Re: [OSM-talk] Rendering of Place Names in Mapnik
I hate to revive a long dead thread, but the issue still exists and Brandon's good idea seems to have been forgotten about. Does anybody have any suggestions for how this idea might be implemented within OSM? -Scott On Fri, Dec 19, 2008 at 5:40 PM, Brandon Aguirre bran...@cloudmade.comwrote: US census has a well-organized system of divisions that might suit the need to create hierarchy within urban agglomerations and poly-centric regions. 1. PMSA- Primary Metropolitan Statistical Area i.e. Greater SF Bay Area is a PMSA 2. SMSA- Secondary Metropolitan Statistical Area possibly SF-Penninsula, Oakland-EastBay and SJ-SouthBay are SMSAs 3. MSA- Metropolitan Statistical Area- the pop. figures and characteristics escape me... PMSA and SMSA account for the SFs and Miamis of the US by statistically recognizing the central core city as the generator/facilitator of activity (Greater So. Fla: 5.8 million, Miami 380,000) Having these two designations also helps in cases like Ft. Lauderdale, which is part of the Miami PMSA and is a 'core city' of it's own SMSA, yet it's population is only 160,000 while the Miami suburb of Hialeah is over 200,000. the automatic deference to FTL would be built in. Making use of that info however, is where my tech skills fail me As a side note- Denver: True City/County Indianapolis: Annexed entire Marion County Louisville: Government Municipal consolidation Jacksonville: Annexed entire St. John's County Broomfield, CO (a tiny suburb of Denver: True City/County Cheers, Brandon Aguirre Community Development CloudMade 503.998.1567 Skype: bragpdx bran...@cloudmade.com www.cloudmade.com On Dec 18, 2008, at 1:46 PM, Beej Jorgensen wrote: As a datapoint, Google renders SF and SJ the same way until you zoom out far enough, and then SJ disappears. FWIW, -Beej ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk -- Scott Atwood The hill isn't in the way, it is the way. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik: Wikipedia article length as a heuristic
I love the idea of using wikipedia article size as a proxy for 'importance'. My idea was to do a google fight between overlapping labels. (San Francisco 227M hits, San Jose 93M hits) Aled On Wed, Dec 24, 2008 at 9:47 AM, Edward Betts edwardbe...@gmail.com wrote: How about we use the length of the Wikipedia article about a city to decide which should get rendered on OpenStreetMap? We could add a tag to each city with the Wikipedia article URL. Here are some figures for the Bay Area: San Francisco: 119958 Oakland: 113560 San Jose: 84823 Mountain View: 26124 Redwood City: 23369 South San Francisco: 20539 Daly City: 17288 Just an idea. -- Edward. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 5:06 PM, Pieren pier...@gmail.com wrote: On Thu, Dec 18, 2008 at 5:12 PM, Elena of Valhalla elena.valha...@gmail.com wrote: ok, not that likely, but I wouldn't use a tag with a different meaning when we can just add another specific one; Is that really so different ? It is a renderer issue. Mapnik decides to not draw one of the names in case of collision. But another renderer could make it differently, drawing a name above the other, or using smaller font, etc... Of course, layer has not to be used in a general context but only in something like this: if (A.place is not equal to B.place) select A or B else if (A.admin_level is not equal to B.admin_level select A or B else //A.admin_level = B.admin_level select highest A or B layer The layer tag is for features which are physically vertically separated. If you have two cities, one of which is physically built above the other, then the layer tag would be appropriate to show which one is on top. Other than that, it's simply incorrect. Thanks, Andy ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Fri, Dec 19, 2008 at 1:00 PM, Andy Allan The layer tag is for features which are physically vertically separated. Agree with features vertically separated. Just sad that you decide to limit the usage to physical objects... Here we talk about tagging nodes separated by long distance. Anyway, we can call it name_level, manik_name_layer, cultural_level or wathever you like, we just need a solution. Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
US census has a well-organized system of divisions that might suit the need to create hierarchy within urban agglomerations and poly-centric regions. PMSA- Primary Metropolitan Statistical Area i.e. Greater SF Bay Area is a PMSA SMSA- Secondary Metropolitan Statistical Area possibly SF-Penninsula, Oakland-EastBay and SJ-SouthBay are SMSAs MSA- Metropolitan Statistical Area- the pop. figures and characteristics escape me... PMSA and SMSA account for the SFs and Miamis of the US by statistically recognizing the central core city as the generator/ facilitator of activity (Greater So. Fla: 5.8 million, Miami 380,000) Having these two designations also helps in cases like Ft. Lauderdale, which is part of the Miami PMSA and is a 'core city' of it's own SMSA, yet it's population is only 160,000 while the Miami suburb of Hialeah is over 200,000. the automatic deference to FTL would be built in. Making use of that info however, is where my tech skills fail me As a side note- Denver: True City/County Indianapolis: Annexed entire Marion County Louisville: Government Municipal consolidation Jacksonville: Annexed entire St. John's County Broomfield, CO (a tiny suburb of Denver: True City/County Cheers, Brandon Aguirre Community Development CloudMade 503.998.1567 Skype: bragpdx bran...