[talk-ph] bulk import of wikipedia POIs to OSM
Over at the main OSM list, there is a discussion on importing POIs from wikipedia articles: http://lists.openstreetmap.org/pipermail/talk/2009-May/036582.html The main discussion revolves around whether we should or shouldn't import given that some of the POIs location (lon/lat) maybe derived from Google Map. I'm not really familiar with wikipedia's POI data particularly for the Philippines. Do we need to add them here? Or there will be major conflicts with existing POIs we have already added. Any ideas? -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] bulk import of wikipedia POIs to OSM
I kinda sit on the fence on both the legal issues and the desirability of bulk importing specifically Philippines data. For kicking off this discussion, I'd therefore suggest that we do not do a bulk import but that we do individually use the Wikipedia map location tool. I have found this very useful for broadly locating something I am looking for even when the exact location is not accurate (often). If list members have not come across this, there are often lat/lon coords in the top right of a Wikipedia article. Click on that and you are given a list of online maps, click on OpenStreetMap map and the on the usual OpenStreetMap Edit link. I then use the landsat imagery to precisely locate the feature I am looking for and manually create a tag or way ... or adding more information to a tag that is already there. I generally use it for getting names for rivers, mountains and towns. Mike At 09:08 AM 6/05/2009, maning sambale wrote: Over at the main OSM list, there is a discussion on importing POIs from wikipedia articles: http://lists.openstreetmap.org/pipermail/talk/2009-May/036582.html The main discussion revolves around whether we should or shouldn't import given that some of the POIs location (lon/lat) maybe derived from Google Map. I'm not really familiar with wikipedia's POI data particularly for the Philippines. Do we need to add them here? Or there will be major conflicts with existing POIs we have already added. Any ideas? -- cheers, maning ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] Is Makati road-complete?
Makati is close to complete and is extremely usable already. But I wouldn't say it's 100% complete especially if we want to use it to find addresses. For instance I recently tagged the street of northern Ecology Village as missing and the residential streets in Magallanes very near the corner of EDSA and NLEX are still missing (Yahoo!'s imagery is outdated). On Wed, May 6, 2009 at 7:15 PM, maning sambale emmanuel.samb...@gmail.comwrote: On Wed, May 6, 2009 at 6:29 PM, Marloue Pidor mur...@mail2engineer.com wrote: I edited some roads in Guadalupe, named some streets. And there are lots of foot path in Brgy. Pitogo that is not in the map yet. I live in Pitogo when I'm in Manila and M. V. Laurilla St. leads to Samar St. I will be editing the area and I will be in Manila early next week. By the way, what is the From your report, it's obvious it's not yet road complete. criteria to consider an area/place to be road complete? When I say road complete, I define it as: Street names are labelled. This means that the map can be used to find an address. Roads for car traffic are present. One way streets and pedestrian streets are present. This means that the map can be used for car navigation Source: http://wiki.openstreetmap.org/wiki/Template:En:Map_status But we can define our own metric if the above definition is not sufficient. murlwe -Original Message- From: maning sambale [emmanuel.samb...@gmail.com] Sent: 5/6/2009 3:30:46 PM To: talk-ph@openstreetmap.org Subject: Re: [talk-ph] Is Makati road-complete? Follow-up to my inquiry here. Is Makati road-complete? In the main map: it has administrative boundaries up to barangay/village level; MRT stations; some building outlines; residential landuse areas; road restrictions; POIS; etc. Although there are few minor errors reported: http://openstreetbugs.appspot.com/?lon=121.02493475009584lat=14.556820060903302zoom=15 I don't go to Makati very often, therefore, I cannot mark a stamp of approval. Any Makati resident who can attest to the completeness of OSm coverage of the City? On Tue, Mar 31, 2009 at 9:37 AM, maning sambale emmanuel.samb...@gmail.com wrote: Of course not, everything changes. There will be new roads, old roads reclassified, Chairman BF changing u-turn slots and oneway streets. Of course, the question is what metrics should we use. We can roughly define complete as: Street names are labelled. Roads for car traffic are present. One way streets and pedestrian streets are present. This means that the map can be used for car navigation Source: http://wiki.openstreetmap.org/wiki/Template:En:Map_status On Tue, Mar 31, 2009 at 9:31 AM, Ed Garcia eppgar...@gmail.com wrote: Hi Maning, Just curious ... when you declare an area as road-complete, would this mean it would be sealed of from additions/edits of ways? On Tue, Mar 31, 2009 at 9:24 AM, maning sambale emmanuel.samb...@gmail.com wrote: I just had a look at Makati today: http://www.openstreetmap.org/?lat=14.554lon=121.0276zoom=14layers=B000FTF I have to say the data is comprehensive, roads, POIs, landuse, administrative boundaries. However, I also saw a couple of unnamed roads: http://www.openstreetmap.org/?lat=14.5543lon=121.0276zoom=14layers=000BFTF Once this, unnamed roads are updated, I think it's time to review the area and declare Makati as road-complete. Can somebody, have an objective look at Makati and evaluate? I suggest we follow this scheme: http://wiki.openstreetmap.org/wiki/Template:En:Map_status Then I propose we start voting whether we declare Makati as road complete -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph -- website administrator: - www.waypoints.ph - reeflife.eppgarcia.com PADI Divemaster #491048 -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
[talk-ph] What level of POIs do we add? (was Click the City)
It's tempting to mark out in OSM all of the Jollibee, McDonald's, and Starbucks branches in the Philippines but I'm wondering if this is wise. Do we mark out every retail store, every bank branch, every Western Union/FedEx/LBC outlet, and every company's office in OSM? My personal take is that we don't need to. One big problem is multi-storey malls like SM Megamall that have hundreds of tenants. It'll be quite unwieldy to maintain and edit overlapping POIs in such a small land area. I feel that the level of detail we should attain is down to the building level. If we know that there's a Jollibee in Megamall Building A, then it's enough for us to map out where Building A is, but not where the Jollibee there is exactly. In the specific case of malls, exceptions will be added for anchor tenants, like in SM Megamall: the SM Department Store, ACE Hardware, SM Supermarket, SM Cinemas, the Megatrade Halls, etc. No need to mark out each and every McDo, KFC, Jollibee, Bench, Penshoppe, Levi's, BPI, Chinabank, etc. in the mall. A comprehensive database of all retail addresses is better suited to something like OpenYellowPages, not OpenStreetMap. What do you guys think? On Tue, May 5, 2009 at 11:31 AM, maning sambale emmanuel.samb...@gmail.comwrote: What I mean is address information (we've discussed it already, and agreed it is important, unless I'm mistaken) not specific business establishment (although it is attached to the address). We also have numerous POIs added already. On Tue, May 5, 2009 at 11:12 AM, Ahmed Farooq ah...@enthropia.com wrote: I don’t think adding business data to OSM is a good idea - that data changes far too often and is far more complex in upkeep. -A -- http://vaes9.codedgraphic.com ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] What level of POIs do we add? (was Click the City)
