-
Notifications
You must be signed in to change notification settings - Fork 822
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add rendering for historic=castle #744
Comments
Would be great. German OSM-Style uses the following symbols: French style uses one symbol for all castles: http://openstreetmap.fr/calendrier#zoom=18&lat=47.55786&lon=10.75013&layers=B0TTT code see here: https://github.com/cquest/osmfr-cartocss/blob/master/amenity-points.mss |
+1 |
-1, it is a bad tagging scheme, castle is a very poor name for "A fortified building in medieval and modern times" ( from http://wiki.openstreetmap.org/wiki/Tag:historic%3Dcastle ). In my region there are many forts and bunkers tagged with it and making it displayable in this style would ensure that this tag scheme will be not fixed for a very long time. |
You can fix them, e.g. there are "fortification" and military=bunker tags. Displaying something will often lead to fixing mistagged objects because people become aware |
I think that fixing http://wiki.openstreetmap.org/wiki/Tag:historic%3Dfort would be a good idea - but as it requires translation from German I am unable to do it. |
Note, It depends on #565. I propose to render it without icon - but as a special building. Like place of worship on the attached image. https://cloud.githubusercontent.com/assets/899988/4741530/2ed1b746-5a17-11e4-982b-ccac76eb2b31.png |
2014-10-30 8:21 GMT+01:00 Mateusz Konieczny notifications@github.com:
in German maps castles are typically divided into ruins and castles and are |
Note, I am no longer considering rendering it as a special building to be a good idea. historic=castle tagging scheme is a poor tagging scheme and IMHO should not be supported. |
How about castles tagged as nodes (11 016 uses) ? and castles without building=* (16 143) ? |
Because historic=castle may be the whole area of the fortification/palace grounds and is comprised of multiple buildings. |
sent from a phone
On 6 Jan 2017, at 03:23, aceman444 ***@***.***> wrote:
Because historic=castle may be the whole area of the fortification/palace grounds and is comprised of multiple buildings
+1, a medieval castle is usually composed of several buildings and often several rings of walls, even the small ones.
|
building=castle would be of use, we can also render historic=castle+building=*, when the proper tagging is chosen. For castle area however or nodes we can find different rendering ideas. School example:
It doesn't mean it's how we want it to be in the future, it's just a starting point for further ideas for castle rendering. |
sent from a phone
On 6 Jan 2017, at 23:37, kocio-pl ***@***.***> wrote:
For castle area however or nodes we can find different rendering ideas.
render a name from z15 and maybe an icon, z17
|
This icon might be good as default symbol for a castle: I guess Tagging discussion is needed and Wiki changes for a start. This can take a long time before it'll be a standard of tagging, and we can use this time to design rendering strategy for all the cases (node, building, area), no matter what tagging scheme will be chosen in the end. |
Currently, |
There is discussion about historic=castle tagging on german forum (https://forum.openstreetmap.org/viewtopic.php?id=57896). People agree that the tag as area means the whole ensemble. This would include the courtyards and even some modern-day additions like a kiosk. Buildings or ruins have to be tagged separately. A technical question arised if there are no surroundings and the sole building is tagged as multipolygon because of a courtyard (example: https://www.openstreetmap.org/relation/22988). Is putting historic=castle on the outer way of the building multipolygon a good idea technically? Because multipolygon rules allow it but I'm unsure whether the new lua branch would support it. |
I see little reason to change the tagging in your example, and no need to add the historic=castle tag to the outer of the MP. Everybody understands the courtyard of the castle belongs to the castle in this example. It would certainly be undesirable to add both historic=castle and building=yes on the outer as well as the MP. Depending on the renderer, you would end up with two buildings stacked on top of each other, with the building derived from the outer blocking the view of the inner court, as it won't have one. |
For me it's not about understanding but good tagging. Currently the courtyard is excluded from the castle area.