@cloudmade.com www.cloudmade.com On Dec 18, 2008, at 1:46 PM, Beej Jorgensen wrote: As a datapoint, Google renders SF and SJ the same way until you zoom out far enough, and then SJ disappears. FWIW, -Beej ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
Brandon Aguirre wrote: US census has a well-organized system of divisions that might suit the need to create hierarchy within urban agglomerations and poly-centric regions. I like this--it's naming a region, so it doesn't seem so far-fetched that it would be in the map. As a renderer, I might render cities by size until zoomed out enough, then ignore the cities and just render the name of the region instead... -Beej ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 1:37 AM, Karl Newman siliconfi...@gmail.com wrote: Maybe we need a tag for cultural value :-P Or use the admin_level tag. Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 12:57 AM, Pieren pier...@gmail.com wrote: On Thu, Dec 18, 2008 at 1:37 AM, Karl Newman siliconfi...@gmail.com wrote: Maybe we need a tag for cultural value :-P Or use the admin_level tag. Pieren That wouldn't work in this case, because as the OP mentioned, San Jose and San Francisco have equal admin_level rankings (county seat) and San Jose is larger in both area and population. As an aside, San Francisco is unique in the USA (as far as I know) in that the city and county have the same extents. Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
Karl Newman wrote: That wouldn't work in this case, because as the OP mentioned, San Jose and San Francisco have equal admin_level rankings (county seat) and San Jose is larger in both area and population. As an aside, San Francisco is unique in the USA (as far as I know) in that the city and county have the same extents. Philadelphia is like this, too. It is the county seat of Philadelphia County http://en.wikipedia.org/wiki/Philadelphia_County,_Pennsylvania (with which it is coterminous) http://en.wikipedia.org/wiki/Philadelphia ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 6:37 AM, Adam Killian vi...@bonius.com wrote: Karl Newman wrote: That wouldn't work in this case, because as the OP mentioned, San Jose and San Francisco have equal admin_level rankings (county seat) and San Jose is larger in both area and population. As an aside, San Francisco is unique in the USA (as far as I know) in that the city and county have the same extents. Philadelphia is like this, too. It is the county seat of Philadelphia County http://en.wikipedia.org/wiki/Philadelphia_County,_Pennsylvania (with which it is coterminous) http://en.wikipedia.org/wiki/Philadelphia Hey, I learned something today. Guess I can stay home now. ;-) Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
El Jueves, 18 de Diciembre de 2008, Karl Newman escribió: Maybe we need a tag for cultural value :-P Or use the admin_level tag. That wouldn't work in this case, because as the OP mentioned, San Jose and San Francisco have equal admin_level rankings (county seat) and San Jose is larger in both area and population. So, cultural_level tag? -- -- Iván Sánchez Ortega i...@sanchezortega.es You will be run over by a bus. signature.asc Description: This is a digitally signed message part. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 7:29 AM, Iván Sánchez Ortega i...@sanchezortega.eswrote: El Jueves, 18 de Diciembre de 2008, Karl Newman escribió: Maybe we need a tag for cultural value :-P Or use the admin_level tag. That wouldn't work in this case, because as the OP mentioned, San Jose and San Francisco have equal admin_level rankings (county seat) and San Jose is larger in both area and population. So, cultural_level tag? Ah, yes, I can see it now. The values could range from Barkada, Arkansas to Paris, France (Apologies to residents of Barkada) Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 4:29 PM, Iván Sánchez Ortega i...@sanchezortega.es wrote: El Jueves, 18 de Diciembre de 2008, Karl Newman escribió: Maybe we need a tag for cultural value :-P Or use the admin_level tag. So, cultural_level tag? Nothing more subjective ? ;-) Say the truth : we need a tag for rendering in case of name collision when the category (city) and admin_level (county seat) are the same (forget population which is even worst in this example). Just an idea : why not reusing the tag layer applying for same place levels ? San Francisco node: layer=1 Daly City node: (layer can be ommited) Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 4:59 PM, Pieren pier...