Add rendering for historic=ruins #331
Comments
|
Yes, there is a need for this. historic=ruins should render on the map. |
matthijsmelissen
changed the title
|
I guess dashed outline (as mentioned here, citing OsmAnd) could work as a visual hint. |
|
I don't think that works for |
|
sent from a phone
what s the difference to archaeological site? Maybe we can use the same rendering? |
|
What about this icon of ruined castle, proposed for Osmic by @MaestroGlanz: |
|
Example rendering on z17: |
|
The separate shape above is implausible, what is it supposed to be? |
|
I agree that ruined castle icon is not working in practice. I guess we can use the idea of decomposed shape for something simpler, like house. |
|
It is directly derived from the castle icon. I could change it to a version where the parts are joined. |
|
Please try, we're still looking for something good enough. |
|
FWIW (different map) for nodes I went for the name and a dot, in a "historical" colour https://map.atownsend.org.uk/maps/map/map.html#zoom=17&lat=53.429972&lon=-1.26303 and for ways the name if present and a "not quite a building" colour https://map.atownsend.org.uk/maps/map/map.html#zoom=19&lat=53.122449&lon=-1.853998 and https://map.atownsend.org.uk/maps/map/map.html#zoom=20&lat=53.2699203&lon=-1.989286 . The actual colours won't transfer to OSM Carto, but I'm not convinced you need a "characteristic" icon - the name normally describes the thing well enough. |
|
I have two other suggestions here: You can check the appearance on I used 16px grid. At least, I think so. |
|
Still not recognizable for me at 14 px. |
|
I created other versions: All together on https://wiki.openstreetmap.org/wiki/User:MeastroGlanz#Icons_I_created Do you have any suggestions of mixing some of the ideas? |
|
I prefer version 3 the most. Btw. are there any icons, which should be created? |
|
Nice try, but none of them work for me at 14px (32 px would be OK). I was thinking about solid, easy to recognize shape like a simple house as a base. |
|
I like Building_ruins_generic6.svg... and Building_ruins_generic4.svg as well. |
|
@PontiacCZ Try to look at 14 px version only, because that's the only thing the user will see. |
|
Yep, that's what I actually did, I compared icons in the "Icon original size" row on the summary page and my Firefox tells me they're 14 px wide. Well, the more I look at the variants the more I like I am not sure, MastroGlanz, what you mean by version 3 - is it But any of those would be better than none. BTW kocio-pl, you have mentioned that "one of them work for me at 14px". Which one is it? |
|
I meant "none" - sorry for the typo. 6 is a nice concept, but the details are too small for me, so it looks just like 2 rectangles, with no connection to buildings. |
|
sent from a phone
On 12. Nov 2020, at 18:24, Joseph E ***@***.***> wrote:
When there are 2 or more ways to tag the same feature, as in this case, we need to wait until mappers and the community decide on which method should be supported, as shown by current usage in different countries.
historically significant ruins are always archaeological sites, and this style tries to avoid rendering several tags with the same meaning. For other ruins we would want a different rendering anyway.
|
I disagree. I've always interpreted a historic=archaeological_site as something known to have undergone a documented archaeological research, and possibly still marked as such in situ, with a information board or a fence. Historic=ruins is used for ruins of lesser, former or unknown archaeological interest, but which are nonetheless important local landmarks. I cannot think of a better tag for several objects I mapped recently, e.g: https://www.wikiloc.com/hiking-trails/fruska-gora-popovica-okretiste-raskrsnica-borove-sume-vaga-raskrsnica-borove-sume-jezero-popovica-o-29555178/photo-18887728 https://www.wikiloc.com/hiking-trails/ps-fruska-gora-170205-crveni-cot-sajlovac-gradac-potoranj-stranputica-tancos-krug-16339306/photo-10245666 https://fruskac.net/en/locations/misc/old-church-saint-george Historic=ruins is documented in OSM wiki, supported as an iD preset, and, with near 150,000 objects, it is one of most popular tags not rendered on Carto. And I don't think those are going away anytime soon. |
|
I made 13 suggestions here. I think, there should be a one fitting: If you want another one, another style, let me know. |
|
These three objects are quite different: 1) "Remnants of quarry cable car station”, 2) "Late Roman-period military outpost hidden in a remote forest”, 3) "ruined late medieval stone church”. Consider that the last could also be tagged abandoned:building or abandoned:amenity. I’d suggest discussing how to tag these on a more appropriate forum, such as help.openstreetmap.org or the Tagging mailing list. |
|
A 40-story skyscraper, a 10,000 m2 shopping mall and a backyard shed are also quite different objects, yet we permit tagging them all with building=yes. Conceptually, historic=ruins are not much different, "here are the ruins of a rather old object", and one can apply other tags to describe them closer. I do read (but do not participate in) tagging and talk mailing lists, and, with respect, I don't remember anything ever solved there. I think that moving discussion to another forum would be a good way to not solve it for another 6 years. With over 100,000 uses, documentation in wiki and support in iD, I believe mappers have already "voted with their feet" about the issue.