On Wed, May 6, 2009 at 10:18 PM, maning sambale emmanuel.samb...@gmail.comwrote: Just my take, 1. If it's on the ground it should be mapped. This is a good guideline in general but of course a line has to be drawn somewhere (pun not intended). I don't think drawing the individual stripes of a pedestrian walkway is productive even if it *is* on the ground. A map is supposed to be a representation--not a facsimile of the real world. That's why we represent roads in OSM using center lines and not (at the moment) as areas. 2. If somebody bothered to map fishball vendors in Luneta, nobody's stopping him. Fishball vendors are probably not a good example for your point since they are too ephemeral to be mapped. But anyway, OSM is a community project and we work on consensus. While we do encourage people to map what they think are important, I don't think we should just let people map things like Location of Mark and Jenny's first kiss, right? 3 years ago (am I that old in OSM?), my only goal is to map major highways around the metro, now it's there. But we want more. We've mapped footways, cycleways, I even saw driveways somewhere. If somebody sees the importance of a certain feature then by all means let them map it. If I can collect data for breeding sites of all the endangered bird species (and I think it is important) I will possibly map it here. I agree there should be some limits and priorities (at the moment) but I am hopeful we will get to a point where we will map individual houses with addresses I disagree that the limits are there just for the moment. I think there should be a limit imposed at all times (the limit can move over time, but there is still a limit). For instance, Wikipedia has guidelines on what is NOT acceptable (http://en.wikipedia.org/wiki/WP:NOT) eventhough the presence of those guidelines run counter to Wikipedia's lofty aim to be the sum of all human knowledge. I think individual features for detached residential houses are OK, but I really don't think we need to place a point for every unit in high-rise condominiums. On Wed, May 6, 2009 at 8:20 PM, Eugene Alvin Villar sea...@gmail.com wrote: It's tempting to mark out in OSM all of the Jollibee, McDonald's, and Starbucks branches in the Philippines but I'm wondering if this is wise. Do we mark out every retail store, every bank branch, every Western Union/FedEx/LBC outlet, and every company's office in OSM? My personal take is that we don't need to. One big problem is multi-storey malls like SM Megamall that have hundreds of tenants. It'll be quite unwieldy to maintain and edit overlapping POIs in such a small land area. I feel that the level of detail we should attain is down to the building level. If we know that there's a Jollibee in Megamall Building A, then it's enough for us to map out where Building A is, but not where the Jollibee there is exactly. In the specific case of malls, exceptions will be added for anchor tenants, like in SM Megamall: the SM Department Store, ACE Hardware, SM Supermarket, SM Cinemas, the Megatrade Halls, etc. No need to mark out each and every McDo, KFC, Jollibee, Bench, Penshoppe, Levi's, BPI, Chinabank, etc. in the mall. A comprehensive database of all retail addresses is better suited to something like OpenYellowPages, not OpenStreetMap. What do you guys think? On Tue, May 5, 2009 at 11:31 AM, maning sambale emmanuel.samb...@gmail.com wrote: What I mean is address information (we've discussed it already, and agreed it is important, unless I'm mistaken) not specific business establishment (although it is attached to the address). We also have numerous POIs added already. On Tue, May 5, 2009 at 11:12 AM, Ahmed Farooq ah...@enthropia.com wrote: I don’t think adding business data to OSM is a good idea - that data changes far too often and is far more complex in upkeep. -A -- http://vaes9.codedgraphic.com -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph -- http://vaes9.codedgraphic.com ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] What level of POIs do we add? (was Click the City)
On Wed, May 6, 2009 at 11:28 PM, Rally de Leon rall...@gmail.com wrote: If you have access to (data of) all locations of Jollibee, McDonalds, 7-Eleven, Ministops, Mercury Drugs and all these big companies, by all means mark them. With the exceptions of those already inside big malls (which will just clutter the map); we only need the mall-building itself as a landmark (POI) together with position of entrance gates and parking lots. Well, my idea is that if a McDo branch exists as a standalone building then there's no problem drawing the building outline and adding name=McDonald's, amenity=fastfood to it. But, as you've said, I don't think that means that we should also place a point for the McDonald's inside Robinsons Galleria. This implies that one *cannot* query the OSM database and get the coordinate locations of *all* the McDonald's branches in Metro Manila (you'll only get partial info). To get the full information, one should look elsewhere (e.g., go to www.mcdonaldsph.com, or look at ClicktheCity.com, or even start an OpenYellowPages) then geocode the obtained addresses using street data from OSM to get the actual coordinates. This actually helps create mashups applications using OSM as a tool. -- http://vaes9.codedgraphic.com ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] POIs Part 2
On Wed, May 6, 2009 at 11:40 PM, Ahmed Farooq ah...@enthropia.com wrote: I have to add – have you seen the Boracay map? While it looks fantastic that it is all mapped out, naming each resort/business there has lead to a map that looks absolutely bloated and is unreadable at times. For example http://www.openstreetmap.org.ph/map/c/11.960549689657205/121.925368309021/17/- “Mango Ray Shenna’s Resort” – are those two places? One? Which building is covered by their name? Multiple buildings or just one? This is a rendering problem, not a problem of data collection. It's theoretically possible to add more metadata to those resorts (like N-star rating) so that a renderer can choose to, say, display only 4- and 5-star resorts so as not to clutter a general-purpose map. This can be tied to zoom levels so that at a low zoom level, only Boracay Regency, Discovery Shores, Friday's, Shangri-La Boracay, etc. will be displayed or labeled. At higher zoom levels, more detail and labels can then be displayed. A POI should be a notable location that a tourist may be interested in – a statue, a building, even a public transit location. But specific businesses (be it fast food joints or specific resorts) only create clutter and a mess – which will only get worse as businesses are created and others go out of business. Having too much data (especially as part of the primary data set) is contrary to the spirit of an open user-maintained map. Well, I generally think that some types or classes of data are good additions to the primary data set. For instance, Asiatype publishes a paper map of the Makati CBD and Ortigas business districts showing the locations of all the buildings. I don't think having those buildings in OSM is too much data. Likewise, I think the location of all resorts in Boracay is a class of data that would be a good addition to OSM. Detached (as in stand-alone) restaurants and other dining places are OK in my opinion; they provide vital landmark info as Rally has said. On the other hand, I oppose the systematic addition of *all* tenants inside shopping malls. Some are OK (like the anchor tenants that I've mentioned in a previous e-mail). -- http://vaes9.codedgraphic.com ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] POIs Part 2