@gmail.com wrote: Nothing more subjective ? ;-) Say the truth : we need a tag for rendering in case of name collision when the category (city) and admin_level (county seat) are the same (forget population which is even worst in this example). Just an idea : why not reusing the tag layer applying for same place levels ? San Francisco node: layer=1 Daly City node: (layer can be ommited) what happens if the place point happens to be in the area of a building that hosts the city hall, is a brige, has two ways that pass through it, crosses a river and below it there are a couple subway lines? ok, not that likely, but I wouldn't use a tag with a different meaning when we can just add another specific one; either based on some hard data (e.g. economic_activity), or clearly marked as a render hint (render:priority, maybe?) -- Elena of Valhalla homepage: http://www.trueelena.org email: elena.valha...@gmail.com ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 1:37 AM, Karl Newman siliconfi...@gmail.com wrote: That's the sort of thing automated renderers have difficulty sorting out. Maybe we need a tag for cultural value :-P (I would hazard a guess that San Jose has a larger economic impact, though.) I have suggested that the number of values for the hamlet/village/town/city hiearchy is incerased. Adding major_* and minor_* for village, town and city could be one way to solve some of the problems. - Gustav ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 7:59 AM, Pieren pier...@gmail.com wrote: On Thu, Dec 18, 2008 at 4:29 PM, Iván Sánchez Ortega i...@sanchezortega.es wrote: El Jueves, 18 de Diciembre de 2008, Karl Newman escribió: Maybe we need a tag for cultural value :-P Or use the admin_level tag. So, cultural_level tag? Nothing more subjective ? ;-) Say the truth : we need a tag for rendering in case of name collision when the category (city) and admin_level (county seat) are the same (forget population which is even worst in this example). Just an idea : why not reusing the tag layer applying for same place levels ? San Francisco node: layer=1 Daly City node: (layer can be ommited) Pieren The discussion was between San Francisco and San Jose. The Daly City strangeness could be fixed by checking population. (Daly City ~100k, San Francisco ~800k). If we're going to tag for the renderer, why not just do it explicitly and call it render_priority (instead of abusing another tag with a different purpose)? Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 8:12 AM, Gustav Foseid gust...@gmail.com wrote: On Thu, Dec 18, 2008 at 1:37 AM, Karl Newman siliconfi...@gmail.comwrote: That's the sort of thing automated renderers have difficulty sorting out. Maybe we need a tag for cultural value :-P (I would hazard a guess that San Jose has a larger economic impact, though.) I have suggested that the number of values for the hamlet/village/town/city hiearchy is incerased. Adding major_* and minor_* for village, town and city could be one way to solve some of the problems. - Gustav You're still missing the point about San Jose--it's larger in both area and population (and probably in economic activity as well), and is located within an hour's drive of San Francisco, but San Francisco is better known around the world and should arguably take priority in rendering. Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 5:24 PM, Karl Newman siliconfi...@gmail.com wrote: You're still missing the point about San Jose--it's larger in both area and population (and probably in economic activity as well), and is located within an hour's drive of San Francisco, but San Francisco is better known around the world and should arguably take priority in rendering. My idea was that major_ could be used for a city of greater importance of some kind. Someone also suggested a value metropolis for the large metropolitan areas. I do not agree that using population or area is a good way to solve this problem. Finding population data for all named places is not easy (this problem extends beyond the largest cities) and population is not necessarily a good way to find the most important place name in an area. - Gustav ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 6:39 AM, Karl Newman siliconfi...@gmail.com wrote: On Thu, Dec 18, 2008 at 6:37 AM, Adam Killian vi...@bonius.com wrote: Karl Newman wrote: That wouldn't work in this case, because as the OP mentioned, San Jose and San Francisco have equal admin_level rankings (county seat) and San Jose is larger in both area and population. As an aside, San Francisco is unique in the USA (as far as I know) in that the city and county have the same extents. Philadelphia is like this, too. It is the county seat of Philadelphia County http://en.wikipedia.org/wiki/Philadelphia_County,_Pennsylvania (with which it is coterminous) http://en.wikipedia.org/wiki/Philadelphia Hey, I learned something today. Guess I can stay home now. ;-) Also, when cities in VA become large enough, they become their own county and seat. Cheers, Adam ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 8:43 AM, Gustav Foseid gust...