should map what the things *appear to be*.
Taking the example of Japanese roads, I would also add "with
reasonably common knowledge". It does leave some space for
uncertainty, but this uncertainty is also present in real life, so it
can appear in OSM as well.


Warin's question also identifies a problem I'd like to discuss. There
seem to be no "formal" agreements on how to create OSM. Things are
documented in the wiki, which is subject to uncontrolled changes and
no review and which is not always read by the mappers. Data is then
used by software development companies in the way they find
reasonable, without any foundation for consistency. It may be cultural
but I am looking for some sort of more robust, maybe even enforced,
agreements. They may be subject to changes, but their mere existence
would help. What do you think?


Cheers,
Kotya




On Wed, Jan 14, 2015 at 1:28 AM, Warin <61sundow...@gmail.com> wrote:
> This comes from the tap discussion but has implications elsewhere.
>
> What is the basic philosophy of OSM tagging at the top level?
>
> Are 'we' tagging for
>
> What things are? eg highways
>
> OR
>
> What things are used for? eg amenity
>
> 
> Explanation? By example;
>
> Highways are used for transport so would be better tagged as
> transport=motorway, sub tags for vehicles etc.
>
> OR
>
> amenity=drinking_water would be better tagged as water=blubber
>
> --
> Is there an FAQ on this? Or has this never been documented/though of?
> Have fun with this  :)
>
>
>
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging

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

Re: [Tagging] Basic philosophy of OSM tagging

[Marc Gemis]

By forced rules: you mean a committee that decides what gets mapped and how
?
So when I want to map something now, I have to file a request to the
committee to start looking for a new tag. And if they like the request they
come back within a few months with a proposal. And this committee is
all-knowing, so they know all the exceptions in the different countries ?
So I don't have to ask for an update when they misunderstood me ?

wrote this half-seriously, half-jokingly

regards

m

On Thu, Jan 15, 2015 at 5:08 PM, Kotya Karapetyan 
wrote:

> Now that the water_tap proposal discussion is over, I'd like to join
> this important discussion.
>
> My opinion: Since OSM is a *map*, we should *map* things. That means,
> we should tag what actually exists on the planet, not what is implied.
> Sometimes things are tagged in real life. For example, motorways are
> marked with special traffic signs, therefore we can tag a road as
> motorway. In other cases, we should use common sense to call the
> things by their names. I am usually asking myself: How would I explain
> to a tourist how to find his way? I will use something like "Pass by
> ..." or "You will see ...". This "..." is then the name (hence tag) to
> be used.
>
> A good example of the contrary is amenity=drinking_water. Though the
> original intention is clear, I believe the solution is suboptimal: a
> tourist wouldn't know what to search for (since it is not drinking
> water but rather its source that is actually visible in a given
> place). A mapper may also have hard times identifying whether a
> specific water source provides potable water or not.
>
> So, my answer would be we should map what the things *appear to be*.
> Taking the example of Japanese roads, I would also add "with
> reasonably common knowledge". It does leave some space for
> uncertainty, but this uncertainty is also present in real life, so it
> can appear in OSM as well.
>
>
> Warin's question also identifies a problem I'd like to discuss. There
> seem to be no "formal" agreements on how to create OSM. Things are
> documented in the wiki, which is subject to uncontrolled changes and
> no review and which is not always read by the mappers. Data is then
> used by software development companies in the way they find
> reasonable, without any foundation for consistency. It may be cultural
> but I am looking for some sort of more robust, maybe even enforced,
> agreements. They may be subject to changes, but their mere existence
> would help. What do you think?
>
>
> Cheers,
> Kotya
>
>
>
>
> On Wed, Jan 14, 2015 at 1:28 AM, Warin <61sundow...@gmail.com> wrote:
> > This comes from the tap discussion but has implications elsewhere.
> >
> > What is the basic philosophy of OSM tagging at the top level?
> >
> > Are 'we' tagging for
> >
> > What things are? eg highways
> >
> > OR
> >
> > What things are used for? eg amenity
> >
> > 
> > Explanation? By example;
> >
> > Highways are used for transport so would be better tagged as
> > transport=motorway, sub tags for vehicles etc.
> >
> > OR
> >
> > amenity=drinking_water would be better tagged as water=blubber
> >
> > --
> > Is there an FAQ on this? Or has this never been documented/though of?
> > Have fun with this  :)
> >
> >
> >
> >
> > ___
> > Tagging mailing list
> > Tagging@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/tagging
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Basic philosophy of OSM tagging

[Michał Brzozowski]

Yes. As much as all this "you can use any tags you want" creates an
environment open for innovation, it creates a horrible mess when you
use it without coordination and on existing features.
Also, no-one seems to ask a question "How it that problem solved in
other maps?" when proposing tags or other solutions.
No need to re-invent the wheel. Look how others do it and you may find
an elegant or just practical solution.

Some people in Poland (the ones who never browse community forums)
maniacally tag every dirt road as highway=track, even if it should be
residential+unpaved (Like short named streets at suburbs) Another
prevalent  thing is tagging every place such as out-patient clinic as
amenity=hospital.
Lack of discipline is very annoying because correcting these is a
Sisyphean task.

Also, I think that editor presets makers should really implement *all*
approved tags (barring some specialized stuff like OSM-3D, indoor
mapping etc) because not featuring a tag makes some people tag things
not exactly correctly, just to map it.

Michał

On Thu, Jan 15, 2015 at 5:08 PM, Kotya Karapetyan  wrote:
> Now that the water_tap proposal discussion is over, I'd like to join
> this important discussion.
>
> My opinion: Since OSM is a *map*, we should *map* things. That means,
> we should tag what actually exists on the planet, not what is implied.
> Sometimes things are tagged in real life. For example, motorways are
> marked with special traffic signs, therefore we can tag a road as
> motorway. In other cases, we should use common sense to call the
> things by their names. I am usually asking myself: How would I explain
> to a tourist how to find his way? I will use something like "Pass by
> ..." or "You will see ...". This "..." is then the name (hence tag) to
> be used.
>
> A good example of the contrary is amenity=drinking_water. Though the
> original intention is clear, I believe the solution is suboptimal: a
> tourist wouldn't know what to search for (since it is not drinking
> water but rather its source that is actually visible in a given
> place). A mapper may also have hard times identifying whether a
> specific water source provides potable water or not.
>
> So, my answer would be we should map what the things *appear to be*.
> Taking the example of Japanese roads, I would also add "with
> reasonably common knowledge". It does leave some space for
> uncertainty, but this uncertainty is also present in real life, so it
> can appear in OSM as well.
>
>
> Warin's question also identifies a problem I'd like to discuss. There
> seem to be no "formal" agreements on how to create OSM. Things are
> documented in the wiki, which is subject to uncontrolled changes and
> no review and which is not always read by the mappers. Data is then
> used by software development companies in the way they find
> reasonable, without any foundation for consistency. It may be cultural
> but I am looking for some sort of more robust, maybe even enforced,
> agreements. They may be subject to changes, but their mere existence
> would help. What do you think?
>
>
> Cheers,
> Kotya
>
>
>
>
> On Wed, Jan 14, 2015 at 1:28 AM, Warin <61sundow...@gmail.com> wrote:
>> This comes from the tap discussion but has implications elsewhere.
>>
>> What is the basic philosophy of OSM tagging at the top level?
>>
>> Are 'we' tagging for
>>
>> What things are? eg highways
>>
>> OR
>>
>> What things are used for? eg amenity
>>
>> 
>> Explanation? By example;
>>
>> Highways are used for transport so would be better tagged as
>> transport=motorway, sub tags for vehicles etc.
>>
>> OR
>>
>> amenity=drinking_water would be better tagged as water=blubber
>>
>> --
>> Is there an FAQ on this? Or has this never been documented/though of?
>> Have fun with this  :)
>>
>>
>>
>>
>> ___
>> Tagging mailing list
>> Tagging@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/tagging
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging

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

Re: [Tagging] Basic philosophy of OSM tagging

[Marc Gemis]

On Thu, Jan 15, 2015 at 6:07 PM, Michał Brzozowski 
wrote:

> Some people in Poland (the ones who never browse community forums)
> maniacally tag every dirt road as highway=track, even if it should be
> residential+unpaved (Like short named streets at suburbs)
>

so how would this be solved by more formal rules ? The wiki now already
discribes how to tag such a road.
If people insist on taking the wrong value in the editor, you cannot do
much about it. In controlled environments you can fire the employee if they
don't map by the rules, in OSM you can only block 1 user account after a
long process. Are there any other crowd-sourced maps ? Otherwise it would
be difficult to find examples of how others do it.

Maybe it's way too easy to start mapping. Maybe you should first follow a
course on how to map before making your first edit. During this course you
can learn about the good mapping habits in your country. But this is
probably also not a popular measure.

I  like your idea of editors adding more tags in their presets, but it
doesn't necessarily have to be the approved ones. Some non approved tags
are used more. So, all the ones with significant usage according to taginfo
should be included.

The more I learn about the tag-approval process, the less suitable it seems
to me for an every growing community. Maybe 10 votes was a lot in the early
days, now it is peanuts. This morning I just thought about it and it is
pretty easy to set up 10 accounts and vote in favour of any tag I like.
Same is true for no-votes. There is no control over who votes... Not that
it was misused in this way yet.

Regarding the tagging:

is it important that there are 100's of amenities on the toplevel ? What's
the problem with that ? I really don't care what the key is under which
e.g. place_of_worship can be found, it could as well have been "aaa001". I
just hope that the editor allows me to look for place of worship and offer
me the right tag. I there anyone browsing all amenities ? On the wiki I
just search for the value and hope to get the right page.

We should have better tools to find the tags we need, the schema is less
important. Someone is working on a tool to show all similar tags, using
synonyms etc. that seems very promising to me

regards

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

Re: [Tagging] Basic philosophy of OSM tagging

[Michał Brzozowski]

On Thu, Jan 15, 2015 at 6:30 PM, Marc Gemis  wrote:

> Maybe it's way too easy to start mapping. Maybe you should first follow a
> course on how to map before making your first edit. During this course you
> can learn about the good mapping habits in your country. But this is
> probably also not a popular measure.

Some time ago I thought "let's be realistic" and that iD, our go-to
editor for newbie mappers, could show some task-oriented
mini-tutorials. Like "I want to map a / change..." so it would show
relevant remarks (that is, like "When drawing buildings, do not place
the outline on the roof, move it to the base. Remember to rectify
corners if the building if applicable" and so on). A large quantity of
mappers are, as the HDYC site ranks them, "hit-and-run mappers" that
only map 5 or 20 edits and get bored - or just OSM is superbly mapped
in their area (:

If you can't beat them (iD), join them...

Michał

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

Re: [Tagging] Basic philosophy of OSM tagging

[Marc Gemis]

Good idea,

Building further on this idea: Right now the main problem we have with iD
in Belgium is that we have better, more recent aerial imagery that cannot
be used in iD. They are working on that but until a new version is released
new mappers might realign house with old images from Bing. It would be nice
that each country could create messages that are displayed when you start
editing in their area. So even experienced mapper could be informed about
news from the local community.


Coming back to the unpaved residential roads vs. tracks: Wouldn't that be
solved by having a map that displays the difference between a paved
residential road, unpaved and track? I know this has been discussed before.



On Thu, Jan 15, 2015 at 8:39 PM, Michał Brzozowski 
wrote:

> On Thu, Jan 15, 2015 at 6:30 PM, Marc Gemis  wrote:
>
> > Maybe it's way too easy to start mapping. Maybe you should first follow a
> > course on how to map before making your first edit. During this course
> you
> > can learn about the good mapping habits in your country. But this is
> > probably also not a popular measure.
>
> Some time ago I thought "let's be realistic" and that iD, our go-to
> editor for newbie mappers, could show some task-oriented
> mini-tutorials. Like "I want to map a / change..." so it would show
> relevant remarks (that is, like "When drawing buildings, do not place
> the outline on the roof, move it to the base. Remember to rectify
> corners if the building if applicable" and so on). A large quantity of
> mappers are, as the HDYC site ranks them, "hit-and-run mappers" that
> only map 5 or 20 edits and get bored - or just OSM is superbly mapped
> in their area (:
>
> If you can't beat them (iD), join them...
>
> Michał
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Basic philosophy of OSM tagging

[Kotya Karapetyan]

Hi Marc,

> By forced rules: you mean a committee that decides what gets mapped and how
> ?
> So when I want to map something now, I have to file a request to the
> committee to start looking for a new tag. And if they like the request they
> come back within a few months with a proposal. And this committee is
> all-knowing, so they know all the exceptions in the different countries ? So
> I don't have to ask for an update when they misunderstood me ?

My vision was more along these lines. There will be a tagging
committee: It will maintain the official OSM tagging scheme. Mapping
will not be changed w.r.t. the current situation. The tagging
functionality as such will remain as it is, however:
1) Software designers can use the official scheme to implement tools
for mapping, display, and routing.
2) The committee will review the existing tags to avoid conflicts, and
decide what tags are to be deprecated, where to adjust official
definition etc.
3) The short formal description of the tags is included in the scheme.
4) The committee has the final word in approving the tags (just to
remove the mess with existing discussion-voting process).
5) Mappers can still implement and use tags as now, however they
should be reviewed by the committee to get into the official scheme.

So something similar to how HTML standard is maintained by W3C:
Google, Microsoft and Mozilla can do and do whatever they like in
their browsers and online tooling, but there is an official standard
based on some proposals. The software developers do not have to stick
to the official scheme, but it helps make things more consistent.

Cheers,
Kotya

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

Re: [Tagging] Basic philosophy of OSM tagging

[Kotya Karapetyan]

On Thu, Jan 15, 2015 at 6:07 PM, Michał Brzozowski  wrote:
> Also, I think that editor presets makers should really implement *all*
> approved tags (barring some specialized stuff like OSM-3D, indoor
> mapping etc) because not featuring a tag makes some people tag things
> not exactly correctly, just to map it.

+1.

I would also like us to discuss a possibility of making the tagging
scheme more error-proof. A couple of issues to address:
1) Conflicting tags should be prevented by means of software.
2) Suggested tags functionality should be implemented.
3) Official tag description should be automatically taken from the
formal approved tagging scheme and shown in the mapping tool, to
minimize misunderstanding. Full tag description stays in the wiki to
include examples, photos, context etc.
4) Invalid tag values should be automatically identified and warnings
should be shown.

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

Re: [Tagging] Basic philosophy of OSM tagging

[David Bannon]

On Thu, 2015-01-15 at 18:07 +0100, Michał Brzozowski wrote:

> Some people in Poland (the ones who never browse community forums)
> maniacally tag every dirt road as highway=track, even if it should be
> residential+unpaved 

Thats is a case of "tagging for the renderer" I'm afraid. They do that
because they want to see the map show the unpaved-ness (sorry) of the
road. Clearly wrong but you can understand why they do it.

I have had my say on the topic many times as has many other people.

It does, IMHO, highlight one more aspect of this philosophy question.
People map and want to see the data they enter used in some way. That
"seeing" is an essential part of the feedback loop. We need to consider
that when looking at how people choose (or invent) tags.

David





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

Re: [Tagging] Basic philosophy of OSM tagging

[johnw]

> On Jan 15, 2015, at 8:33 PM, Mateusz Konieczny  wrote:
> 
> > 4 lane ‘tertiary" road that handles 5 times the vehicle traffic, traveling 
> > on to connect with 2 major trunk roads -  
> > intersects the narrow two lane “secondary road”  that is one of the small 
> > roads coming down from the “suburbs” into the city
> 
> Interesting. I was unaware about so drastic changes in meaning of tag by 
> editors. This one cripples Japanese road data - 
> and I am unsure what could be done about it.


http://www.openstreetmap.org/edit#map=20/36.33615/139.21688 


The “secondary road” I speak of (I was mistaken, it isn’t a primary).  To the 
north, it was recently widened (again) and nice, but below rt. 17, if it didn’t 
have a shield # - it would be an unclassified road (as it doesn’t even have a 
center line and a very awkward turn at a narrow level crossing). 



http://www.openstreetmap.org/edit#map=19/36.40643/139.07303 
 

The “tertiary” road intersection. it connects the primary and the trunk 
intersection to the west (rt 17 / 6), and then (after being secondary for a 
block) meets a “primary road” is tertiary again for several KM as it heads down 
to the rt 50 trunk road and on to “primary” rt 2. This kind of situation occurs 
a lot in cities where old roads meet new bypass roads. I had tagged and 
retagged it as a secondary, but reverted it to tertiary because I have decided 
to follow the OSM:JA rules in this case (after writing my long-winded example).

http://wiki.openstreetmap.org/wiki/Japan_tagging 


They follow their shield system, which places roads into categories which often 
do line up as you expect, but in cities, where there are newer major roads that 
connect the older ones, or in old neighborhoods that are bypassed by newer 
roads, it does not represent the road system very well sometimes.  Notice there 
is no mention of the road size or traffic flow consideration - just shield 
numbers. 

http://www.openstreetmap.org/edit#map=16/36.9819/139.3064 
 

One of the roads in my region, Gunma Route 1, runs through what is now the 
largest national park in Japan, and some pieces are completely closed to all 
public road traffic (it is not continuous), besides the park shuttle bus. I 
assumed I could drive on it (as it is labeled as a primary road, no tolls), and 
drove up there to find (a) it is a narrow service road and (b) roped off, and 
only the park shuttle bus is allowed in (as opposed to private busses and 
taxis, as with other “limited access” park roads in Japan).  I re-tagged the 
private section as a service road, and hastily made the park entrance months 
ago.  it still has it’s ref=1. 



Honestly I don’t mind them being screwed up (besides the rt1, because you 
cannot drive on it whatsoever) - because it is the way that Japanese people 
expect maps to be presented to them. They have pretty rigid ways of 
representing maps - and they have adapted OSM to fit their national mapping 
scheme and expectations:

The two unconnected sections of rt. 1 -  the thin yellow portions are not 
drivable by the public, but still labeled as primary regional roads because of 
the shield type/number. 
http://www.mapion.co.jp/m/basic/36.93219243_139.3159987_5/t=print/size=640x740/icon=home,139.31599869871812,36.93219243059891/
 

 

The “tertiary” intersection 
http://www.mapion.co.jp/m/basic/?t=print&lat=36.4033877201575&lon=139.0776209799907&scl=9&icon=home,139.0776209799907,36.4033877201575/
 



As you can see, rendered width of the roads is the defining feature of use 
(check Google Maps, it uses Zenrin data up close to show the width too) and the 
classification below “trunk" is less important,  whereas in OSM, the 
classification is most important.  The width and standards of most roads 
changes drastically as they go along, and rendering the width change is easer 
than reclassifying the roads section by section to reflect changing road 
standards (width, traffic volume, shoulder, curve radius, expected hazards, 
etc). 


BTW - the way Mapion renders different neighboorhoods, borders, and other 
things are quite interesting, especially as it drastically changes with the 
zoom level. 
http://www.mapion.co.jp/m/36.43371388_139.05561388_8/v=m1:群馬県前橋市関根町28/ 

 might have some good ideas for -carto in the future. 


Javbw


___
Ta

Re: [Tagging] Basic philosophy of OSM tagging

[Pieren]

On Thu, Jan 15, 2015 at 9:13 PM, Kotya Karapetyan  wrote:

> There will be a tagging
> committee: It will maintain the official OSM tagging scheme.

Historical contributors and leaders will tell you that there is no
"official committee" in OSM. But, to be a litle provocative, I would
say we already have two committees for tagging scheme:
- the JOSM presets maintainers
- the DWG

Pieren

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

Re: [Tagging] Basic philosophy of OSM tagging

[moltonel 3x Combo]

On 14/01/2015, Warin <61sundow...@gmail.com> wrote:
> Are 'we' tagging for
>
> What things are? eg highways
> OR
> What things are used for? eg amenity

Why does it have to be one exclusive_or the other ? Both what things
are and what they are used for is important. There's normally always a
way to tag both, even if the tags to do that can be a bit bysantine.
New tags may want to focus on one aspect or the other, there's no hard
rule.

The highway tag is a poor example for the nature/use dichotomy. The
interpretation differs by country (hopefully not within a country).
Some countries place more importance on use (amount and type of
traffic...) than on nature (surface, width...). Other countries adhere
rather strictly to the national road refs, even if the road changes
classification for purely administrative reasons (so neither nature
nor use). What the highway tag is a good example of, is the organic
evolution of osm's tagging system.

One good example of nature vs use is church buildings reconverted to
shops/pubs/tourist offices (all 3 can be seen in Dublin) and
place_of_worships in various incongruous buildings. The building tag
is now clearly documented as pertaining to the nature of things, but
this hasn't always been the case and a lot existing building tags have
dubious values.

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

Re: [Tagging] Basic philosophy of OSM tagging

[moltonel 3x Combo]

On 16/01/2015, Pieren  wrote:
 Historical contributors and leaders will tell you that there is no
> "official committee" in OSM. But, to be a litle provocative, I would
> say we already have two committees for tagging scheme:
> - the JOSM presets maintainers
> - the DWG

AFAIK the DWG's influence on tags is fairly light. But all the editors
(not just josm), the renderings (especially the "default" mapnik one,
but also the specialized 3d and indoor ones), the search engines
(Nominatim), the routers (OSRM) all carry a lot of weight. Anybody
spending time on the wiki and mailing lists (and to a lesser degree
also on forum, irc, help, etc) carry some weight too.

Of course not everybody has the same weight... But that crowd is
varied enough that there's IMHO no reason to fear one overreaching
commitee.

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

Re: [Tagging] Basic philosophy of OSM tagging

[Mateusz Konieczny]

Note also massive influence of JOSM validator and other
rule-checker tools.

2015-01-16 13:14 GMT+01:00 moltonel 3x Combo :

> On 16/01/2015, Pieren  wrote:
>  Historical contributors and leaders will tell you that there is no
> > "official committee" in OSM. But, to be a litle provocative, I would
> > say we already have two committees for tagging scheme:
> > - the JOSM presets maintainers
> > - the DWG
>
> AFAIK the DWG's influence on tags is fairly light. But all the editors
> (not just josm), the renderings (especially the "default" mapnik one,
> but also the specialized 3d and indoor ones), the search engines
> (Nominatim), the routers (OSRM) all carry a lot of weight. Anybody
> spending time on the wiki and mailing lists (and to a lesser degree
> also on forum, irc, help, etc) carry some weight too.
>
> Of course not everybody has the same weight... But that crowd is
> varied enough that there's IMHO no reason to fear one overreaching
> commitee.
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Basic philosophy of OSM tagging

[althio althio]

"Kotya Karapetyan"
> 2) Suggested tags functionality should be implemented.

I have seen that in the Android editor Vespucci.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Basic philosophy of OSM tagging

[Martin Koppenhoefer]

> Am 15.01.2015 um 20:52 schrieb Marc Gemis :
> 
> It would be nice that each country could create messages that are displayed 
> when you start editing in their area.


+1, but will not solve all problems obviously, as the topic is quite complex. 
We'd either cover only the most important points ( +1 ), or have a very long, 
trying to be exhaustive, list, that no one will read or could remember when 
mapping;-)

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

Re: [Tagging] Basic philosophy of OSM tagging

[Martin Koppenhoefer]

> Am 16.01.2015 um 10:56 schrieb Pieren :
> 
> I would
> say we already have two committees for tagging scheme:
> - the JOSM presets maintainers
> - the DWG


I'd add 
- the osm-carto-style sheet maintainer
- the ID developers

plus the Nominatim developers, osm2pgsql developers (at least for relation 
types) and to a smaller extent every other data consumer (eg cycle map, wheel 
map, open sea map, ...)

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

Re: [Tagging] Basic philosophy of OSM tagging

[Martin Koppenhoefer]

> Am 16.01.2015 um 12:56 schrieb moltonel 3x Combo :
> 
> The building tag
> is now clearly documented as pertaining to the nature of things, but
> this hasn't always been the case

at least in the past 7 years (and back then there have been hardly any 
buildings mapped) it has always been the building type to be expressed with the 
tag (and 99% have been "yes"). Now "type" is not completely clear either, it 
can refer to the usage class or to the construction system for instance, but in 
OSM I've always seen it as usage type (eg hotel, church, ...), eventually 
combining formal aspects if you do it in a detailed way.

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

Re: [Tagging] Basic philosophy of OSM tagging

[althio althio]

Warin <61sundow...@gmail.com> wrote:
> What is the basic philosophy of OSM tagging at the top level?
> ...
> Is there an FAQ on this? Or has this never been documented

I do not have a FAQ on philosophy, only this and that...

A few entries about 'how to create/propose/use' tags:
http://wiki.openstreetmap.org/wiki/Proposal_process#Significance
http://wiki.openstreetmap.org/wiki/Any_tags_you_like#Syntactic_conventions_for_new_tags
http://wiki.openstreetmap.org/wiki/User:Joto/How_to_invent_tags
http://wiki.openstreetmap.org/wiki/Duck_tagging

New tagging scheme should also be designed so that Tagging good
practices make sense.
http://wiki.openstreetmap.org/wiki/Good_practice
http://wiki.openstreetmap.org/wiki/Tags
http://wiki.openstreetmap.org/wiki/Namespace

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

Re: [Tagging] Basic philosophy of OSM tagging

[johnw]

> On Jan 16, 2015, at 6:22 AM, David Bannon  wrote:
> 
> On Thu, 2015-01-15 at 18:07 +0100, Michał Brzozowski wrote:
> 
>> Some people in Poland (the ones who never browse community forums)
>> maniacally tag every dirt road as highway=track, even if it should be
>> residential+unpaved 
> 
> Thats is a case of "tagging for the renderer" I'm afraid. They do that
> because they want to see the map show the unpaved-ness (sorry) of the
> road. Clearly wrong but you can understand why they do it.

I had a quick comment, especially as I have done exactly that when I first 
started. I’m going back to correct some of my more glaring errors as I clean up 
what I had done int he past. 

The biggest thing to me is that is exactly what the mappers are working for - a 
more accurate map (the output of the renderer) and it takes a long time to get 
the idea that we’re tagging for a dataset, not an output for the dataset. 
“Those people talking about relations and semicolons have to worry about that 
stuff, I’m just tracing imagery in iD!” I’m guessing is the mindset. And 
especially where I started mapping, in central Japan which had almost no 
cleanup done since a data import, bringing any order to the spaghetti of badly 
aligned and outdated “unclassified roads” that covered everything was better 
than doing nothing. 

since there is not many obvious visible uses (to new mappers) for the mapping 
data beside the output rendered onopenstreetmap.org 
, then tagging for the renderer is the only 
possible verifiable way to check to see what they are doing is (remotely) 
correct. Train stations were confusing as hell to me, as a couple broken areas 
made me completely flummoxed as to how to get a “station” to “show up” 
correctly. I worked for hours trying whatever I could to get the station to 
show up, assuming I was bad at tagging, and I just needed to find the right way 
to make it show up properly. 

Maybe if there were some different overlays that can be put onto the OSM site - 
“Usage” “surface” etc - it would be easy to see what is tagged for purpose of 
use (driveway) and surface tagged or untagged (paved , ground, gravel, 
untagged) - it might make checking existing work (for novices such as myself) 
much easier, especially since I’m not mapping in JSOM or aware of all these web 
tools everyone seems to know how to use. 

As an easy (well, easier) fix - it might be a good idea for iD to show, color 
coded, what option is chosen for a road - white for paved, grey striped for 
gravel and friends, and brown for various ground / soil/ mud etc.   it might 
make it easier to tag, and easier to keep noobs like me from tagging everything 
dirt as a track when they want to convey that it’s not a paved road. 

and it would show that there is a difference between what you see in -carto and 
what you see in the renderers. 

Javbw

> 
> I have had my say on the topic many times as has many other people.
> 
> It does, IMHO, highlight one more aspect of this philosophy question.
> People map and want to see the data they enter used in some way. That
> "seeing" is an essential part of the feedback loop. We need to consider
> that when looking at how people choose (or invent) tags.
> 
> David

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

Re: [Tagging] Basic philosophy of OSM tagging

[Clifford Snow]

On Tue, Jan 20, 2015 at 8:34 PM, johnw  wrote:

> As an easy (well, easier) fix - it might be a good idea for iD to show,
> color coded, what option is chosen for a road - white for paved, grey
> striped for gravel and friends, and brown for various ground / soil/ mud
> etc.   it might make it easier to tag, and easier to keep noobs like me
> from tagging everything dirt as a track when they want to convey that it’s
> not a paved road.


+1


-- 
@osm_seattle
osm_seattle.snowandsnow.us
OpenStreetMap: Maps with a human touch
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging