Proposal:Shrubbery

This proposal is currently beeing re-opened. Work in progress

shrubbery
Proposal status:	Rejected (inactive)
Proposed by:	Cartographer10, JeroenHoek
Tagging:	maintained=no/yes/rarely/semi/intense
Applies to:
Definition:	A series of (sub-)tags to indicate how maintained a natural feature is. It is initially meant for mapping shrubbery as natural=scrub but can be used on any natural feature.
Draft started:	2021-02-21
RFC start:	2021-07-15
Vote start:	2021-08-03
Vote end:	2021-08-17

Proposal

Three (sub-)tags for natural=scrub are proposed that can be used to further refine these bushy areas based on objective criteria.

The following tags are proposed:

Key	Value	Definition
maintained	no, yes, rarely, semi, intense	Indicates whether, and broadly to what extent, an area of natural=scrub is maintained(i.e., manicured, pruned, or otherwise shaped and maintained by humans).
scrub:density	sparse, medium, dense	Indicates the effective density of the shrubbery.
scrub:shape	user defined	Indicates the shape an area of natural=scrub is sculpted into. This will in practice often be combined with scrub:density=dense and maintained=intense, and is typically used for hedges, shrubberies, and topiaries.

Status

An earlier version of this proposal proposed a new natural=* tag-value shrubbery. While many mappers did support the idea of a separate tag for the type of bushy greenery cultivated by humans, there was significant opposition too and the proposal was rejected. This iteration therefore aims to remove the biggest concern by approaching the problem from these basic principles:

• Accept that the de-facto scope and practical use of natural=scrub applies to all bushes, scrubs, shrubberies, and hedges
• Introduce optional, supplementary tags that document reasonably objective and observable physical properties of areas mapped with natural=scrub

By taking this approach, we hope to address the critical issues highlighted in the first iteration:

• Scope too wide: The scope and definition have been narrowed down significantly. By reworking this proposal to use sub-tags of natural=scrub representing physical properties of this base-tag, we hope to placate the mappers who feared that introducing a separate tag would lead to massive re-tagging and a huge grey area between natural scrub and cultivated shrubbery.
• What about barrier=hedge?: Instead of submitting the potential for these tags to provide a successor to barrier=hedge with area=yes in a separate follow-up proposal, it is now included here.
• If hedges can be tagged with this, can their shape be described?: We've taken this suggestion and tentatively suggest scrub:shape=*.

The results of the first voting round have been archived, and can be found here.

Rationale

Landscaped, cultivated, or manicured (i.e., trimmed and pruned) bushes, shrubs and shrubbery are a common feature in many built-up areas. These bits of greenery tend to serve aesthetic as well as pragmatic functions — usually acting as decoration, often acting as natural barriers between different uses of public space, and sometimes being nothing more than a way to fill some gaps in the layout of some grander landscaping scheme.

In OpenStreetMap this type of manicured greenery is overwhelmingly tagged with natural=scrub. However, this tag in itself does not tell data consumers anything about the nature of the feature. Originally intended for wild scrublands, it has come to be used for any and all scrubby/bushy/shrubby areas, including those actively managed and manicured by humans (what one might call shrubbery). For users orienteering themselves on a map the ability to distinguish between neatly trimmed bushes and hedges and wilder unkempt or even fully natural areas of shrubs can be of great value, and adds to the possibility to get a general feel for an area when consulting the map during (preparations for) a visit.

This proposal seeks to define maintained=* within the scope of natural=scrub. This will eventually allow renderers to distinguish between wild and man-managed natural=scrub, and give data consumers a tag with a clear unambiguous meaning. This tag could conceivably also be used with other features in the natural=* namespace and other vegetative features; the exact semantics of this falls outside of the scope of this proposal.

We also seek to offer mappers an alternative to barrier=hedge in combination with area=yes. While a barrier hedge mapped as linear feature can clearly be tagged as barrier=hedge, for shrubbery/hedges mapped as areas, the use of the barrier=* namespace is contentious, and different mappers will have varying interpretations. Instead, the sub-tag scrub:density=* can be used to describe the density of shrubbery that data users can then use to interpret the shrubbery areas differently. This makes the data more objective and allows for different interpretation depending on the needs of the data user. Also, for renderers this has the advantage that it removes the confusion about whether a barrier=hedge should be rendered as a line or area where for areas rendering in Carto was dropped.

For a more in-depth explanation about why not being in the barrier=* namespace should not be a problem for hedges mapped as areas, please see the → appendix.

Naming

There are several reasons why this key and value has been chosen:

• Discussion on the proposed tag landuse=shrubs revealed that the landuse tag is not preferred for this. It is not seen as an actual land use compared to, for example, a residential or industrial area.
• This tag describes a natural feature that fits within the definition of natural=*: “The natural key is used to describe wide variety of physical geography, geological and landcover features, including ones that have been modified or created by humans”. See also natural=tree, natural=water, or natural=beach which all can be both naturally occurring as well as human made. For a more in-depth explanation please see the → appendix.
• Another option was to keep tagging these shrubs as natural=scrub adding a sub-tag via scrub=*. The manicured, decorative shrubbery we seek to map are no sub type of scrub like what is the case with grassland=* or wetland=*
• Another option was to keep tagging these shrubs as natural=scrub + managed=*. The problem here is that managed is not the right key. For example, a nature reserve can be managed but the scrub in it can be left untouched.
• leisure=garden with some secondary tags was also considered. However, one might not see these shrubs and bushes as a leisure activity. Additionally, other renderers might incorrectly assume that these shrubs and bushes are gardens that people can visit as leisure.
• natural=shrubs was also considered. It is argued that shrubs is more commonly known compared to shrubbery. However, there are a few reasons to oppose for shrubs in favour of shrubbery. The word is very similar to the word "scrub" and "shrub" (both already accepted natural values). This creates the risk for confusion between the values. Secondly, all values in the natural key are singular while shrubs is plural. This makes shrubs not preferred for consistency reasons. At last, shrubbery puts more emphasis on the decorative and human factor included with the shrubs and bushes intended to map.
• natural=shrubbery remains the preferred tag for this proposal's authors and has its share of supporters, but introducing a new tag was one of the biggest points of criticism received in the first draft.
• The values no/yes/rarely/semi/intense are partly based on the current values for the key maintained=* (some other values exist, but their values make it seem likely that for some operator=* is meant instead).

The choice for the namespaced scrub:density=* and these three initial values was made based on the existing wood:density=* and scrub:density=* usage. The existence of these is also the reason we choose to use a namespace as well.

Tagging

This section introduced tagging guidelines.

Draw an area of natural=scrub and use the following tags to further define the scrub.

For the definition of maintained, heavy inspiration was found in a document by the Portland Bureau of Planning & sustainability

Key	Value	Definition
maintained=*	intense	intensely maintained areas of natural=scrub are typically actively pruned and maintained and kept in good shape. This includes neatly clipped box hedges, decorative shrubbery, and topiaries. These typically fulfil ornamental roles, or are used as barrier areas between urban features such as roads and recreation areas.
	semi	Like intense, semi-maintained areas of natural=scrub are typically planted and maintained by human, but are left to grow in a more natural form, and may appear unkempt or even natural. These may still be trimmed or weeded, and usually won't be allowed to grow wild beyond their boundaries.
	rarely	The natural features are wild and natural growth, however they are sometimes trimmed or otherwise restrained (for example to remove overgrowth to keep paths clear). Minimal intervention is the defining characteristic here.
	yes	The natural features shows clear signs of maintenance but the extend of which it is maintained is unknown. One of the more specific values above is preferred when observable.
	no	Natural growth. These are not (or no longer) maintained in any way.
scrub:density=*	sparse	Sparsely planted shrubbery. Generally speaking able-bodied persons could walk through them (although this is usually not desirable).
	medium	Shrubbery with a medium density. Able-bodied persons might be able to barge through it, but it takes some effort and might result in scratches or soiled clothing.
	dense	Densely planted shrubbery. Effectively impassible without damaging the shrubbery (which at this point often acts as a hedge) and getting your clothes torn.
scrub:shape=*	user defined	To indicate the shape of shrubbery. Typically this would only apply to scrub tagged with the tag maintained=intense. Examples include box for the typical box hedge, and any kind of figure one might prune a topiary into, such as dragon.

• The natural=scrub can of course also be further defined with suitable existing tags such as leaf_cycle=* and leaf_type=*, genus=*, and height=*.
• If distinct, separate trees are planted in a shrubbery area, the area could be tagged as natural=scrub + maintained=* with the trees as either natural=tree or natural=tree_row.
• For a linear hedge that acts as a barrier, use barrier=hedge instead.
• For vegetation where the bushes act as ground-cover for a forest or clump of trees, the tags landuse=forest or natural=wood are preferred.

maintained=* could of course be used for other natural=* features as well, although the exact semantics will likely vary.

Examples

Some examples of the different tag combinations are illustrated below.

Density	Intensely maintained (maintained=intense)	Maintained (maintained=semi)	Natural/wild (maintained=rarely)	Natural/wild (maintained=no)
dense
medium
sparse

Examples for which natural=scrub does not apply:

Example	Description
	This is an example of natural=heath

Opportunities for rendering

Any rendering style that wishes to support the tag with a mininal of fuss can opt to employ the same rendering as natural=scrub at first. Renders can also choose to render it differently based on the provided tags. Below we show an example. For Carto in particular, maintained=intense offers a chance to provide the much needed alternative to barrier=hedge with area=yes.

Example

Renderers can choose to distinguish between the different maintained=* values of which an example can be seen below. maintained=semi/intense could also be rendered similar in a shade and colour to natural=scrub, but with a pattern that evokes a slightly less ‘wild’ appearance.

Features/Pages affected

If this proposal is approved, wiki pages for maintained=*, scrub:density=*, and scrub:shape=* will be created.

Also affected are barrier=hedge and of course natural=scrub. barrier=hedge will be amended so that it can only be mapped on ways .

External discussions

The topic of this proposal was already partly discussed in the proposal for landuse=shrubs.

Comments

Please comment on the discussion page.

Voting

Voting closed

Voting on this proposal has been closed.

It was rejected with 25 votes for, 18 votes against and 2 abstentions.

Instructions for voting

• Log in to the wiki if you are not already logged in.
• Scroll down to voting and click 'Edit source'. Copy and paste the appropriate code from this table on its own line at the bottom of the text area:

To get this output	you type	Description
I approve this proposal.	{{vote|yes}} --~~~~	Feel free to also explain why you support proposal.
I oppose this proposal. reason	{{vote|no}} reason --~~~~	Replace reason with your reason(s) for voting no.
I have comments but abstain from voting on this proposal. comments	{{vote|abstain}} comments --~~~~	If you don't want to vote but have comments. Replace comments with your comments.

Note: The ~~~~ automatically inserts your name and the current date.
For full template documentation see Template:Vote. See also how vote outcome is processed.

• I oppose this proposal. By expanding the definition of a tag, countless uses of the tag that were made with the original definition will suffer information loss. Besides, the difference between scrubland and shrubbery is not only about being managed or not, but also about native versus alien or cultivated shrubs. --Dafadllyn (talk) 18:51, 3 August 2021 (UTC)
• I have comments but abstain from voting on this proposal. In my opinion, this creates a trolltag. The term "scrub" is specifically about wild plants, so including landscaped plantings (shrubbery) would re-define the term "scrub". It's a bit like calling a house cat a "tamed tiger", which, while sort of correct, is a dance around the fact that it's just a "cat". My preference would be to approve the original natural=shrubbery, however, I will not oppose this proposal if it finally gets us to unambiguous tagging for shrubbery that we can live with. --ZeLonewolf (talk) 19:00, 3 August 2021 (UTC)
• I oppose this proposal. In my mind scrub is, by definition, unmaintained, and I believe that this was the original meaning of natural=scrub tag in OSM. Thus I would not want to see the tag redefined and its meaning extended to cover maintained shrubberies. I would say that shrubberies are a distinct concept and so need to have their own top-level tag instead. Rjw62 (talk) 19:16, 3 August 2021 (UTC)
• I oppose this proposal. Like other above votes, I agree that natural=scrub is not fit for this because of how I, and many other people, think scrub is, unmaintained. I also am voting no because of how the shrub key was meant for scrubs out in the wild. I like the scrub:density=* key and I think it would be helpful for both shrubbery and scrub. If the key was different (natural=shrubbery) I think I might have voted yes. --BubbaJuice (talk) 19:41, 3 August 2021 (UTC)
• I approve this proposal. I prefer natural=shrubbery but I support this so that there finally is a way to tag the shrubbery that currently has no good tagging --Cartographer10 (talk) 19:46, 3 August 2021 (UTC)
• I approve this proposal. I too prefer natural=shrubbery but I support this because it is better than the status quo. --JeroenHoek (talk) 06:34, 4 August 2021 (UTC)
• I approve this proposal. People's opinions on this topic are divided strongly. Let this be the compromise. --501ghost (talk) 20:28, 3 August 2021 (UTC)
• I approve this proposal. This is not a grammar exercise. Tag is just a placeholder for whatever object or property we choose. --Mashin (talk) 21:22, 3 August 2021 (UTC)
• I approve this proposal. I'm fine with this. But won't it be a good idea to use direct meanings in maintained=* and scrub:density=* instead of general words? Like scrub:density=passable to scrub:density=impassable, and something in the lines of maintained=shape or maintained=topiary. ---- Kovposch (talk)

• I oppose this proposal. I'm sorry, because your proposal seems to have needed a lot of work to improve scrub tagging, and seems very well crafted to me. Nevertheless, the proposal treats the issue of using natural=scrub for areas of wood/forests in a single phrase, seemingly without much debate, whereas this issue is somewhat fundamental for mapping forests. The community was never, AFAIK, given a chance to take a decision about this issue: you have supporters and opponents to this custom, but no debate took place and no decision have been taken yet, because there never was a real debate. Treating this issue the way the proposal does it does not seem democratic enough to me. --Penegal (talk) 04:10, 4 August 2021 (UTC)

@Penegal: Could you explain what you mean by using natural=scrub for wood/forests? I don't think we are changing anything with regard to that in our proposal. --JeroenHoek (talk) 06:33, 4 August 2021 (UTC)

@JeroenHoek: After a clear cut, the land generally goes through natural regeneration; for several years, often one or two decades, it will be covered with young trees, and often also with other unwanted scrubs. The land will look like a scrubland for the time to main public, but foresters still consider it forest. Some mappers use natural=scrub to map such areas, reasoning that the area looks like a scrubland and is often left untouched for the whole time, until it can again be considered landuse=forest/natural=wood when the trees have grown enough. This mapping custom is disputed, and the debate is for now unresolved. Yet, your proposal states that For vegetation where the bushes act as ground-cover for a forest or clump of trees, the tags landuse=forest or natural=wood are preferred.; this sounds to me as a try to close this issue, without real, open debate. That being said, my interpretation may be incorrect: maybe you did not want to take part in this debate, but the proposal formulation tells me that you include in the proposal the interpretation Do not use natural=scrub for forests, an interpretation that has not yet be formally approved. --Penegal (talk) 17:11, 4 August 2021 (UTC)

Are you refering to this sentence "For vegetation where the bushes act as ground-cover for a forest or clump of trees, the tags landuse=forest or natural=wood are preferred."? that was not intended as any response on the dispute you are refering to. It was more meant as "if the forest floor is covered in bush/shrubbery then look at natural=wood or landuse=forest" instead, for example as on this image https://images.fineartamerica.com/images/artworkimages/mediumlarge/2/pine-forest-floor-anne-oreilly.jpg . So just a clarification which maybe I should have left out. Before you pointed it out, I didn't know of the dispute you refer to and that is also not something this proposal covers. --Cartographer10 (talk) 17:27, 4 August 2021 (UTC)

Oh, it seems that there was a confusion. Your picture is what most mappers would map with natural=wood/landuse=forest, whatever the bushes: they are merely understory, not a landcover by themselves, and should not be treated apart from the forest, as they are part of it. OK, I got it wrong, thanks for the clarification. --Penegal (talk) 17:56, 4 August 2021 (UTC)

• I approve this proposal. A welcomed refinement to natural=scrub, which allows to explicitly say whether it is managed or not. --Penegal (talk) 18:02, 4 August 2021 (UTC)
• I approve this proposal. --FERINA365 (talk) 09:24, 4 August 2021 (UTC)
• I oppose this proposal. In OSM natural=scrub is about natural areas being natural and unmaintained. --Reino Baptista (talk) 09:43, 4 August 2021 (UTC)

@ZeLonewolf:, @Rjw62:, @BubbaJuice:, @Reino Baptista: (and others who understandably raise this argument): Have you seen the actual, de facto use of natural=scrub though? A lot of it is used for maintained shrubbery. With this proposal you can at least mark them as such. A separate tag, natural=shrubbery, was voted down by mappers who believe the opposite to what you state: a separate tag should not be created. So what do you suggest instead? Without a solution to tag these maintained shrubs, natural=scrub (without any additional tags to help clarify its level of wildness) will remain the default for years to come, which to me sounds worse than this solution. --JeroenHoek (talk) 12:09, 4 August 2021 (UTC)

I agree with your frustration, and hence my abstain vote (unlike the others tagged) was specifically to not prevent this version from proceeding if there was community support for such a compromise. Quite simply, I would suggest there isn't a community consensus on how to tag shrubberies, and even worse, there isn't even community consensus that shrubbery is a separate thing from scrub. And so, I would suggest that there isn't a magical tag to be discovered that will satisfy the >75% of voters needed to get that green "Approved" label. Personally, in the absence of a community consensus, I will continue to keep using natural=shrubbery when tagging new shrubbery areas as I like this tag and it makes sense to me, and we have ATYL, and it can be easily changed later. Should this edition of the proposal fail, my recommendation would be to simply document natural=shrubbery on a page, document the fact that it is in use (because it IS in use, with > 1200 tags and growing), and note the failed attempt at proposal process approval. --ZeLonewolf (talk) 12:53, 4 August 2021 (UTC)

Well, that is certainly an approach we have considered, and it may be the best way forward should this proposal fail. We do see one huge downside though: indefinite lack of rendering in Carto seems likely without an approved proposal for quite a while (which also impacts the area-hedges rendering bug). This means that promoting natural=shrubbery might cause some strive amongst mappers working on the same area where one camp sees a bunch of shrubbery areas that aren't rendered and wish to retag, and the other wishes to build towards eventual support by numbers — which is a fair way to do it as you say. For a tag that has such a huge visual impact, this is a scenario we would hope to avoid. We've played with the notion of double-tagging with something like natural=scrub with something like actually:natural=shrubbery, but that too seems too confrontational. Getting support in (built-in) JOSM/iD presets and editor rendering is also a lot harder without approval. --JeroenHoek (talk) 13:35, 4 August 2021 (UTC)

In my opinion, it would be better for these areas to not render in openstreetmap-carto, than for them to (wrongly) render as wild scrub. We've already gained over 1000 tags so far despite it not rendering yet. So I think there is plenty of growth opportunity to justify adding tagging support in renderers that are being actively developed. I'm not really interested in catering to the whims of openstreetmap-carto as it is not being actively developed/maintained anymore and so it shouldn't hold back advancements in tagging. --ZeLonewolf (talk) 16:10, 4 August 2021 (UTC)

Quote JeroenHoek: "Have you seen the actual, de facto use of natural=scrub though? A lot of it is used for maintained shrubbery." That might hold for the Nederland, but it is quite the opposite eg. in Austria: query overpass-turbo https://overpass-turbo.eu/s/1am3 to get a list of top contributors, then each one can be quried like this https://overpass-turbo.eu/s/1am4 (replace Hungerburg with user account name). This is not a problem of understanding the prose of the Wiki, it is a problem with local customs in certain areas, Netherlands, Berlin also likely. No idea how mappers in India or … fare. It is also not a problem that is plaguing OSM for 10 years. In the Netherlands, its 2 1/2 years of explosion. I for one will welcome the proposal failing, as it would prevent a much needed tagging initiative, that targets micro-mapping pratices in a way that makes the work of cartographers easier, not more convoluted. "natural=shrubbery" was much better at this, but "shrubbery" is a leisure:garden:type, not a natural feature, so it got rightfully opposed. --Hungerburg (talk)

• I oppose this proposal. I don't consider shrubbery a subtype of scrub. Like some other voters, scrub is generally thought of as unmaintained vegetation. I like the author's preference of natural=shrubbery. --Dr Centerline (talk) 21:45, 4 August 2021 (UTC)
• I oppose this proposal. The proponents explain, that the new tags will heal a common error in tagging, namely, that planted, manicured shrubs are mapped as "natural=scrub". I consider the damage therefrom much less, than the damage of turning "natural=scrub" into a synonym for the failed "natural=scrubbery". As of now "scrub" has no accepted subkeys. The onesided approach of adding only stuff that applies to its "shrubberiness" (read the tagging section) does nothing but bolster ongoing misuse. No effort has been made to add anything useful to the "natural=scrub" key in its widely perceived original meanaing but the tag "maintained=no". Which is kind of dearth. --Hungerburg (talk) 21:46, 4 August 2021 (UTC)

@Hungerburg: So what are you going to do to solve the ongoing misuse? This has been going on for over a decade. Because this will only get cemented in further without any accepted way to tag the difference. --JeroenHoek (talk) 06:51, 5 August 2021 (UTC)

A number of values to the natural and landuse keys share the same fate. --Hungerburg (talk) 19:30, 5 August 2021 (UTC)

I don't quite see how that answers the above question. Can you explain? If you mean that you consider this problem (and of other tags like landuse=forest perhaps?) unsolvable, then I don't see why you would bother to vote at all. --JeroenHoek (talk) 05:56, 7 August 2021 (UTC)

I see the problem, and I see where it comes from, I just do not consider the proposed tags a viable solution. Others have been more detailed at that. Thank you for clarifying on the talk page, that, once passed, the proposal will make all mappings of uncultivated scrubland incomplete. The number goes into the millions. One more reason to oppose. --Hungerburg (talk) 16:11, 7 August 2021 (UTC)

Aren't they already incomplete? If the current situation is that shrubbery is mostly mapped as natural=scrub, it means that natural=scrub de-facto means "some kind of scrubs or shrubbery, i.e. anything higher than grass and lower than woods". --Westnordost (talk) 22:33, 8 August 2021 (UTC)

@Westnordost: I looked at usage in my local area: Out of 1240 scrubs, at most 10 are of the truly maintained kind, and quite small at that: a hedge drawn as an area, gardening in the center of a roundabout. Not much of an issue, I'd say. Also, quite recent. It started around late 2018 early 2019: https://wiki.openstreetmap.org/wiki/File:ScrubByNumbersNLvsAT.png - Especially the NL chart should give a good indication: numbers explode, area grows slowly. It is visible in the AT chart too, to a much smaller degree. I bet, in Austria 90+ % are of the documented kind, while in the Netherlands perhaps a quarter only. Not making maintained default to no, so to keep the original meaning of the tag, will cause a huge information loss on the one side, while adding nothing the micro-mapped scrubs anywhere else; (remember, they were wrong from the start.) It is just creating extr work, that will not happen. --Hungerburg (talk) 09:03, 9 August 2021 (UTC)

On the key values: I mapped some scrub today (eg. https://www.openstreetmap.org/way/973058107 a very small one, only 1531.5 m², but I consider this one important for way finding) - I did not tag maintained - according to the proposal, it would be "maintained=rarely", which I consider pure nonsense. I cannot map density, because density only applies to planted shrubs. These are not planted.

All in all, looking at the examples ZeLoneWolf posted to the "scrubs" article, reading through talks here and elsewhere, a simple barrier=shrubs for areas would cover 90% of the cases, that this proposal is set out to heal and might be much less controversial while quite up to the point at that. --Hungerburg (talk)

• I approve this proposal. I prefer natural=shrubbery, but it was rejected because many people wanted natural=scrub. Now natural=scrub will be voted and it will be rejected again because many people want natural=shrubbery. I'm voting yes anyway, in hoping that any tagging at all manages to get the necessary majority. The question is whether the necessary 80/20 majority is really useful in such cases. --Shaun das Schaf (talk) 08:48, 5 August 2021 (UTC)
• I oppose this proposal. REASON: too many elements which I consider "not OK". I have trouble calling the examples dense-intense ; dense-semi ; and medium-intense "scrub". Dense-intense is hedge (KIS: natural=hedge), dense-semi and medium-intense are areas of shrubs (KIS: natural=shrubs). Apart from that, I think maintained as a general key to detail some naturals is fine, though I have some doubts about the practical usability of the proposed value set for mappers and data users. The other two proposed keys use reverse namespacing, I don't like that. If you want to indicate density of a feature, just (KIS:) use density=* on the feature. No need to specify that density of scrub in a scrub area is scrub:density, that's just redundant. scrub:shape is typically aimed at using scrub for hedge, which I rather wouldn't. Lastly, using scrub for hedge areas and beds of shrubs colours the Carto map, but not in a right way. I have done it for a while, but stopped because it's just tagging for a renderer AND doesn't even come out right! Tagging the outline of a hedge area (that has a barrier function) as barrier=hedge is better. Too bad the hedge area is no longer supported by OSM Carto, but that is no reason to tag for the renderer or to redefine scrub, even though many mappers have done so. The simplest solution for hedge areas would be to support natural=hedge on areas. --Peter Elderson (talk) 09:09, 5 August 2021 (UTC)

There is no such thing as natural=hedge (at least not in wide use) just a barrier=hedge - and it works fine on areas AFAICT. --Stefanct (talk) 19:13, 6 August 2021 (UTC)

barrier=hedge is a way feature. If it is a closed way, you get a hedge outline on OSM Carto, not an area. That's why I would sugggest to simply tag a hedge area as natural=hedge, because that's what everybody calls it. Whether its function is barrier, decorative, separation, or filler, doesn't matter. It is not uncommon for a particular string to be a value for different keys. --Peter Elderson (talk) 19:46, 6 August 2021 (UTC)

• I oppose this proposal. I support the consensus of using natural=scrub as top level key. Even the subkeys and semantics chosen for the keys. However, definitions are poor and invite to a broad interpretation, don't describe the scope or intend well. Use of namespaces is redundant and unnecessary complex for subkeys which can easily be used on other natural tags. Lot's of the examples given are incorrect, they are natural=heath. So concept and allocation yes, implementation very poor. --Bert Araali (talk) 11:36, 6 August 2021 (UTC)
• I approve this proposal. Subtags for shape and density are definitely useful. As for the main tag, I could live with either natural=shrubbery or the currently proposed solution. If this proposal fails, I'll probably continue using the shrubbery tag. --Tordanik 17:41, 6 August 2021 (UTC)
• I have comments but abstain from voting on this proposal. First of all, thanks for your persistence! I recognize the need for better tags in this area and would very welcome a decent outcome of your efforts. However, I concur with some of the critique mentioned so far (notably: the bolstering the misuse of the existing natural key, the unnecessary use of a namespace, the (broad and unverifiable) definition of values). Especially the density does not make sense to me. If it is not easily passable it should be a barrier=hedge. I like the idea of the maintained key as it pretty much deals with the key issue here (which also is a problem for some people apparently :) However, there exists managed=* already that fits quite well apart from its documented peculiar values. You mention a conflict with it in the proposal that really isn't one AFAICT. Maybe we should try extending/refining managed=* next(?). --Stefanct (talk) 19:13, 6 August 2021 (UTC)

barrier=hedge rendering is currently broken when applied to areas in Carto, and will remain so. This is a threat to existing features mapped as area hedges (as explained in the proposal) leading to gradual data degradation (well-meaning mappers just change them to natural=scrub without any tags indicating the hedge-y properties). If you wish to continue that battle, you can you are welcome to try (this change happened one and a half year ago though, so you will be late to the party). The density-key is a way to sidestep this issue completely, and helps move beyond the binary 'hedge or shrubs' distinction. Of course, anyone who simply wishes to have an alternative to barrier=hedge for areas, could just use natural=scrub with scrub:density=dense and maintained=yes. --JeroenHoek (talk) 05:56, 7 August 2021 (UTC)

• I approve this proposal. In OSM we have to deal with the fact that basic tagging systems and keys/values/terms from the past turn out to be insufficient, inaccurate or otherwise problematic as time goes by - there are enough examples and the landuse/landcover or forest/wood issues may be the most well-known. There is - as the first proposal has shown - obviously no community consensus on terminology and our antiquated proposal process unfortunately might be a factor in the inability to find satisfactory solutions. Nevertheless, it is clear that natural=scrub is de facto used for all forms of wild and managed scrub and shrubbery (even if it was not originally intended that way in theory - but how should a common, botanically inexperienced mapper see the difference in case of doubt?) The proposal offers extensions/solutions for issues that have been discussed regularly for many years and that I have also constantly faced. So better an unsatisfactory solution than none at all for decades... --Supaplex030 (talk) 11:55, 7 August 2021 (UTC)
• I approve this proposal. --Timmy_Tesseract (talk) 10:05, 8 August 2021 (UTC)
• I approve this proposal. --Whatismoss (talk) 11:30, 8 August 2021 (UTC)
• I approve this proposal. --Lkw (talk) 20:02, 8 August 2021 (UTC)
• I approve this proposal. --Mstriewe (talk) 20:50, 8 August 2021 (UTC)
• I approve this proposal. To be honest, I don't really like it because it adds a lot of subjectivity. I'd prefer a tag that is boolean (shrubbery or shrubs, maintained=yes/no) because that is also easier to decide than between 4 categories. But anyway, some kind of approved categorization is better than none, I guess.. --Westnordost (talk) 22:24, 8 August 2021 (UTC)

• I approve this proposal. I prefer natural=shrubbery but what I like more is having something over nothing, then it's just a matter of wording really Crimsonfox (talk) 08:39, 9 August 2021 (UTC)
• I oppose this proposal. I've convinced myself otherwise after reading some of the other oppose votes, I think the difference between scrub and shrubbery is too different and it'll make a mess. My preference is natural=shrubbery Crimsonfox (talk) 11:31, 12 August 2021 (UTC)
• I approve this proposal. Nice improvement over the existing situation, but very clear definitions of the different maintained categories and density categories will be essential. --LeifRasmussen (talk) 14:33, 9 August 2021 (UTC)
• I approve this proposal. I've been micromapping in city enviroments for a few months, and there are multiple areas where I've used scrub. I think this proposed feature would work much better. --Aetiusjp (talk) 15:39, 9 August 2021 (UTC)
• I oppose this proposal. same reason as Dr Centerline pointed out --Strubbl (talk) 19:35, 9 August 2021 (UTC)
• I oppose this proposal. natural and maintained is a contradiction in terms and retroactively changing the meaning of a widely used tag is a very very bad idea. natural=scrub should be used for natural scrub, artificially maintained hedges should have a dedicated tag. --Nop (talk) 19:37, 9 August 2021 (UTC)

@Nop: What about natural=tree? The majority of mapped stand-alone trees are planted by man. Should this tag be limited to wild trees only? --JeroenHoek (talk) 06:59, 10 August 2021 (UTC)

@JeroenHoek: Yes, exactly. Actually, the abuse of natural=tree is a very good and sad example why I am voting with no. In its original meaning, natural=tree marked only significant trees, most of them decades or even centuries old and naturally growing. Then people started abusing it for micromapping everything: artificially planted, geographically irrelevant trees down to flower pot level, carelessly invalidating about 50000 existing meaningful entries in the process. I still see that as a form of vandalism by changing the meaning of a widely used tag. No need to repeat the same mistake again for natural=scrub. I would have been in favour of your previous proposal with a dedicated main tag. --Nop (talk) 11:24, 14 August 2021 (UTC)

• I oppose this proposal. Same reason as in the irst voting --Aeonesa (talk) 23:28, 9 August 2021 (UTC)

@Aeonesa: You didn't give a reason in the first voting round. Could you clarify why you oppose both proposals? --JeroenHoek (talk) 06:53, 10 August 2021 (UTC)

@JeroenHoek:As well as you accept votes without any statement, why they approve your proposal, you should also accept votes against your proposal, without any statement. There are written down enough reasons against your proposal. I agree to the most of them.--Aeonesa (talk) 13:27, 10 August 2021 (UTC)

@Aeonesa: I didn't make up that rule, it is part of the proposal process (have a look). You can vote for without comment, but when voting against an explanation is expected. --JeroenHoek (talk) 17:39, 11 August 2021 (UTC)

• I oppose this proposal. This is way to subjective. No one can accurately tell how often a scrub (hedge) is maintained or it's density. Let alone would two people have the same way of evaluating such things. I've tromped through a ton of scrub (hedges) in my life and I can't even give an accurate evaluation of either for any of them that I've gone through. Especially if I was going off the description for dense scrub (hedges). That it involves getting your clothes torn. I've torn my clothes on some pretty non-dense scrub (hedges) before. No one is to go through a scrub (hedge) to see how easy it is to rip their clothes so they can tag it with scrub:density=dense or not either. I also agree with other people that using a natural tag on something that's maintained is a little weird. Since that's not it's intended or original purpose. All that said, I'd probably support a simple maintained=yes/no tag without the other subjective values. Also, why not just allow for barrier=hedge to be mapped as areas? That would solve the problem this was created to. Adamant1 (talk) 03:28, 10 August 2021 (UTC)

@Adamant1: Will you be raising the issue of allowing barrier=hedge with Carto? Not having the reference renderer drawing them is one thing, but having the reference renderer draw them in a broken manner has an impact on its use. --JeroenHoek (talk) 06:56, 10 August 2021 (UTC)

@JeroenHoek: no. It's already been discussed and rejected on their Github page several times. So I don't feel a need to. Your free request it if you think hedges should be rendered/mapped as areas. In the meantime, I'm not really concerned with what they render. Nor should the potential of Carto rendering a tag or not be relevant to the merits of this proposal. Not that "Openstreetmap-carto" is Carto anyway. It uses CartoCSS as part of it's rendering back end, but outside of that they have no connection to each other. BTW, I can almost guarantee that if they aren't willing to render hedges as areas there's almost zero chance they will render this. You can't just create a whole new tagging scheme as a way to protest or get around them not rendering something. Adamant1 (talk) 01:10, 12 August 2021 (UTC)

@Adamant1: The Openstreetmap-Carto maintainers have explicitly stated to be willing to consider tagging alternatives for barrier=hedge with area=yes. That this proposal provides such tagging is an added benefit of this proposal that fits well into its scope. --JeroenHoek (talk) 11:40, 13 August 2021 (UTC)

@JeroenHoek: I'm aware of what they said. "considering tagging alternatives" is way different then "we will render anything that isn't barrier=hedge mapped as an area" though. I'm sure your aware of the many times they have said approval of a tag isn't enough on it's own for them to render the tag. There's also a pretty well established history of them not rendering (or removing rendering of) subjective, shoehorned tagging schemes like this one. So whatever they have said about "considering tagging alternatives" is irrelevant to the proposal. It still would be if they gave their explicit approval of the proposal though, because their opinions don't dictate what tags should or shouldn't be used in OpenStreetMap. It's not like they wouldn't also "consider" something more thought out and less subjective then this either. In the meantime, any proposal were the only response to criticism is "but, but, muh rendering!!" should be dead on arrival. Adamant1 (talk) 01:09, 14 August 2021 (UTC)

• I oppose this proposal. I still dont see the natural key beeing used for this kind of micro mapping individual plants or sub-square-meter areas as in the examples Flohoff (talk) 12:48, 10 August 2021 (UTC)
• I approve this proposal. --Roef (talk) 12:51, 10 August 2021 (UTC)
• I approve this proposal. There have already been a few comments on how this can improve the de facto status (instead of the "shrubbery"-tag) and I generally concur. I see this as a good opportunity to move forward. --kartonage (talk) 16:06, 10 August 2021 (UTC)

• I oppose this proposal. Ezekielf (talk) 17:25, 10 August 2021 (UTC) I was in favor of the previous natural=shrubbery proposal, but I do not support expanding the meaning of natural=scrub beyond wild or relatively unmaintained areas of scrubland. I consider the combination natural=scrub + maintained=yes/semi/intense to be a troll tag because scrubland is fundamentally un-maintained. I find the idea of using maintained=* on other natural=* features similarly problematic. Under the logic of this proposal, an urban park or garden with many trees could be tagged as natural=wood + maintained=yes/semi/intense, but these features are better tagged as leisure=park or leisure=garden. A maintained area of land covered in trees has a very different character from an area of woods/forest. Similarly, an area of neatly trimmed shrubbery has a very different character from an area of scrubland. In both cases using the same top level tag would not make for a useful data set.
• I oppose this proposal. streckenkundler (talk) 20:07, 10 August 2021 (UTC) scrub:density=*: is not detectable. It is 100% subjective. There is no general and easily usable scale of what you can or should take as a reference for a certain value. scrub:shape=*: What is that supposed to be...? maintained=*: Here you very quickly get to the point where a distinction is not possible or you could have multiple values in a smaller area. This is also extremely dependent on the plant material used or found. Whether a certain amount of maintenance then takes place cannot be assessed in the vast majority of cases. For users with a love of detail (micromapping), this would also open the door to a high degree of fragmentation in a small area.
• I approve this proposal. --Fizzie41 (talk) 07:07, 11 August 2021 (UTC)
• I approve this proposal. --Hike39 (talk) 09:44, 11 August 2021 (UTC)
• I oppose this proposal. There is too much of a difference between scrubland and a small area of trimmed bushes for these to be lumped into the same tag. I am in favor of a separate tag, of which natural=shrubbery seems to be the best option. The main argument for this proposal seems to be that currently many areas of maintained shrubbery are tagged with natural=scrub ( I have done so myself), but this proposal/compromise does not solve this problem. Any already mapped shrubbery will have to be fixed anyway by adding maintained=* tags, in which case you could just as easily change it to natural=shrubbery. --Msondergaard (talk) 07:57, 12 August 2021 (UTC)
• I approve this proposal. --EneaSuper (talk) 11:02, 12 August 2021 (UTC)
• I approve this proposal. To prevent senseless chaos: landuse=shrubs Natural=shrubbery landcover=shrubs ... Jo Cassel (talk) 11:56, 14 August 2021 (UTC)
• I approve this proposal. Not 100% convinced, but probably a step ahead. --EinKonstanzer (talk) 19:37, 15 August 2021 (UTC)
• I oppose this proposal. It would be better to use a separate landuse key for urban green areas overgrown with shrubs and bushes. I don't think it's a good idea to use the established first level key natural=scrub for such urban areas. Overall, I have the impression that something is being decided too quickly without sufficient international discussion. --Dx125 (talk) 13:01, 16 August 2021 (UTC)
• I approve this proposal. --OSMRogerWilco (talk) 16:14, 17 August 2021 (UTC)