@gmail.com wrote: On Thu, Dec 18, 2008 at 5:24 PM, Karl Newman siliconfi...@gmail.comwrote: You're still missing the point about San Jose--it's larger in both area and population (and probably in economic activity as well), and is located within an hour's drive of San Francisco, but San Francisco is better known around the world and should arguably take priority in rendering. My idea was that major_ could be used for a city of greater importance of some kind. Someone also suggested a value metropolis for the large metropolitan areas. I do not agree that using population or area is a good way to solve this problem. Finding population data for all named places is not easy (this problem extends beyond the largest cities) and population is not necessarily a good way to find the most important place name in an area. - Gustav Sure it is. If a lot of people want to live in a place, in general that should make it more notable. Besides, I was only suggesting using population as a tiebreaker for equal place key values. It's not the final answer, but it's objective and it goes a long way toward fixing the problem. I don't like your _major and _minor suffixes because they imply different population, which is not how you described it. Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 5:12 PM, Elena of Valhalla elena.valha...@gmail.com wrote: ok, not that likely, but I wouldn't use a tag with a different meaning when we can just add another specific one; Is that really so different ? It is a renderer issue. Mapnik decides to not draw one of the names in case of collision. But another renderer could make it differently, drawing a name above the other, or using smaller font, etc... Of course, layer has not to be used in a general context but only in something like this: if (A.place is not equal to B.place) select A or B else if (A.admin_level is not equal to B.admin_level select A or B else //A.admin_level = B.admin_level select highest A or B layer Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 5:58 PM, Karl Newman siliconfi...@gmail.com wrote: Sure it is. If a lot of people want to live in a place, in general that should make it more notable. Besides, I was only suggesting using population as a tiebreaker for equal place key values. It's not the final answer, but it's objective and it goes a long way toward fixing the problem. I don't like your _major and _minor suffixes because they imply different population, which is not how you described it. How do you suggest we find population for places? I can tell that a town is a regional center, without having to know it's population. Maybe major/minor is not the best names, however. How many values should we have for populated places? We have 4 now (hamlet/village/town/city). Should we add more? Reduce to fewer? Maybe just one? - Gustav ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 9:05 AM, Gustav Foseid gust...@gmail.com wrote: On Thu, Dec 18, 2008 at 5:58 PM, Karl Newman siliconfi...@gmail.comwrote: Sure it is. If a lot of people want to live in a place, in general that should make it more notable. Besides, I was only suggesting using population as a tiebreaker for equal place key values. It's not the final answer, but it's objective and it goes a long way toward fixing the problem. I don't like your _major and _minor suffixes because they imply different population, which is not how you described it. How do you suggest we find population for places? I can tell that a town is a regional center, without having to know it's population. Maybe major/minor is not the best names, however. How many values should we have for populated places? We have 4 now (hamlet/village/town/city). Should we add more? Reduce to fewer? Maybe just one? - Gustav Well, in the US, the population for cities is posted on the city limit signs. Or widely available in the internet, etc. Understand that this wouldn't be necessary for most places. If the population tag is missing, then we just get the behavior we have now, which isn't terrible but could be improved. I would argue for more granularity in place values. It's dominated on the low end of population (everything over 100k population is a city). And to me, there's plenty of room to subdivide town, too--there's a big difference in a place with 10,000 people vs. 99,999 people. Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
