From OpenStreetMap Wiki
Jump to navigation Jump to search

Moved values of "grassland=" to grassland=*

According to the usage in the wiki I moved the list with values of "grassland=" to grassland=*. --Rudolf (talk) 14:59, 6 May 2014 (UTC)

"Areas" or "Natural areas"

In this edit. Correct definition is "Areas", not "Natural areas". Xxzme (talk) 17:34, 19 April 2015 (UTC)

I would say this is the right place to discuss the matter, so thanks for creating the discussion here. I've tweaked the heading. My understanding is that this has been discussed elsewhere. I see you getting into arguments with User:Lyx : User talk:Xxzme#natural.3Dgrassland. I can't read Russian but I'm told there's discussion of this on this russian forum discussion and we've had complaints about your conduct on there. So there's people who do not agree with your edit (people preferring "Natural areas" over "Areas")
"Areas" was the definition discussed and voted upon during the proposal back in 2011. Then shortly afterwards User:T99 changed it to say "Natural areas".
Now on the one hand it might seem a bit cheeky to be making a change immediately after the proposal finishes, however I think it's quite common. Tags often need more exhaustive/improved documenting after a proposal process has finished. He/She probably considered it be a minor change refining the definition.
And nobody raised objections at the time. And then that remained for over 3 years. This means lots of mappers (and exponentially increasing community of mappers) will read and understand how to map based on this. ...So changing it back to the wider definition "Areas" at this stage is not necessarily the right thing to do. You could say there are 12 people, the people who voted originally, who prefer it to say "Areas". However that's not strictly true. The finer points of that wording were not particularly discussed back then. The wording on the proposal shouldn't be your sole justification for changing it.
Basically we should discuss this particular change afresh, and entirely separately from the 3 year old proposal vote. Should the natural=grassland definition say "Areas" or "Natural areas"? Am I missing any more discussion elsewhere?
-- Harry Wood (talk) 12:11, 21 April 2015 (UTC)
It is definetly cheecky (IMO) to change defition only at single page without contacting authors. See also Key:natural definition: "Where vegetation is dominated by grasses (Poaceae) and other herbaceous (non-woody) plants, except for ornamental grass (see landuse=grass), mowing for hay, etc. (see landuse=meadow) and grazing (see landuse=pasture).".
I can definetly say there wasn't agreement about to map it outside/inside place=* boundaries - I will vote against that definition.
Probably we want to start new proposal process and start new voting instead of changing definitiong back and forch to "accepted" and "more logical" definition.
At wiki, I just sync content: if it wasn't stated as "Natural area" and it wasn't stated as "natural area" at Key:natural I will copy accepted defition from the proposal.
I wasn't picky about changed defition if this tag wasn't accepted previously or it was in "draft" stage and so on, nothing special here.
> And nobody raised objections at the time. At the same time, Key:natural had different defition since Lzhl edit. It was corrected by PeterIto (2012-06-19).
While T99 intention may be correct, it is nohow okay to redefine tag after authors agreed about "wider definition".
We should start new proposal about 1. what do we mapped already, 2. how do we want to define this tag (wide meaning "Area" or narrow "Natural area")
with best regards Xxzme (talk) 14:08, 21 April 2015 (UTC)
this russian forum discussion is about how to map "grass" [note: title "газон" (something close to landuse=grass) isn't equal to "трава" *=grass]. Russian wiki wasn't updated to reflect all changes in Landuse. Particularly there people unaware about problem with landuse=grass is not about "natural" feature, but about "landuse" (this topic was covered at English Landuse). Here you can see me explaining changes happened at English wiki and Proposed_features/landcover. Mainly we discuss how to verify existing objects landuse=grass and what tag to use when we speak about "natural grass" (wich is simple to verify).
We actually haven't spoken about "Natural areas" over "Areas". It was just me verifing tag page and updating defition acccording definition in the proposal.
BushmanK has same nickname at forum, he haven't said any concerns about what each tag should be and reasoning behind his changes.
His comment refers to modifed version by T99. At the time he ignores proposal and definition at Key:natural. Xxzme (talk) 14:30, 21 April 2015 (UTC)
My reasoning was clearly described in comment to this rollback: Any existing and widely used tag should be used in uniform way (as possible). Otherwise, entities, mapped by the same tag in different countries, would be really different without any real reason. It's really bad practice to change or broaden the meaning of tag, when it's already in use for the long time.
Personally, I don't care about mapping lawns (especially, I don't care about discussion on Russian forum, where it turned into something philosophical, regarding to entities, which can't be verified). But I do care about broadening the meaning of this tag (which belongs to class of natural objects) which will lead to tagging man-made lawns or whatever.
If someone wants to have strict mapping scheme with crystal clear definitions, new tags should be used (as it happened, for example, with leaf_cycle and leaf_type) instead of changing usage of existing tags, especially, in a manner used by Xxzme - changing the description on a single page. --BushmanK (talk) 18:08, 21 April 2015 (UTC)
"It's really bad practice to change or broaden the meaning of tag, when it's already in use for the long time" I agree there's always problems doing this, but I don't think it should be absolutely labelled as bad practice. If, as a community, we can agree on changes to the docs, then hopefully it's not impossible to make changes (improvements!). But we can be a bit less sudden in our changes too. We can recommend people change over to a new tagging approach (if there's wide agreement on it), rather than just suddenly changing the definition. That also depends on how big the change is, and whether anyone really disagrees with it. If we're just "clarifying" the meaning of a tag, then we can be bold. There's shades of grey with all these things.
But you're saying you disagree with this change, and that you believe the mapping community have been following and should be following a narrower (natural) definition. ...And just to be clear, you're saying people should use a different tag, landuse=grass for tagging man-made lawns?
For me, the fact that this change is disputed, and does seem to be a change from what we might call the "de facto" accepted definition (because people were happy with it saying "natural" over the past three years), we should revert, at least for the moment. And then discuss whether the definition should be wider
-- Harry Wood (talk) 13:19, 23 April 2015 (UTC)
I'm not trying to solve two problems here, therefore I'm intentionally avoiding forking discussion towards proposition of landuse=grass or anything else, so, please, don't push it there. I'm only trying to prevent breaking the existing data consistency. Current scheme can be far from ideal (which is typical for old tags), but changing it is the whole other deal.
I disagree with this change by these main reasons:
First, it was made on single language page, which will lead to mapping different objects by different communities, which is definitely bad for data consistency, as I already described above. Even users of iD, will see different descriptions depending on language, set in their browsers. For sure, I understand, that any clarification or change, supported by the whole OSM community is legit with no question.
Second - because this change broadens the meaning towards man-made objects, related to grass, while keeping the grassland a value of natural=* tag, it will, for sure, make a lot of confusion for users.
Third - this change was not made with intention to clarify anything. The only reason was "political" - it was made in the middle of discussion of lawns, grasses and so on, where Xxzme (using nickname d1g) is just trying to support his own view. His change of grassland definition is solely serving his private view, not the community view. This is not how Wiki should be edited. --BushmanK (talk) 16:52, 23 April 2015 (UTC)
I also have to add, that for English-speaking community, some tags are much more self-described, while in Russian-speaking community there is a tendency to pick any barely similar tags for mapping objects with almost nothing common with original meaning (because people are somehow afraid of making new tags for new entities).
Therefore, being a value of natural=* key, grassland value will not necessarily be used for natural objects only, if this requirement wouldn't be clearly described. Good example of broadening the meaning of tags in Russian-speaking community is natural=cliff, which is widely used for any kinds of slopes instead of explicitly vertical rock exposures. Therefore, any shortening of Russian description potentially leads to unacceptable broader usage of tag. --BushmanK (talk) 18:47, 21 April 2015 (UTC)
Oh definitely I think we have a lot of work to do nail down the definitions of tags a lot more clearly for the benefit of foreign language speakers. Actually just adding the word "natural" in the sentence is not really as clear as it should be. If this distinction between natural=grassland and landuse=grass is important, then it could be documented more clearly and explicitly -- Harry Wood (talk) 13:22, 23 April 2015 (UTC)
If I were to encounter a way tagged natural=grassland (anywhere in the world), I could only be certain that it must contain (mostly) grass. The keys landuse and natural describe orthogonal attributes: what is there, and what is the area used for; even if it's sometimes highly likely the landuse implies some "what is", or the other way round. Alv (talk) 18:25, 21 April 2015 (UTC)
> a way tagged natural=grassland (anywhere in the world), I could only be certain that it must contain (mostly) grass. For me it also irrelevent is this grass "natural" or "not". How can one distingush "Natural grass" from "grass" or "Natural area" from "Area". See also managed=*. Xxzme (talk) 19:44, 21 April 2015 (UTC)
This problem can't and must not be fixed solely by changing tag description, especially, on the one language page only (which serves only someone's selfish interests). Need better tagging? Introduce new tags. --BushmanK (talk) 20:08, 21 April 2015 (UTC)
As mentioned above, we might solve problems by introducing changes carefully with consultation, and placing suggested changes to tagging practice on the page rather than suddenly redefining things in a way which people disagree with. We don't necessarily need to introduce new tags in order to fix a definition. Actually it seems Xxzme is arguing for fewer tags. A consolidation if you like. He's saying we should use natural=grassland for all kinds of grass. Is that right Xxzme? So are you also suggesting that we should be phasing out landuse=grass for lawns then? -- Harry Wood (talk) 13:37, 23 April 2015 (UTC)
Term "natural" was never defined in the proposal or at tag page.
After long debate that you cannot formalize "lawn" or use landuse=grass to determine physical grass we ended up with following tagging (most) users agree:
previous landuse=grass tag will be untouched (since they are rendered ATM)
natural=wood / natural=grassland will mean "there grass/woody area"
managed=yes will further clarify "real maintenance"), managed=no will mean "natural" for those who care about "real maintenance"
operator=* will mean if there operator company or not (operator=no - "there no company", operator=yes "this object operated by unknown company") - it is orthogonal to managed=* and IMO very few mappers will use that tag
surface=grass will mean "surface", as before. Most mappers agree that previous tags surface=*[1] landcover=*[2] to determine whatever actual grass present or there actually "mud" "dirt" "ground" etc. Xxzme (talk) 13:57, 23 April 2015 (UTC)

This topic is not new:

Revert to the original state?

Hi, I propose to revert the page to the state before Xxzme edits. It is very diffucult to understand what it means in the current state. Any comments? Chrabros (talk) 15:07, 10 November 2015 (UTC)

I have that impression too. I find his additions confusing. I have no idea why he mentions that "the idea ,,, is not new", and why he distinguishes simple and precise mapping. --Fkv (talk) 21:10, 5 July 2016 (UTC)


The photo does not fit here, because there are bushes in the background. It's certainly a Central European meadow, which would quickly turn into woodland if it were not regularly mowed. --Fkv (talk) 21:27, 5 July 2016 (UTC)


The description doesn't say how to distinguish natural=grassland from natural=fell, and whether natural=grassland is a superset or relative complement of natural=fell, natural=heath, landuse=meadow and landuse=grass. --Fkv (talk) 21:36, 5 July 2016 (UTC)

The definition for natural=fell is not very precise. It is used for alpine (high-altitude) areas covered by grasses, dwarf shrubs, bushes, shrubs and wetland, such as bogs or marshes, sand, scree or bare bedrock. It's also used for hill areas in Scotland with similar characteristics. The one certain thing is that natural=fell is not usually tagged on areas covered with trees. Many areas of natural=fell could also be mapped with areas of natural=grassland when they are covered with grasses. However, check that there are not areas of natural=wetland such as marsh, or patches of rock or bushes, first. --Jeisenbe (talk) 05:47, 10 August 2019 (UTC)