Right - it isn't just about the rendering, but about the underlying data. A POI should be considered a universal point of interest - ala a landmark of building that would be included in a 'Guide to XXX City'. Loading up on all the businesses is changing from being a map to a yellowpages - which is fine if that is the idea of the project. Keeping this data upto date is far harder than actual pathways - and once businesses are no longer operating, the data will be bad - and that is a dangerous path to go down. -A From: Eugene Alvin Villar [mailto:sea...@gmail.com] Sent: Wednesday, May 06, 2009 11:58 AM To: Ahmed Farooq Cc: maning sambale; OSM Subject: Re: POIs Part 2 On Wed, May 6, 2009 at 11:40 PM, Ahmed Farooq ah...@enthropia.com wrote: I have to add - have you seen the Boracay map? While it looks fantastic that it is all mapped out, naming each resort/business there has lead to a map that looks absolutely bloated and is unreadable at times. For example http://www.openstreetmap.org.ph/map/c/11.960549689657205/121.925368309021/17 / - Mango Ray Shenna's Resort - are those two places? One? Which building is covered by their name? Multiple buildings or just one? This is a rendering problem, not a problem of data collection. It's theoretically possible to add more metadata to those resorts (like N-star rating) so that a renderer can choose to, say, display only 4- and 5-star resorts so as not to clutter a general-purpose map. This can be tied to zoom levels so that at a low zoom level, only Boracay Regency, Discovery Shores, Friday's, Shangri-La Boracay, etc. will be displayed or labeled. At higher zoom levels, more detail and labels can then be displayed. A POI should be a notable location that a tourist may be interested in - a statue, a building, even a public transit location. But specific businesses (be it fast food joints or specific resorts) only create clutter and a mess - which will only get worse as businesses are created and others go out of business. Having too much data (especially as part of the primary data set) is contrary to the spirit of an open user-maintained map. Well, I generally think that some types or classes of data are good additions to the primary data set. For instance, Asiatype publishes a paper map of the Makati CBD and Ortigas business districts showing the locations of all the buildings. I don't think having those buildings in OSM is too much data. Likewise, I think the location of all resorts in Boracay is a class of data that would be a good addition to OSM. Detached (as in stand-alone) restaurants and other dining places are OK in my opinion; they provide vital landmark info as Rally has said. On the other hand, I oppose the systematic addition of *all* tenants inside shopping malls. Some are OK (like the anchor tenants that I've mentioned in a previous e-mail). -- http://vaes9.codedgraphic.com ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] talk-ph Digest, Vol 10, Issue 8
think individual features for detached residential houses are OK, but I really don't think we need to place a point for every unit in high-rise condominiums. On Wed, May 6, 2009 at 8:20 PM, Eugene Alvin Villar sea...@gmail.com wrote: It's tempting to mark out in OSM all of the Jollibee, McDonald's, and Starbucks branches in the Philippines but I'm wondering if this is wise. Do we mark out every retail store, every bank branch, every Western Union/FedEx/LBC outlet, and every company's office in OSM? My personal take is that we don't need to. One big problem is multi-storey malls like SM Megamall that have hundreds of tenants. It'll be quite unwieldy to maintain and edit overlapping POIs in such a small land area. I feel that the level of detail we should attain is down to the building level. If we know that there's a Jollibee in Megamall Building A, then it's enough for us to map out where Building A is, but not where the Jollibee there is exactly. In the specific case of malls, exceptions will be added for anchor tenants, like in SM Megamall: the SM Department Store, ACE Hardware, SM Supermarket, SM Cinemas, the Megatrade Halls, etc. No need to mark out each and every McDo, KFC, Jollibee, Bench, Penshoppe, Levi's, BPI, Chinabank, etc. in the mall. A comprehensive database of all retail addresses is better suited to something like OpenYellowPages, not OpenStreetMap. What do you guys think? On Tue, May 5, 2009 at 11:31 AM, maning sambale emmanuel.samb...@gmail.com wrote: What I mean is address information (we've discussed it already, and agreed it is important, unless I'm mistaken) not specific business establishment (although it is attached to the address). We also have numerous POIs added already. On Tue, May 5, 2009 at 11:12 AM, Ahmed Farooq ah...@enthropia.com wrote: I don't think adding business data to OSM is a good idea - that data changes far too often and is far more complex in upkeep. -A -- http://vaes9.codedgraphic.com -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph -- http://vaes9.codedgraphic.com -- next part -- An HTML attachment was scrubbed... URL: http://lists.openstreetmap.org/pipermail/talk-ph/attachments/20090506/b445d2d9/attachment-0001.htm -- Message: 2 Date: Wed, 6 May 2009 23:42:51 +0800 From: Eugene Alvin Villar sea...@gmail.com Subject: Re: [talk-ph] What level of POIs do we add? (was Click the City) To: Rally de Leon rall...@gmail.com Cc: OSM talk-ph@openstreetmap.org Message-ID: a07a5a700905060842g4b8e4782q2a8287fc0ed29...@mail.gmail.com Content-Type: text/plain; charset=iso-8859-1 On Wed, May 6, 2009 at 11:28 PM, Rally de Leon rall...@gmail.com wrote: If you have access to (data of) all locations of Jollibee, McDonalds, 7-Eleven, Ministops, Mercury Drugs and all these big companies, by all means mark them. With the exceptions of those already inside big malls (which will just clutter the map); we only need the mall-building itself as a landmark (POI) together with position of entrance gates and parking lots. Well, my idea is that if a McDo branch exists as a standalone building then there's no problem drawing the building outline and adding name=McDonald's, amenity=fastfood to it. But, as you've said, I don't think that means that we should also place a point for the McDonald's inside Robinsons Galleria. This implies that one *cannot* query the OSM database and get the coordinate locations of *all* the McDonald's branches in Metro Manila (you'll only get partial info). To get the full information, one should look elsewhere (e.g., go to www.mcdonaldsph.com, or look at ClicktheCity.com, or even start an OpenYellowPages) then geocode the obtained addresses using street data from OSM to get the actual coordinates. This actually helps create mashups applications using OSM as a tool. -- http://vaes9.codedgraphic.com -- next part -- An HTML attachment was scrubbed... URL: http://lists.openstreetmap.org/pipermail/talk-ph/attachments/20090506/1466730a/attachment-0001.htm -- Message: 3 Date: Wed, 06 May 2009 17:04:36 +0100 From: Ronny Ager-Wick - Develo Ltd. r...@develo.ltd.uk Subject: Re: [talk-ph] What level of POIs do we add? (was Click the City) To: OSM talk-ph@openstreetmap.org Message-ID: 4a01b514.5020...@develo.ltd.uk Content-Type: text/plain; charset
Re: [talk-ph] talk-ph Digest, Vol 10, Issue 8
Noel, This is my opinion... It is feasible but not practical, the names of the establishments will be used as our points of interest (POI). My point is, those establishments is on the ground and should be mapped and identified. OSM will not become as you say it commercially-oriented instead, it will be very useful to have those POIs intact. For example, if you are looking for Mercury Drug because it is your preferred drugstore then if the scenario is all the names of these drugstores were removed, you will be guessing and hoping that the next drugstore is the one you preferred. Its really hard to use the map that way. In my case I downloaded the OSM tiles into my Palm (osm2palm) and I use it whenever I'm in Manila the establishment's names were very useful for me to identify my location in the map. murlwe -Original Message- From: noel mondragon [noel_pylo...@yahoo.com] Sent: 5/7/2009 8:23:11 AM To: talk-ph@openstreetmap.org Subject: Re: [talk-ph] talk-ph Digest, Vol 10, Issue 8 My suggestion is that we can POI like drugstores, fastfood but we will not put the name like McDonalds or Jollibee or Mercury Drugstore??or resort? Is it that feasible. It will make the OSM as commercially-oriented if we put commercial establishments with names Comments? thanks. -noel --- On Wed, 5/6/09, talk-ph-requ...@openstreetmap.org talk-ph-requ...@openstreetmap.org wrote: From: talk-ph-requ...@openstreetmap.org talk-ph-requ...@openstreetmap.org Subject: talk-ph Digest, Vol 10, Issue 8 To: talk-ph@openstreetmap.org Date: Wednesday, May 6, 2009, 8:55 AM Send talk-ph mailing list submissions to talk-ph@openstreetmap.org To subscribe or unsubscribe via the World Wide Web, visit http://lists.openstreetmap.org/listinfo/talk-ph or, via email, send a message with subject or body 'help' to talk-ph-requ...@openstreetmap.org You can reach the person managing the list at talk-ph-ow...@openstreetmap.org When replying, please edit your Subject line so it is more specific than Re: Contents of talk-ph digest... Today's Topics: 1. POIs Part 2 (Ahmed Farooq) 2. Re: What level of POIs do we add? (was Click the City) (Eugene Alvin Villar) 3. Re: What level of POIs do we add? (was Click the City) (Ronny Ager-Wick - Develo Ltd.) -- Message: 1 Date: Wed, 6 May 2009 11:40:05 -0400 From: Ahmed Farooq ah...@enthropia.com Subject: [talk-ph] POIs Part 2 To: 'Eugene Alvin Villar' sea...@gmail.com, 'maning sambale' emmanuel.samb...@gmail.com Cc: 'OSM' talk-ph@openstreetmap.org Message-ID: 01f601c9ce60$ee0643a0$ca12ca...@com Content-Type: text/plain; charset=us-ascii Hi All, I have to add - have you seen the Boracay map? While it looks fantastic that it is all mapped out, naming each resort/business there has lead to a map that looks absolutely bloated and is unreadable at times. For example http://www.openstreetmap.org.ph/map/c/11.960549689657205/121.92536830902 1/17 / - Mango Ray Shenna's Resort - are those two places? One? Which building is covered by their name? Multiple buildings or just one? A POI should be a notable location that a tourist may be interested in - a statue, a building, even a public transit location. But specific businesses (be it fast food joints or specific resorts) only create clutter and a mess - which will only get worse as businesses are created and others go out of business. Having too much data (especially as part of the primary data set) is contrary to the spirit of an open user-maintained map. -A From: talk-ph-boun...@openstreetmap.org [mailto:talk-ph-boun...@openstreetmap.org] On Behalf Of Eugene Alvin Villar Sent: Wednesday, May 06, 2009 11:28 AM To: maning sambale Cc: OSM Subject: Re: [talk-ph] What level of POIs do we add? (was Click the City) On Wed, May 6, 2009 at 10:18 PM, maning sambale emmanuel.samb...@gmail.com wrote: Just my take, 1. If it's on the ground it should be mapped. This is a good guideline in general but of course a line has to be drawn somewhere (pun not intended). I don't think drawing the individual stripes of a pedestrian walkway is productive even if it *is* on the ground. A map is supposed to be a representation--not a facsimile of the real world. That's why we represent roads in OSM using center lines and not (at the moment) as areas. 2. If somebody bothered to map fishball vendors in Luneta, nobody's stopping him. Fishball vendors are probably not a good example for your point since they are too ephemeral to be mapped. But anyway, OSM is a community project and we work on consensus. While we do encourage people to map what they think are important, I don't think we should just let people map things like Location of Mark and Jenny's first kiss, right? 3 years ago (am I that old in OSM?), my only goal is to map major
Re: [talk-ph] What level of POIs do we add? (was Click the City)
Guys, just merging the thoughts on qualifying POIs ... how about, in addition to a POI having a street level entrance, we consider if the POI has prominent building signage? Some POIs may be inside a building but have very visible signage outside or in front of the building that serve as very good landmarks. :) ed On Thu, May 7, 2009 at 9:55 AM, Marloue Pidor mur...@mail2engineer.comwrote: In this case http://openstreetmap.org/?lat=7.078087lon=125.614181zoom=18layers=B000FTF Only KFC, Gerry's Grill and Pizza Hut have street level/accessible entrance. So if we agree on that rule of thumb to mapped only if they have a separate entrance. Should I remove Jollibee and McDonald's? Or retain the data as long as it is not cluttered? And in another case http://openstreetmap.org/?lat=7.062681lon=125.593989zoom=18layers=B000FTF All of these food chains do not have a street level entrance. murlwe -Original Message- From: Ronny Ager-Wick - Develo Ltd. [...@develo.ltd.uk] Sent: 5/7/2009 12:05:36 AM To: talk-ph@openstreetmap.org Subject: Re: [talk-ph] What level of POIs do we add? (was Click the City) Eugene, I think what you suggest sound sensible. Could we use this as a rule of thumb: If it has a door out to street level, then it can be mapped as a POI. This way, a shop or fast foot outlet in a mall can be mapped only if they have a separate entrance. All other shops are *inside* the mall. This way it's also consistent with for example a Jollibee that is located on the ground floor of an office building, as it will have its own entrance. Ronny. Eugene Alvin Villar wrote: On Wed, May 6, 2009 at 11:28 PM, Rally de Leon rall...