It certainly could, but that would not render in Carto either. |
|
sent from a phone
On 13. Nov 2020, at 09:49, DujaOSM ***@***.***> wrote:
A 40-story skyscraper, a 10,000 m2 shopping mall and a backyard shed are also quite different objects, yet we permit tagging them all with building=yes.
we do not advocate or encourage tagging them the same
|
Certainly. But we render them nonetheless. |
|
@DujaOSM - the meaning of If you think |
|
@imagico I'm afraid you're asking for shrubbery. How can I prove or disprove that
Besides, what this has to do with rendering or not rendering in Carto? I really did not come here for that discussion, but to report an issue that irked me for a while (and I found this one by searching). In my opinion, the default and flagship map style of the project ought to render all reasonably common and on-wiki documented map features, regardless of any ongoing disagreement about the exact scope. I don't really care which visual style is applied as long as the object is somehow rendered on the map. |
Well - that is not going to happen as @jeisenbe explained. Again - we invite you to argue for feature addition under the premises and goals of this style but if you don't want to do that it will have to wait until someone else does. Yesterday i presented some numbers from taginfo (supplemented by @dieterdreist) that could be used (in combination with the discussion in #4238) as a starting point for sorting out how different tags in the historic/abandoned/ruined structures field are practically used. This is all no rocket science, taginfo and overpass turbo are your friends here (also geofabrik's regional taginfo instances can help you with analyzing regional differences). Sometimes you might need to run a planet file through osmium for some more detailed stats. And yes, geographic knowledge is useful here as well as the ability and willingness to discuss tagging with other mappers on the usual channels. You write:
Well - we do and we take our responsibility to produce a style of decent cartographic quality and pursuit of our documented goals seriously. So our perspectives here differ. If that causes incomprehension and annoyance on your side i can relate but this does not change the situation in substance. |
|
"An engineer is a person that makes something work. An expert is a person that can thoroughly explain why something cannot work." Going back to my engineering business, then. Sorry for wasting your time. |
|
Am Fr., 13. Nov. 2020 um 17:09 Uhr schrieb DujaOSM <[email protected]
:
Going back to my engineering business, then. Sorry for wasting your time.
although this was funny, I have been left with the impression that you are
frustrated about the outcome of this discussion. I am sorry for this. I
have myself been asking for rendering this tag many years ago, but from my
current understanding, it should not be rendered, because it is not applied
consistently, and because the case that is literally described in the wiki
"The tag historic=ruins is used with ruins that are of historic importance,
where it is not possible, or not appropriate, to be more specific about the
type of structure that is now in ruins." is completely covered by
historic=archaeological site. Just nest archaeological sites if you have
one inside the other.
Also note that "historic importance" is a subjective judgement, and not
useful as a requirement because it will only lead to edit wars and will be
hard to decide about. "Significance" is already determined through the
mapper (if she thinks something is not sufficiently significant, she will
not map it). Our main criterion is "verifiable on the ground", "importance"
on the other hand seems to open a way to dismiss things that are verifiably
existing on the ground, as "unimportant".
|
|
We can argue all day about differences between Accidentally, I've just stumbled upon this:
This morning I found out the location of those Roman-era ruins so I wanted it to map them and show to my friends, so that's what brought me here. I'm now going off to tag those for the renderer. Building=whatever will do, I suppose. |
|
just go with "historic=archaeological_site"
|
|
I gather that historic=ruins is not rendered because there's already historic=archeological_site. So, in your views, should the ruins of a church destroyed by Second World War bombings tagged with historic=archeological_site? |
|
I don't know how things work in Germany, but there'a a couple of places from the early 1900s that are archeology sites near where I live in California. I'm pretty sure there's a submerged town from the late 40s a few miles away that's also an archeology site. It doesn't really have anything to do with age of the site, more if it can be excavated for artifacts or other physical objects for the purpose of analyzing them. So I'd assume a building bombed out in WW2 would count. At the end of the day it's probably still archeologists doing the excavation and analysis if someone wants to study the area. |
No, historic=archeological_site is rendered here because it has been rendered in this style predecessor already. No tags indicating ruined/historic structures are specifically rendered so far because the tagging used for those is complicated and not very well defined and no one so far has invested the time to look at this in more depth and analyze how the various tags in that context (see #331 (comment), #331 (comment)) are used and come up with a rendering concept that reflects that. This issue is open, work on solving it is welcome. What is needed here is:
Neither of these tasks is rocket science. But it requires some ambition beyond the level of i want this tag rendered, i don't care how. |
|
Yes, historic=ruins is a tag that is used for all kind of things, from the insignificant abandoned and collapsed shed on a field to historic ruins of bigger buildings and likely of more general interest. My advice is to avoid the tag where alternatives are available and prefer historic=archaeological_site over historic=ruins whereever possible. It also has more subtags for additional description and classification. There are also other (more specific feature-) tags under the historic key that sometimes may fit well. |
|
@dieterdreist - please don't turn this into a tagging discussion. Use number for historic=ruins and historic=archaeological_site are fairly similar - see #331 (comment). And i would not say from my superficial impression that one of them is necessarily used more consistently than the other. We do not tell people here to use one tag over the other, just because it would be convenient for us because we happen to render one of them for historic reasons and not the other. If there are reasons to definitely avoid rendering historic=ruins based on practical use of the tags please present them. Otherwise my suggestion to analyze the tagging practice in this context and to develop a rendering scheme based on that stands. |
|
@imagico Propose a rendering rule, I would say, then we can discuss it. I suggest like this:
The goal must be to render ruins, that
Historical value defined as
The second one is not razor sharp, but we can not get this information from the tagging directly anyway. The question is, where is such a meaning implied by the tags and how. This would be also a topic for @westnordost and streetcomplete: Is this ruin really a historical landmark? With result as last_checked-tag. Or is_landmark=%date by %user. There is certainly a better solution. |
Neither of those things are implicit in how the tag is currently being used. All the Wiki article for the tag says is "The tag historic=ruins is used with ruins that are of historic importance." There's nothing about the ruins having to be designated as such by a public authority or that they have to be on some listing somewhere (I don't know how anyone could confirm either thing in most cases anyway). |
|
Why don't we redirect this issue towards rendering combinations like historic=ruins + ruins=castle;fort;church. That way we at least have historically significant poi's covered. |
|
Ran into this problem. Gave up and searched for an issue. Why? How does this work? Does anything change? |
|
Current state of this issue and what needs to be done to solve it is clearly documented in #331 (comment). |
|
@BertMule With regard to the first part of #331 (comment) :
I've done that for usage in UK/IE, and the resulting processing can be seen at https://github.com/SomeoneElseOSM/SomeoneElse-style/blob/master/style.lua (search for "ruins"). To summarise, it really doesn't make sense to render My implementation is mostly in lua (which this style uses for some processing, but not much). I suspect you would be able to do at least some of that in select statements in the .mml / .mss files that this style uses, but I certainly won't be volunteering to do that. However, you (or anyone else) wants to take this work on I'd be happy to explain how to do the "analyze tag use in OSM" part (and to share what I did myself previously). To see what this looks like try an overpass query like this one and, click through to OSM, and move the map slightly to get the "map=" layer and co-ordinate details. Then use that same value at map.atownsend.org.uk to see the results. |
Although we of course need to look at global use and not just at a small region it would be helpful - not only for OSM-Carto but also for data users in general - if you'd publish the results of your analysis so people would not need to try to infer that from the tag interpretation logic you use after reconstructing that from some lengthy spaghetti lua 😉. Learning (and teaching) how to determine quantitative data on use of tags is valuable. But at least as important is learning how to analyze this data to determine established mapping practice and de facto meaning of tags. The latter in my experience is mostly learned by example - having people experienced in the field demonstrate how they look at the data, analyze it and come to their conclusions. I try to do that here as much as possible, but it would definitely be of high value if others with a different perspective and different communication style would demonstrate their approach more often as well. |
Neither POI or area/multipolygon ruins are currently rendered. http://www.openstreetmap.org/note/114642 brought this to my attention, as the castle named in the note is already mapped as both POI and multipolygon.
For POI, perhaps just the name rendered, and for areas - well I'll let you pick a suitable colour. The wiki also suggests a lot of ways are also tagged as ruins - perhaps for these render a bit like generic barriers?
The text was updated successfully, but these errors were encountered: