Proposal talk:Dog barrier

New proposal

barrier=dog

barriers can't be made out of dogs. Suggest barrier=dog_fence. This would also follow the style deer_fence with 6 usages. Barriers can be wood, metal etc and have height values.
Coupled with optional sign I believe it would work.

--TonyS (talk) 09:16, 26 May 2020 (UTC)

How does such special dog-fence look like? And also: what is the key difference to the tagging with barrier=fence, fence_type=* and an additional "dog-warning-area-tag"?

To the barrier-tagging in general: Not all areas contain a fence, some dogs may be held back by a long leash or chain. I don't think a barrier tag (or even fence specific tag) is the right way.

--Hauke-stieler (talk) 10:17, 26 May 2020 (UTC)

I have a link to a barrier=dog example in the main body of the text. --Floridaeditor (talk) 14:46, 28 May 2020 (UTC)

This is still a hazard, not a "barrier". I feel lost on your change of the proposal. -- Kovposch (talk) 19:42, 28 May 2020 (UTC)

I agree that "barrier=dog" is incorrect. The value of barrier=* is the type of obstacle which blocks a road, street or boundary: e.g. barrier=bollar, barrier=bollar, barrier=log, barrier=chain, barrier=fence. A dog is not like any of these barriers, since it is a living creature which can move out of the way. I don't believe the location of dogs should be mapped in OpenStreetMap, see Good practice. --Jeisenbe (talk) 16:13, 28 May 2020 (UTC)

I don't understand what's with barrier=dog, but about "location" of dog, the old version of this proposal was to indicate hazard as in hazard=* (34+11=45 instances) - issue would then be to distinguish signed, from unsigned, surveyed conditions.

hazard=dog	hazard=dogs

-- Kovposch (talk) 19:42, 28 May 2020 (UTC)

Sorry that I've missed the discussion. In my contribution on barrier/talk I had described a significant barrier, not a transient feature, not a potential hazard, not a warning sign. In my favourite village in the Cretan mountains the shepards have blocked a track by chaining a dog on each side of he road, with chains long enough to meet in the middle. This effectively stops sheep, hikers and cyclists, while cars simply drive through. It's not at all a transient feature, the dogs have little boxes made of concrete and they (or their successors) are there for some 15 years. My idea was to make OSM using hikers aware for this barrier as well as for the fence on both sides. Next time I will take a photo of this construction. I don't know whether it's a unique feature or used also elsewhere. I guess it's cheaper than a cattle grid, after a dog counts almost for nothing in that region. Some may remember that there were also "dog barriers" at the inner-German border (cf. German Wikipedia), some of them designed as cable-runs, other by keeping dogs between two fences. --GerdHH (talk) 16:17, 1 December 2020 (UTC)

Old proposal

Bad tagging. Questionable proposal.

You propose to make this dog=hazard and specifically refer to the wiki page for the dog=* key. So what does that have to say?

• dog=yes dogs are allowed
• dog=no dogs are not allowed
• dog=leashed dogs are allowed only when on a leash

Which leads us, logically, to:

• dog=hazard dogs are allowed only if they are a hazard

If there were any merit to the proposal, it would be better as hazard=dog even though hazard=* seems to be about hazards to road traffic.

--Brian de Ford (talk) 22:39, 12 May 2020 (UTC)

+1, "dog=hazard" could also be interpreted a a hazard for dogs, not the dog being the hazard. As said above, dog=* ist used as an access tag, so mixing other meanings in gets confusing.--Polarbear w (talk) 08:10, 13 May 2020 (UTC)

I will change the tag to dog_warning=yes. --Floridaeditor (talk) 12:49, 13 May 2020 (UTC)

Proposal makes no sense

Dogs are not fixed items: they move around. Dogs are not permanent or semi-permanent: they die, or are sold, or are given away, or are impounded. Dogs are too ephemeral to be mapped.

You may argue that the owner of the property has a history of replacing dangerous dogs when they die. But that owner can move or die. Legal authorities may take note of the habit of keeping dangerous pets and ban the owner from keeping dogs.

I'd expect anyone with a dangerous dog to have warning signs on the property or risk the possibility of prosecution for negligence if somebody is harmed.

You may argue that the owner doesn't have such a sign therefore we must map it. But if there is no such sign it's not verifiable. Even if you saw the dog and were threatened by it, perhaps it belongs to a visitor. If there isn't a sign we can't map it; if there is a sign we do not need to map it.

--Brian de Ford (talk) 22:39, 12 May 2020 (UTC)

I disagree; see construction=*, which changes all the time. Also, I think we should map if the sign is there, this would be like not mapping a road's name because the sign is there. --Floridaeditor (talk) 12:49, 13 May 2020 (UTC)

With construction=*, you mention that the place will have to be reviewed. No such invitation to review your proposed tag.

--Nospam2005 (talk) 12:52, 24 May 2020 (UTC)

The caption under the picture of the dog park in the proposal says: "A dog hazard node could go on the outside of this dog park's entrance, as dogs may get out and bite pedestrians". Thus the proposal recommends to map common sense, the potential hazard is implicit in the feature of the dog park. Similarly, we would need to map hazard=car to each highway since somebody could have an accident there. --Polarbear w (talk) 08:25, 13 May 2020 (UTC)

I agree, I will change that example. --Floridaeditor (talk) 12:53, 13 May 2020 (UTC)

The restriction to nodes makes no sense. I have tagged a (right of) way across a farmyard with hazard=dangerous_dogs. It is the whole way that is dangerous. Likewise, restricting this to the front of a property is ridiculous. In the case of a path through a farmyard, it has no obvious connection with the "front" of a property, whatever that might mean. —Preceding unsigned comment added by Messpert (talk • contribs) 13 May 2020

Yes, that is a good point. I suppose it will apply to nodes, ways and areas now. Thanks for the feedback. --Floridaeditor (talk) 12:49, 13 May 2020 (UTC)

Usage of the term "hazard"

I don't know if the word "hazard" is the best term to describe the situation. Most signs are saying "beware of the dog", something like "warning: security dog" or similar things. So maybe a tag like one of the following might be better?

• dog:warning=yes|no
• dog:danger=yes|no
• dog:danger_warning=yes|no
• dog:beware=yes|no
• dog:beware_warning=yes|no

Or just plain tags like dog_beware=yes (or similar)? I currently don't know which is the best but I pefer the dog:beware_warning tag as it is -- at least in my opinion -- the least ambiguous one.

--Hauke-stieler (talk) 10:23, 13 May 2020 (UTC)

But why not use hazard=dogs or maybe hazard=dangerous_dogs? We already have the hazard tag: why invent a new dog tag? —Preceding unsigned comment added by Messpert (talk • contribs) 13 May 2020

Because a hazard is - as e.g. the key hazard_type=* partly describes it - a larger and especially natural danger. The hazard key itself is also just a very old proposal draft, so nothing official. Furthermore, an area with a dangerous dog is neither large and important for most people or traffic, nor is it a natural thing. That's why I personally would distinguish between a hazard and unnatural (man made) dangerous things. As there is no general "danger" tag, I think a dog specific tag (for example dog:beware_warning=yes|no) is fine. --Hauke-stieler (talk) 11:56, 13 May 2020 (UTC)

I will change the tag to dog_warning=yes. --Floridaeditor (talk) 12:49, 13 May 2020 (UTC)

patch of wood where dogs are allowed to be unleashed

Would you also use this for a patch of forest where dog owners are allowed to walk their dogs unleashed? It is forbidden to cycle through it since it changed status.

The dog owners are only supposed to let their dog roam, if they are 90% (ballpark figure) the dog doesn't pose danger to humans or other dogs. i.e. the dog is socialised properly. --Polyglot (talk) 12:34, 13 May 2020 (UTC)

Not how I originally intended it to be, but it could work. --Floridaeditor (talk) 12:49, 13 May 2020 (UTC)

Harbours in Greece

When I traveled to Greece 30 years ago, there were many stray dogs in the harbour areas, I don't think this has changed. Would you tag those as well? On the area? They are mostly not harmful, but they could be. Let's say I didn't feel entirely safe there, even though I didn't have any actual bad encounters. --Polyglot (talk) 12:34, 13 May 2020 (UTC)

Yes, you would add it to the harbor area. This is a great example. --Floridaeditor (talk) 12:49, 13 May 2020 (UTC)

Animal Behaviour

Not a fan of adding transient behaviours based on one or two personal subjective observations.

When to remove a transient behaviour?

If we have dogs, what about cats, lions, tigers, snakes . . . or humans? Don't come here there is a nasty man? Why is he nasty? . .. . .
A whole host of issues about freedom of speech and consequential legal actions - one persons hazardous dog is another persons cuddly pet.

A highway with two lanes is a matter of fact, a hazardous dog is a matter of opinon and should not be in OSM.
--TonyS (talk) 11:14, 24 May 2020 (UTC)

I don't think mapping a dog_warning=* area just by subjective observations is in deed not a good idea. If there's a sign warning of dogs, it's then more like a fact. This also complies with the verifiability rule from our good practice page. So, mapping this area without any sign is in my opinion false data, since it's not verifiable.

--Hauke-stieler (talk) 12:03, 24 May 2020 (UTC)

As consequence, sign=dog_warning could be an option. Not dog_warning=*. --Nospam2005 (talk) 12:58, 24 May 2020 (UTC)

sign=dog_warning

sign=dog_warning would turn this proposal into a on the ground verifiable proposal. sign=dog_warning not traffic_sign=dog_warning as it's not about traffic. Other values could be used for similar not traffic related signs. --Nospam2005 (talk) 13:07, 24 May 2020 (UTC)

This would be good to tag the location of such sign using a single node. However, I don't think this is suitable for an area. I also think that the location of such sign is of lower importance, but the fact that dangerous dogs are walking within an area, is of higher importance. So I wouldn't replace the proposed dog_warning=* by the your sign tag, instead the sign tag could be some extra detail one could map. --Hauke-stieler (talk) 13:12, 24 May 2020 (UTC)

I second this. --Floridaeditor (talk) 20:39, 24 May 2020 (UTC)