Gustav wrote: How many values should we have for populated places? We have 4 now (hamlet/village/town/city). Should we add more? Reduce to fewer? Maybe just one? Perhaps something could be done similar to boundary with so many admin_levels and some sort of default mapping from the existing 4 places to their new numeric equivalent (a bit like footway and some combination of tags including highway=path are equivalent as far as I can tell). This would allow areas around the world to use intermediate levels should they wish to, if their societal structure makes such use appropriate. I'd probably add suburb somewhere between town and village, and allow perhaps 2 spare levels between each of those 5 categories, and perhaps a couple either side as well, though can't imagine what gets smaller than hamlet - isolated house perhaps? Ed ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Thu, Dec 18, 2008 at 6:34 PM, Ed Loach e...@loach.me.uk wrote: Perhaps something could be done similar to boundary with so many admin_levels and some sort of default mapping from the existing 4 places to their new numeric equivalent (a bit like footway and some combination of tags including highway=path are equivalent as far as I can tell). This would allow areas around the world to use intermediate levels should they wish to, if their societal structure makes such use appropriate. That could be an option, but it is not backwards compatible. It would, however, make it easior to adapt to various cultures. The place name structure must work in a number of different situations, just to mention a few: - A metropolis like Tokyo or Los Angeles, often constisting of what is considered a number of smaller cities. - An island like Crete, where the mountains are literally scattered with little towns and hamlets quite close together (example from Douglas Furlong in another thread). - Norwegian rural areas, where villages are just areas where the houses are somewhat closer than outside of villages. I'd probably add suburb somewhere between town and village, and allow perhaps 2 spare levels between each of those 5 categories, and perhaps a couple either side as well, though can't imagine what gets smaller than hamlet - isolated house perhaps? Regions within a town or city is of course another problem, with only suburb available today. - Gustav ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
Karl Newman wrote: As an aside, San Francisco is unique in the USA (as far as I know) in that the city and county have the same extents. Oh, so THAT'S why San Francisco's unique! I've always wondered. ;) -Beej, proud Bay Area citizen ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
As a datapoint, Google renders SF and SJ the same way until you zoom out far enough, and then SJ disappears. FWIW, -Beej Strange. In the catholic hierarchy, San Jose (Jesus' father) is clearly above San Francisco (merely an italian saint) Lucas ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Wed, Dec 17, 2008 at 4:20 PM, Karl Newman siliconfi...@gmail.com wrote: I looked a bit at the osm.xml file for Mapnik. It currently orders them by the place tag hierarchy (city, town, suburb, village, hamlet/locality), but there doesn't seem to be any sorting within equal hierarchies (which is probably why you see Daly City instead of San Francisco at certain zooms). Ideally it could do this by population, with a bonus for capitals. The documentation for Mapnik indicates that it supports value comparisons, but it looks like the population would have to be stored in a column. Anyway, that seems like it would be the way to go. I don't think simple population ranking is necessarily the best option. There are other, more subjective factors that are important as well. For instance, San Francisco is smaller in terms of both population and land area than San Jose, and both are the county seat of their respective counties, yet due its economic, cultural, and historical importance, San Francisco should trump San Jose in case of a rendering collision. -Scott -- Scott Atwood Cycle tracks will abound in Utopia. ~H.G. Wells ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Rendering of Place Names in Mapnik
On Wed, Dec 17, 2008 at 4:33 PM, Scott Atwood scott.roy.atw...@gmail.comwrote: On Wed, Dec 17, 2008 at 4:20 PM, Karl Newman siliconfi...@gmail.comwrote: I looked a bit at the osm.xml file for Mapnik. It currently orders them by the place tag hierarchy (city, town, suburb, village, hamlet/locality), but there doesn't seem to be any sorting within equal hierarchies (which is probably why you see Daly City instead of San Francisco at certain zooms). Ideally it could do this by population, with a bonus for capitals. The documentation for Mapnik indicates that it supports value comparisons, but it looks like the population would have to be stored in a column. Anyway, that seems like it would be the way to go. I don't think simple population ranking is necessarily the best option. There are other, more subjective factors that are important as well. For instance, San Francisco is smaller in terms of both population and land area than San Jose, and both are the county seat of their respective counties, yet due its economic, cultural, and historical importance, San Francisco should trump San Jose in case of a rendering collision. -Scott That's the sort of thing automated renderers have difficulty sorting out. Maybe we need a tag for cultural value :-P (I would hazard a guess that San Jose has a larger economic impact, though.) Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk