[Tagging] Tagging buildings that people work in

bkil bkil.hu+Aq at gmail.com
Sat Jun 1 09:06:13 UTC 2019

Sorry if I didn't make myself clear in formulating the questions, I'll try
to rephrase my inquiries again below.

On Wed, May 29, 2019 at 8:09 PM ET Commands <etcommands at gmail.com> wrote:

> > Date: Fri, 24 May 2019 20:34:52 +0200
> > From: bkil
> > Subject: Re: [Tagging] Tagging buildings that people work in
> >
> > landuse=* seemed appropriate for most use cases I have encountered. Why
> do
> > we need to tag this on a building resolution?
> Because landuse is for the entire property a building sits on, not the
> building itself.
You've described the difference between specifying the high level landuse
in an area (that may be even a few blocks large) compared to the proposed
micro-mapping on buildings. This is correct, but I would like to know the
reason, meaning what advantage would such a resolution carry to the map

> What data consumers did you have in mind?
> Mapmakers.
Being a mapmaker is a profession - they make maps for a given audience, for
a given user group or a given task. Specifically, by data consumer we mean
a downstream project that offers services based on OSM data, commonly in a
slippy-map format, but sometimes aggregating statistics or joining data
from various places, providing better insight or an index.

For example, given the height of a building or the number of floors or its
colour, there exist applications that can show a 3D map or even generate
game scenery from real cities. By adding a proper digital elevation model,
you could envision the use of this information for radio antenna link
planning as well. Tagging the kind of sports played in a recreation centre
or in a bar makes it available for map queries (where is the nearest place
I could play pool at?).

It's not a problem if there doesn't exist any data consumers for a new
notation and you don't want to implement any either - I'm just curious
whether you could come up with a plausible one to be created in the future.

Reading this answer together with the next one makes me feel as if we
should map this for the general renderer and "just because we can"
(although we probably can't as per my other points), but I hope I'm reading
you incorrectly. We could use line laser scanners to reproduce roads to
submillimeter accuracy, but we don't do that - instead we take a model of a
real life entity and represent it in a way that serves a given real life
purpose like pedestrian and automotive navigation. Buildings have all kinds
of imperfections (also accumulating with time), but we usually represent
them as simple boxes (that mind you is already too detailed to the taste of

Another drawback is that Mapnik already has too many colours and symbols by
only showing the items of common interest, hence they had to hide a few
already. I'd definitely not like to compromise even more items of common
interest for occupancy on the general overview maps. I couldn't think of
who would look at specialized maps only showing these, but as I've asked
previously, please do share.

> What common interest does this annotation serve?
> It allows you to symbolize "occupied" buildings differently from
> "unoccupied" ones.
For example, people drawing their own back yards and their own garage there
does not serve common interest, however if you add a roof (rain cover) that
resides in public property where I could stand under if it is raining, it
is of common interest.

Another example is mapping your own water tap does not serve common
interest, but at the same time mapping a water tap found on the street adds
great value, as I could go there and have some water when I'm hiking in the
summer when it's hot and I'm thirsty.

Do people want to go inside an unoccupied building to seek shelter? Do
people want to see these kind of buildings as an attraction? Do people want
to stay away from occupied buildings due to possible danger? Should we
search for survivors in occupied buildings after a disaster (not good
enough because we should also search in residential buildings)? Does this
help mobile cell phone base station planning (although I guest they simply
look at their spatial network stats and be done with it)?

In general, I really like features, new tags and micromapping, but please
help me out a little here. I'd still need some convincing to see why this
would be a useful addition.

If it does not stand on its own as a feature, then as you've mentioned
yourself, it's just a placeholder, a kind of FIXME that you expect others
to map in more detail for it to be usable. See the answer I've given in the
relevant section as to why I view this as a bad thing.

> What is the verification criteria? Do I need to station next to the
> My personal criteria is not meant to be that exact.  For example, I can

You may have a given heuristic in your head right now, but you will need to
provide a pseudo-algorithm and document it in detail in the wiki for others
to follow. A tagging scheme is only useful if everybody is tagging the same
way (or at least in a compatible way).

It also happens with well documented, established tags that people engage
in light "tagging wars" - each visiting the same place tags it in a
slightly different, but still kind of valid way. Imagine how often this
happened if you couldn't give an exact algorithm with your proposal.

> Do you consider weekend houses occupied if they are only occupied
> > intermittently or even seasonally? How do I verify this?
> Note that my question was in reference to buildings people work in, not
> live in.
This goes against the definition most people understand, so even if
$SUBJECT could be tagged, a better word should definitely be found to
describe the phenomenon:


> Note that we usually do not add fixme kind of tagging for the sole purpose
> > of marking the absence of regular information, as by definition, a blank
> > map is missing an infinite amount of information and we would definitely
> > not like to store so many fixme's.
> I was not advocating the use of fixme's.  Knowing that a building is
> "occupied" is having more knowledge than simply knowing that a building
> exists.

One of the replies given to you has hinted at using note=* (which is for a
different concept, leaving an advisory for the next editor, not to mark
something as missing) in a fixme=* fashion. See my answer above regarding
whether this is of common interest. If it is not, mapping these is
equivalent to a fixme - expecting others to add real value to the map
because we couldn't.

It is not necessary to know everything about a feature in order
> to map it.  OpenStreetMap will never be "complete," because there will
> always be more information that can be added to features.
Exactly, and this is why we don't list the kind of information missing from
a given feature. We don't add fixme=height, despite the fact that
everything that can usually be mapped also has a height, - we simply add
height=* when it is known and leave it out otherwise.

True.  But abandoned buildings are not the only buildings that people do
> not work in.  An example is storage buildings.
Yes, I have included that too as an example in my last response, but I
would still need the reason behind why we would need to map where potential
storage buildings are located. Although, I think we could extend landuse to
mark areas with mostly storage buildings if this is your real major use
case, although you still haven't shared your use case with us.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20190601/4cad7ce2/attachment-0001.html>

More information about the Tagging mailing list