The building tag would stay on the MP. |
This is just one interpretation, I don't think you can really say that either way is good or bad in this particular example, or that people would not generally understand the courtyard is part of the castle. Undoubtedly, this has been discussed at length in the German forums, but unless you are planning to calculate metrics on castle sizes including courtyards, I still see very little reason this tagging would need changing. As to a specific argument in favor of maintaining the current tagging: if some style would like to emphasize castle buildings by use of a specific color instead of using a "castle" POI, only putting the tag on the outer would not allow this or at least make it much more elaborate to accomplish. I am not saying that your suggestion is bad here, I just want to illustrate with the above example that either way has consequences for the potential cartography. It is up to you to make the choice for the tagging, and than accept the possible outcome of the cartography in your favorite OpenStreetMap style. |
Castle buildings undoubtedly need further tagging development. For example by documenting building=castle mentioned above with already 1400 uses. |
That's exactely the point. Every database object must make sense when queried individually. I.e. when someone asks for castle areas (to draw a map, calculate average area, or whatever), the object returned must represent that part of observable reality - the castle area doesn't have a hole. Same by the way with schools, where the school yard gets excluded from the school grounds. As for putting the tag on the outer ring, I don't like it either, but it saves a duplicate polygon and it is valid according to joto. This only occurs for single, stand-alone buildings without surroundings. |
This is still only one possible interpretation / definition. I also think that, while I applaud consistent tagging and concise Wiki definitions of real world objects in OSM, with the current state of OSM however, your ambitions as voiced in the bold quote are sky high. There is still a massive amount of inconsistencies all over the OSM database and Wiki. Based on the experience of developing my own renderer, I also know that the number of possible - and often totally unexpected - exceptions to certain Wike "rules" is just enormous. There is always an edge case defying a certain rule. Again: I am not against your tagging suggestion. Go ahead and tag it like that if you want to! It is OSM after all. I am just saying there are a lot more nuances to all of this than many people think. |
Regarding this issue, here is an example of a castle which is currently not rendered at all http://www.openstreetmap.org/way/211046305. The "castle" is a kind of open area with ruins; currently only the individual buildings inside the castle area are rendered, which is a bit confusing. |
Would you like to try preparing the code or move this idea forward in some other way? I don't think it's gonna be done if nobody will step up to do it. It sounds like a strong motivation for me that you can't accept current state of things. We should decide which type needs rendering (we're finally ready to select subtypes) and in which way (icons, minimum zoom level etc). Some icons that might be helpful are here: |
Just to chuck in another rendering option, here's what I went for for various "historic" areas in another map style: https://map.atownsend.org.uk/maps/map/map.html#zoom=18&lat=53.007133&lon=-0.624572 (that's https://www.openstreetmap.org/way/371069082 in OSM BTW) |
I would suggest to have the name and an icon at the castle node or area. I think no extra colouring or other fancy stuff is needed. I'm not a developer, so I cannot contribute here, sorry. I mapped a few castles at OSM, I "repaired" some of them: I turned single castles mapped as multiple ones (tagging for the renderer because name of buildings are shown?) into single ones by using a surrounding extra way - as described in the wiki. With the result: They are lost on openstreetmap.org now. What would you suggest here, how could I contribute to get castles rendered? And yes, to my opinion, the situation is unacceptable! I'm not saying that in an angry way, it is just like I think the current situation is. |
Sure, it's OK that you say what you need. It would just help if you be more precise - which castle type with which icon from which zoom level? |
2018-01-30 7:55 GMT+01:00 kocio-pl <notifications@github.com>:
Sure, it's OK that you say what you need. It would just help if you be
more precise - which castle type with which icon from which zoom level?
I would render the name label for every castle_type, z17+
The more prominent ones from z15 with icon and z16 with additional label.
Loooking at taginfo, the values for the second category can be these:
stately
defensive
fortress
palace
kremlin
defensive;stately
castrum
shiro
citadel
fortification
chateau
water_castle
|
We still need more specific decisions to prepare the code.
How would we select these (which tag or set of tags for example)?
Tags are not the problem, the icons are. So - which icons would we use and which types would they depict (of course there might be more types depicted by the same icon, just like we do with some shops for example)? |
I saw that currently many things are rendered starting with z16, so I would also use z16 for the castles. To my opinion there is no need to have dependencies on the "importance" of the castles. I was looking at the icons also, but I have the impression that they all look quite "European". I was trying to draw a new one using "inkscape", I'll continue trying during the weekend. |
I tried some icons, they don't look so different to the ones as your examples provided in https://github.com/gmgeo/osmic (mine are a little more ugly :-) ) I would suggest row 10, colum 10, the one with the "mixed roofs" on the two towers. The mixed roofs "soften" the icon a little to make it more compatible to the different versions of castles. I checked the web for fortified buildings, they don't look so dramatically different on the globe. |
So you suggest to use just one icon for all the types? I guess we can use more of them to show their slightly different meaning. |
Initial proposition: https://github.com/gmgeo/osmic/blob/master/tourism/castle-defensive-14.svg
https://github.com/gmgeo/osmic/blob/master/tourism/castle-fortress-14.svg
https://github.com/gmgeo/osmic/blob/master/tourism/castle-manor-14.svg
https://github.com/gmgeo/osmic/blob/master/tourism/castle-stately-14.svg
https://github.com/gmgeo/osmic/blob/master/tourism/castle-palace-14.svg
Notes:
What do you think? |
|
I would suggest to use just one icon. After the renderer will display castles, things can be refined in later versions. Hopefully some castle experts will contribute after the first most simple version will appear on openstreetmap.org I updated the german wiki: to explain the tagging in greater detail. I also went through this thread again: |
Manor icon is interesting, but just doesn't work for me at 14 px and a palace icon doesn't work for me at all. |
@kocio-pl I understand. I'm putting Gist, in case of someone want to tune up this icons. |
Did you also have a look here?: |
Better on 14 px than scaled up version, but still not too clear for me. BTW: do you try to align the lines with the pixel matrix? |
There is some discussion on talk-de that the less-used historic=fort was merged before the more-used historic=castle was ready. There is fear that this might lead to retagging for the renderer. Consequently, this issue here should be accelerated, maybe with an interim solution. |
I see. I think this is quite close to being ready (I have some new ideas to test soon, but it's just about tuning visual differences in Osmic shapes), so I wouldn't worry about that. |
I like this version of manor shape and I'd like to test how it looks on the map. Palace/stately icon is good in general, but I get lost with readability details like windows and connected roofs - I'm struggling to recognize it as a building. |
+1 manor shape |
Gist links:
|
+1 manor |
Can different icons for different castle types be added? |
Castles are important landmarks, so I think we should render them. See also https://trac.openstreetmap.org/ticket/2247.
The text was updated successfully, but these errors were encountered: