[Tagging] Feature Proposal - RFC - Shrubbery V2

Bert -Araali- Van Opstal bert.araali.afritastic at gmail.com
Thu Jul 15 16:01:28 UTC 2021


I admire your persistence. The effort you put in this proposal is above 
average and shows your dedication and support for OSM.

Sorry for starting a new sub-thread, but it seemed appropriate to 
express as subjective as possible my comments on the large content in 
this proposal rather then biased personal opinions and the reactions on 
those as in the initial opening.

This mail might be rather long, but long proposals ask for long answers, 
out of respect for the large amount of effort put into it.

Abandoning the concept of a new top level tag surely has my preference.
Still, some issues raised by me and others in the previous versions 
still persist.

1. semantics: "cultivated", used in a context of "growing" applies to 
the land or in a context referencing the land and crops or plants grown 
upon it as a group. It is used in some definitions across OSM correctly. 
You came close to a much better and correct semantic term: manicured, it 
might be the gem we are all looking for. It is a common term used in 
gardening and the shaping of vegetation in general. Any English 
dictionary can be referenced to support this. So I suggest to use 
"manicured" as the attribute key instead of "cultivated". Confusion 
throughout the proposal as you use "landscaped", "cultivation" and 
"manicured" in the same context to refer to the practice and vegetation 
you are really targetting: "landscaped" refers to land AND vegetation; 
"cultivation" refers to land OR land with it's vegetation in a grouped 
context; "manicured" refers to vegetation ONLY. Please don't mix them.

2. scope:
2a. you describe it to be used as attribute tags (you use the term 
sub-tags which is fine for me) for natural=scrub. It can be used however 
on a much larger scope of OSM top level tags. Essentially they can be 
used on any tag applicable to vegetation. natural=shrub, natural=heath, 
natural=grass, natural=tree_row, landuse=forest; barrier=hedge etc... In 
order to get support and approval for this proposal and not make the 
scope to wide, you describe this as intentional. I fear however this 
will have the opposite effect as it is strongly related to f.i. 
natural=shrub. Why not be more bold and describe and test it's more 
general applicability? I am resistive to any proposals that try to 
address or invent tagging schemes that are generally applicable, to 
address specific or niche needs. It grows the need for the use of 
namespaced keys instead of more general applicable simple keys.
2b. Manicured goes beyond a practice by humans. Also animals manicure 
vegetation for the purpose of producing food, protection of their 
nesting places, and even decorative purposes. Often the difference is 
not obvious for some mappers. To avoid ambiguity and support subjective 
verifiability, I would prefer a definition that avoids terms referencing 
only human practices.

3. gender neutrality: the use of "man" is generally discouraged. Use 
"humans" instead.

4. barrier=hedge. Despite you say that you want to limit the scope, you 
include a new sub-tag scrub:density, with the intention to make 
barrier:hedge less contentious. I am afraid you are going to achieve the 
opposite. First of all this is a completely different issue, not related 
at all to the shape or anicure of vegetation. It's a rendering related 
issue, we don't tag for the renderer. Secondly, even if you disagree 
with the fact that it is purely a rendering issue, you propose a tag 
that is not specifically targetted at areas and highly subjective. The 
appendix made it clear what you intend to tag. Density in regard if a 
barrier can be safely passed. Density is not deterministic to express 
this. A thorny hedge or hedge made of toxic plants can be very effective 
as a safe non-passable barrier, even at low density, yet very 
ineffective for privacy. A very dense hedge made of vegetation with very 
flexible branches easy and safely to pass but being very effective for 
privacy.
The choice to use a similar namespaced key for this like existing 
wood:density and scrub:density seems to avoid ambiguity, but I think 
it's the contrary because they originate and are used in a different 
context.

5. appendix.
5a. You conclude that natural=tree is pre-dominantly used for trees 
planted by humans. This might be the case in your area but I doubt this 
can be used as a global justification. In areas where I map it is most 
dominantly used for trees that have grown naturally and often very old. 
They are the remains, wherever located currently of natural untouched 
landscapes and often protected by legislation, so the opposite of your 
statement.
5b. The rendering as an area in Carto is not related to barrier=hedge 
only. The same issue applies to barrier=wall. You conclude that the 
motivation to do this by carto brakes rendering of those barriers. 
Although this is true the key problem and reason, of course to my 
analysis, is that bariier tags are out of laziness often applied to 
areas which require different rendering, like landuse and amenity tags. 
The issue is nt the tagging, but the mapping practice. Linear walls and 
hedges should be separately mapped by us in OSM, same to walls and 
hedges which are intended as larger areas. Even if they coincide or 
overlap with other areas, they should be mapped separately.  It's very 
difficult to resolve a bad or inconsistent mapping practice with a tag. 
So this one will not do it either.
5c. We don't bow for Carto's demands, we don't map for the renderer. I 
am happy and appreciated Carto's comments, as it painfully illustrates 
inconsistencies in mapping practices in OSM and our failure to do 
something about it.

Scrub:shape is the only remaining. Why again a namespaced tag for this 
purpose?

Greetings,

Bert Araali


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20210715/1a0e7751/attachment.htm>


More information about the Tagging mailing list