@gmail.com wrote: If you have access to (data of) all locations of Jollibee, McDonalds, 7-Eleven, Ministops, Mercury Drugs and all these big companies, by all means mark them. With the exceptions of those already inside big malls (which will just clutter the map); we only need the mall-building itself as a landmark (POI) together with position of entrance gates and parking lots. Well, my idea is that if a McDo branch exists as a standalone building then there's no problem drawing the building outline and adding name=McDonald's, amenity=fastfood to it. But, as you've said, I don't think that means that we should also place a point for the McDonald's inside Robinsons Galleria. This implies that one *cannot* query the OSM database and get the coordinate locations of *all* the McDonald's branches in Metro Manila (you'll only get partial info). To get the full information, one should look elsewhere (e.g., go to www.mcdonaldsph.com, or look at ClicktheCity.com, or even start an OpenYellowPages) then geocode the obtained addresses using street data from OSM to get the actual coordinates. This actually helps create mashups applications using OSM as a tool. -- http://vaes9.codedgraphic.com ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph ___ Get the Free email that has everyone talking at http://www.mail2world.com Unlimited Email Storage POP3 Calendar SMS Translator Much More! ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph -- website administrator: - www.waypoints.ph - reeflife.eppgarcia.com PADI Divemaster #491048 ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] What level of POIs do we add? (was Click the City)
My 2 Centavos on Commercial Establishments, It is nice to see them on the map but I find that the work to maintain these is going to get very difficult and a bulk load is just asking for problems. Using the Wikipedia data before I found many were out of location or just plain wrong, Thus the work to fix is quite a lot and is discouraging at times. If we just take Makati for an example the police station moved, I thought it had and checked the site. Now i live down the road from this and it was not updated. I have tried to move it and checking now at least it is no longer out side the city hall which is shown but the police and the fire station have both disappeared from where they are meant to be.. Now moving down the road from Makati City hall we have McDo but Jollibee which is has more customers is not labeled, nor is the fast food shop across the street. So do we want the head ache.. If you work for the establishment EG: McDo Philippines then go ahead and keep it up to date... If you don't then don't add them... Places of Cultural Value all should be added first.. (and they don't move..) Moving further out we have at Jupiter and Makati Ave a McDo listed but once again in the wrong place. Now does it mean that since I live here close I have to double check every week the data? What happens when I move again? Do I only add the ones I like? And then add bias to the maps? Better to exclude.. What happens if I dont like and add it in the wrong place on purpose? These are my 2 Centavo's ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] What level of POIs do we add? (was Click the City)
Just a thought, imagine osm tokyo japan contemplating on this issue. That would be one hell of a mapping problem. hehehe But I digress, please move on ... On 5/7/09, Marloue Pidor mur...@mail2engineer.com wrote: It seems to be a good idea but most of these food chains are located in front and in most cases on the ground floor/street level and most of them have a visible signage. Take for example Glorietta, most of the establishments have street level entrance Suggestion: maybe we can add these mall POIs in such a way that it is not confusing and overwhelming. When I say overwhelming, think of Mall of Asia if we add all establishments with street level entrance it would literally darken the perimeter of the building with all the POIs. Well this is just my thoughts and I could be wrong. :) murlwe -Original Message- From: Ed Garcia [eppgar...@gmail.com] Sent: 5/7/2009 10:21:44 AM To: mur...@mail2engineer.com Cc: r...@develo.ltd.uk;talk-ph@openstreetmap.org Subject: Re: [talk-ph] What level of POIs do we add? (was Click the City) Guys, just merging the thoughts on qualifying POIs ... how about, in addition to a POI having a street level entrance, we consider if the POI has prominent building signage? Some POIs may be inside a building but have very visible signage outside or in front of the building that serve as very good landmarks. :) ed On Thu, May 7, 2009 at 9:55 AM, Marloue Pidor mur...@mail2engineer.com wrote: In this case http://openstreetmap.org/?lat=7.078087lon=125.614181zoom=18layers=B0 00FTF Only KFC, Gerry's Grill and Pizza Hut have street level/accessible entrance. So if we agree on that rule of thumb to mapped only if they have a separate entrance. Should I remove Jollibee and McDonald's? Or retain the data as long as it is not cluttered? And in another case http://openstreetmap.org/?lat=7.062681lon=125.593989zoom=18layers=B0 00FTF All of these food chains do not have a street level entrance. murlwe -Original Message- From: Ronny Ager-Wick - Develo Ltd. [...@develo.ltd.uk] Sent: 5/7/2009 12:05:36 AM To: talk-ph@openstreetmap.org Subject: Re: [talk-ph] What level of POIs do we add? (was Click the City) Eugene, I think what you suggest sound sensible. Could we use this as a rule of thumb: If it has a door out to street level, then it can be mapped as a POI. This way, a shop or fast foot outlet in a mall can be mapped only if they have a separate entrance. All other shops are *inside* the mall. This way it's also consistent with for example a Jollibee that is located on the ground floor of an office building, as it will have its own entrance. Ronny. Eugene Alvin Villar wrote: On Wed, May 6, 2009 at 11:28 PM, Rally de Leon rall...@gmail.com wrote: If you have access to (data of) all locations of Jollibee, McDonalds, 7-Eleven, Ministops, Mercury Drugs and all these big companies, by all means mark them. With the exceptions of those already inside big malls (which will just clutter the map); we only need the mall-building itself as a landmark (POI) together with position of entrance gates and parking lots. Well, my idea is that if a McDo branch exists as a standalone building then there's no problem drawing the building outline and adding name=McDonald's, amenity=fastfood to it. But, as you've said, I don't think that means that we should also place a point for the McDonald's inside Robinsons Galleria. This implies that one *cannot* query the OSM database and get the coordinate locations of *all* the McDonald's branches in Metro Manila (you'll only get partial info). To get the full information, one should look elsewhere (e.g., go to www.mcdonaldsph.com, or look at ClicktheCity.com, or even start an OpenYellowPages) then geocode the obtained addresses using street data from OSM to get the actual coordinates. This actually helps create mashups applications using OSM as a tool. -- http://vaes9.codedgraphic.com ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph ___ Get the Free email that has everyone talking at http://www.mail2world.com Unlimited Email Storage - POP3 - Calendar - SMS - Translator - Much More! ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph -- website administrator: - www.waypoints.ph - reeflife.eppgarcia.com PADI Divemaster #491048 span id=m2wTlpfont face=Arial, Helvetica, sans-serif size=2 style=font-size:13.5px___BRGet the Free email that has everyone talking at a href=http://www.mail2world.com target=newhttp://www.mail2world.com/abr font color=#99Unlimited Email Storage #150; POP3 #150; Calendar #150; SMS #150; Translator #150; Much More!/font/font/span -- cheers,
[talk-ph] Fwd: [OSM-talk] Nominations for State of the Map Travel Scholarships
-- Forwarded message -- From: Mikel Maron mikel_ma...@yahoo.com Date: Thu, May 7, 2009 at 1:25 PM Subject: [OSM-talk] Nominations for State of the Map Travel Scholarships To: t...@openstreetmap.org The OpenStreetMap Foundation is excited to announce a program to cover full travel and accomodations costs for 15 mappers to attend State of the Map. We're seeking nominations from the community for potential mappers. Generally, we are seeking people from places where costs would prohibit attendance, developing countries, and places that are interesting geopolitically. The ideal candidates for funding are from countries with a small OSM community, perhaps just a few mappers in total. They have made a significant start at mapping their city, either through Yahoo imagery or with their own GPS, and are directly familiar with the process of OSM. They may have started communicating among themselves, and made plans and scoped out the process for their local district. But, the community is nowhere near critical mass, and they need the inspiration and support to take OSM to the next level. We need to act fast. State of the Map is just over two months away, tickets and visas need to be arranged. In order to allow enough time for all the arrangements, the nomination period will be short, one week only, ending next Wednesday, May 13. From the nominations received, we'll review the list and choose 15 mappers to approach with the offer. Depending on their availability to attend, we'll work our way through the list. We only recently secured funding for this program, so the process has to be quick. Please send your nominations to sotm.scholars...@gmail.com. For each nomination, include the mappers name, OSM user name, email address, location, and a paragraph or two on why they'd be great to have at SOTM. And also, please forward this message to other relevant local OSM lists. As for regions, here are a few regions that seem to fit the bill, but nominations are not limited to these places at all. * Eastern Europe: the Caucasus, Russia, Bulgaria, Romania. * Arab States: Lebanon, Jordan, Egypt. Egypt is particularly interesting, as the ban on GPS units there was recently lifted. * South Asia: India, Pakistan. While both countries have seen significant activities, relative to size and population they are in the very early stages. * Southeast Asia: Phillipines, Vietnam, Thailand. * South America: Colombia, Bolivia, Peru. Promising leads in government for the release of data for use in OSM. * East Africa: Kenya is a hotspot for mapping right now (Ushahidi, AgCommons, MapMaker..) Many thanks to the Open Society Institute for helping make this happen. -Mikel ___ talk mailing list t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] [OSM-talk] Nominations for State of the Map Travel Scholarships
We are mentioned here: * Southeast Asia: Phillipines, Vietnam, Thailand. Anybody willing to represent OSM-PH? I guess it's worth trying to nominate one from the group. On Thu, May 7, 2009 at 1:39 PM, maning sambale emmanuel.samb...@gmail.com wrote: -- Forwarded message -- From: Mikel Maron mikel_ma...@yahoo.com Date: Thu, May 7, 2009 at 1:25 PM Subject: [OSM-talk] Nominations for State of the Map Travel Scholarships To: t...@openstreetmap.org The OpenStreetMap Foundation is excited to announce a program to cover full travel and accomodations costs for 15 mappers to attend State of the Map. We're seeking nominations from the community for potential mappers. Generally, we are seeking people from places where costs would prohibit attendance, developing countries, and places that are interesting geopolitically. The ideal candidates for funding are from countries with a small OSM community, perhaps just a few mappers in total. They have made a significant start at mapping their city, either through Yahoo imagery or with their own GPS, and are directly familiar with the process of OSM. They may have started communicating among themselves, and made plans and scoped out the process for their local district. But, the community is nowhere near critical mass, and they need the inspiration and support to take OSM to the next level. We need to act fast. State of the Map is just over two months away, tickets and visas need to be arranged. In order to allow enough time for all the arrangements, the nomination period will be short, one week only, ending next Wednesday, May 13. From the nominations received, we'll review the list and choose 15 mappers to approach with the offer. Depending on their availability to attend, we'll work our way through the list. We only recently secured funding for this program, so the process has to be quick. Please send your nominations to sotm.scholars...@gmail.com. For each nomination, include the mappers name, OSM user name, email address, location, and a paragraph or two on why they'd be great to have at SOTM. And also, please forward this message to other relevant local OSM lists. As for regions, here are a few regions that seem to fit the bill, but nominations are not limited to these places at all. * Eastern Europe: the Caucasus, Russia, Bulgaria, Romania. * Arab States: Lebanon, Jordan, Egypt. Egypt is particularly interesting, as the ban on GPS units there was recently lifted. * South Asia: India, Pakistan. While both countries have seen significant activities, relative to size and population they are in the very early stages. * Southeast Asia: Phillipines, Vietnam, Thailand. * South America: Colombia, Bolivia, Peru. Promising leads in government for the release of data for use in OSM. * East Africa: Kenya is a hotspot for mapping right now (Ushahidi, AgCommons, MapMaker..) Many thanks to the Open Society Institute for helping make this happen. -Mikel ___ talk mailing list t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
Re: [talk-ph] [OSM-talk] Nominations for State of the Map Travel Scholarships
i would like to nominate Maning Sambale to represent the OSM-PH. :-) On Thu, May 7, 2009 at 1:47 PM, maning sambale emmanuel.samb...@gmail.comwrote: We are mentioned here: * Southeast Asia: Phillipines, Vietnam, Thailand. Anybody willing to represent OSM-PH? I guess it's worth trying to nominate one from the group. On Thu, May 7, 2009 at 1:39 PM, maning sambale emmanuel.samb...@gmail.com wrote: -- Forwarded message -- From: Mikel Maron mikel_ma...@yahoo.com Date: Thu, May 7, 2009 at 1:25 PM Subject: [OSM-talk] Nominations for State of the Map Travel Scholarships To: t...@openstreetmap.org The OpenStreetMap Foundation is excited to announce a program to cover full travel and accomodations costs for 15 mappers to attend State of the Map. We're seeking nominations from the community for potential mappers. Generally, we are seeking people from places where costs would prohibit attendance, developing countries, and places that are interesting geopolitically. The ideal candidates for funding are from countries with a small OSM community, perhaps just a few mappers in total. They have made a significant start at mapping their city, either through Yahoo imagery or with their own GPS, and are directly familiar with the process of OSM. They may have started communicating among themselves, and made plans and scoped out the process for their local district. But, the community is nowhere near critical mass, and they need the inspiration and support to take OSM to the next level. We need to act fast. State of the Map is just over two months away, tickets and visas need to be arranged. In order to allow enough time for all the arrangements, the nomination period will be short, one week only, ending next Wednesday, May 13. From the nominations received, we'll review the list and choose 15 mappers to approach with the offer. Depending on their availability to attend, we'll work our way through the list. We only recently secured funding for this program, so the process has to be quick. Please send your nominations to sotm.scholars...@gmail.com. For each nomination, include the mappers name, OSM user name, email address, location, and a paragraph or two on why they'd be great to have at SOTM. And also, please forward this message to other relevant local OSM lists. As for regions, here are a few regions that seem to fit the bill, but nominations are not limited to these places at all. * Eastern Europe: the Caucasus, Russia, Bulgaria, Romania. * Arab States: Lebanon, Jordan, Egypt. Egypt is particularly interesting, as the ban on GPS units there was recently lifted. * South Asia: India, Pakistan. While both countries have seen significant activities, relative to size and population they are in the very early stages. * Southeast Asia: Phillipines, Vietnam, Thailand. * South America: Colombia, Bolivia, Peru. Promising leads in government for the release of data for use in OSM. * East Africa: Kenya is a hotspot for mapping right now (Ushahidi, AgCommons, MapMaker..) Many thanks to the Open Society Institute for helping make this happen. -Mikel ___ talk mailing list t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- -- cheers, maning -- Freedom is still the most radical idea of all -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ -- ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph ___ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
[OSM-legal-talk] Legal review by ITO lawyer
Just to let you know that we have now received an update to the legal review we did of the ODbL license from our lawyer. We have forwarded this to the ODC for an initial response and are in discussion on a few points. Some of the issues have been resolved, some are issues for the OSMF not ODC, some are minor and can in our opinion be ignored and some remain. We should be in a position to publish the results tomorrow. Regards, Peter Miller ITO World Ltd ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
[OSM-legal-talk] ODbL RC and share-alike licensing of Produced Works
Hi, Mike Collinson wrote: The new text is available at http://www.opendatacommons.org/licenses/odbl/ and includes diff versions so that you can see clearly what changes are made. With 0.9, we identified the problem of produced works not being releasable under CC-BY-SA (or any other share-alike license, say GFDL or even GPL where included in software) because of the reverse engineering clause which collides with the no restrictions may be added clause in these share-alike licenses. I think that most of us were quite clear that this would be a total show stopper, and several suggestions were made to overcome the problem. Can someone explain how this has been resolved? I see that the 1.0 release candidate has a provision to name someone who decides which licenses are deemed compatible, but it seems to me that this only applies to databases, not produced works. As I understand it, the follwoing would have to happen to be able to release a produced work under CC-BY-SA with ODbL RC 1.0: * When applying the ODbL, OSMF is authorized to define catalogue of compatible licenses; * OSMF decrees that CC-BY-SA is compatible; * User makes a derived database by copying all of OSM, or a part of it, and says this copy is now CC-BY-SA, not ODbL * User is then able to create a produced work from his CC-BY-SA copy of the database and license that produced work CC-BY-SA Am I reading this correctly? I think this would overcome the problem, but it would effectively dual-license the whole of OSM under ODbL and CC-BY-SA. Has that been discussed and found to be a good idea? Or does OSMF not have the intention of declaring CC-BY-SA a compatible license, and if not, how will CC-BY-SA licensed produced works be made possible? Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
[OSM-legal-talk] OSMF lawyer call
Dear all Today the license working group had a 2 hour phone call with OSMFs council to run through a large number of open issues, use cases and so on which had been generated on the wiki and this list.. We found it super useful and we're planning another soon. Grant is collating the replies to all the issues we managed to cover (which was a substantial portion of what we felt was the most important) and will get back to this list shortly, and the wiki, with those answers. Best Steve ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] ODbL RC and share-alike licensing of Produced Works
On 6 May 2009, at 16:04, Frederik Ramm wrote: Hi, Mike Collinson wrote: The new text is available at http://www.opendatacommons.org/licenses/odbl/ and includes diff versions so that you can see clearly what changes are made. With 0.9, we identified the problem of produced works not being releasable under CC-BY-SA (or any other share-alike license, say GFDL or even GPL where included in software) because of the reverse engineering clause which collides with the no restrictions may be added clause in these share-alike licenses. I think that most of us were quite clear that this would be a total show stopper, and several suggestions were made to overcome the problem. Can someone explain how this has been resolved? I don't think it has. In our call with the OSMF lawyer today he wanted more time to think this one through. Grant can you collate Frederiks thoughts for our next run through with the OSMF lawyer? Best Steve ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
[OSM-legal-talk] Produced Work
Hi We've put together a practical definition for the OSMFs point of view on what a substantial extract is, or isn't http://wiki.openstreetmap.org/wiki/Open_Data_License/Substantial_Defined And we'd like help similarly with building a practical definition of Produced Work. Here's how the license RC1 defines it: Produced Work – a work (such as an image, audiovisual material, text, or sounds) resulting from using the whole or a Substantial part of the Contents (via a search or other query) from this Database, a Derivative Database, or this Database as part of a Collective Database. Thoughts on more practical definitions? Best Steve ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-talk] Wikipedia POI import?
this will create many duplicates. You will need to do some checking before a poi is added. so many mass imports are done cleanup is a lot of work. checking should be done against points, ways, polygons. in osm tags are somtimes on building polugon or on a point. If we have both the map is too cluttered with trahs better import less but have better quality. On 5 May 2009, at 11:00 , Russ Nelson wrote: Any reason not to go through Wikipedia and import everything with a coordinate as a POI, with a url=http://wikipedia.org/NAME link, and name=NAME where NAME is the name of the Wikipedia entry? If I do this under a special username, then there is no problem backing out the import if somebody has a better idea later. -- Russ Nelson - http://community.cloudmade.com/blog - http://wiki.openstreetmap.org/wiki/User:RussNelson r...@cloudmade.com - Twitter: Russ_OSM - http://openstreetmap.org/user/RussNelson ___ 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] Wikipedia POI import?
available to OSM editors. One way to do that is to have a second API which consists of a cached copy of everything that map renderers might use, all merged into one read-only OSM-compatible api. So when somebody asks to edit an area, the editor also shows them the read- only elements, so they know not to enter anything already available to map users. could be very useful for other data too. one idea is to make osm itself or certain object a readonly layer and provide a editor for newbies where they can not destroy things easily. advanced mode will allow full edits. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Tue, May 05, 2009 at 09:14:30PM +0200, andrzej zaborowski wrote: Some notes: There's already the wikipedia=NAME tag (wikipedia=LA:NAME for non-english wikipedias, where LA=en,de...) in use in some places, so I'd recommend using that. Shouldn't that be wikipedia:LA=NAME ? Jochen -- Jochen Topf joc...@remote.org http://www.remote.org/jochen/ +49-721-388298 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Tue, May 05, 2009 at 11:59:51PM +0200, Iván Sánchez Ortega wrote: El Martes, 5 de Mayo de 2009, Russ Nelson escribió: On May 5, 2009, at 4:37 PM, Richard Fairhurst wrote: We have the sweat-of-the-brow doctrine instead. Fine enough, and who sweated hardest to click in a particular point on a Google Map? Google? Or the Wikipedia editor [...] OMG, I'm practically DRIPPING with sweat off my brow just thinking about it. OK, but *you* tell that to the Google Lawyer Legion(tm)(beta). I don't think we have to worry about that. Google hasn't sued Wikipedia yet. And Wikipedia has been distributing all those points in bulk for years. Google itself has taken the Wikipedia points and put the into a Google Earth layer. So they can't argue that they didn't know. Now that Wikipedia is switching to CC-BY-SA I don't see any legal reason why we can't exchange data. Come on, people, don't be afraid all the time of something that theoretically maybe could eventually happen. We should not in anticipatory obedience cede the ground to everybody who claims any kind of right just because he thinks he should have it. If there is a reasonable claim, we have to act, of course, but not because of some theoretical possibility where the alleged copyright holder didn't even say anything. If you really want to be whiter than white I suggest you scroll around the existing OSM maps and start emailing all those people doing large scale imports and not having any documentation about where that data was coming from. Ask them for a full pedigree of this data back to the guy doing the surveying. Written, signed and in triplicate. This is no super-white project and it can't be. Its a community project and there is not much control about what people are doing. If and only if there is a problem, we have to handle it (and do so quickly), but not before. Wikipedia has been having those issues for years and has been handling them well. Sure there are letters from laywers etc. all the time and they look at those cases, document them, maybe delete data or lock access temporarily. That works. Judges see the effort they put in to protect copyright where there is some copyright to protect. I suggest we follow the same strategy instead of announcing that we are white where we aren't and coming up with insurmountable hurdles in other cases. Jochen -- Jochen Topf joc...@remote.org http://www.remote.org/jochen/ +49-721-388298 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Languages
On Tue, May 05, 2009 at 08:45:26PM +0100, Emilie Laffray wrote: I was actually thinking about anything that carries a name. In the example you have given me, you have partially answer the question that I was asking: name is expressed in the local language. If you want to add translation, you need to append :isolanguage to the named field. What might be interesting and worth a discussion: A tag to describe the default language of this object, e.g. language=en. This could als be several tags, e.g. name=België - Belgique - Belgien name:nl=België name:fr=Belgique name:da=Belgien (and some more) language=nl;fr;da- this would be new (I hope I got the languages right, sorry if not) Comments? I like the idea of the language element. I would like to add an extra precision in this case. I think the order of language should be by importance. Great to have a supporter :) I think the people who enter the name and the language-tag to the database should decide, because they know the political situation. As a general rule the importance of languages sounds reasonable to me. I even had a second idea. When you look at Google Maps you see that there's an English translation (for places which have an English translation) in a second row. This is something that I also applied to the OpenStreetBrowser[1]. I thought it might be interesting to apply a tag second_languages to an element (or to an administrative boundary), so that the application can decide which languages should be put in the second row. And if this languages is already part of the main-name it will be ignored. This could even be inherited from the next-higher administrative boundary. Examples: name=Sibiu name:de=Hermannstadt language=ro second_languages=en;de (in the administrative boundary) - FIRST: Sibiu SECOND: Hermannstadt name=België - Belgique - Belgien name:nl=België name:fr=Belgique name:da=Belgien name:en=Belgium (and some more) language=nl;fr;da- this would be new second_languages=en;da (in the administrative boundary) - FIRST: België - Belgique - Belgien SECOND: Belgium (da is already part of the main name, so it will not be put in second row) Just a few thoughts ... Things like different alphabet might be also interesting to look at even if it is likely that it can be subsumed under translation. Yes. I was thinking if it wouldn't be better to not use the english translation as a second row, but rather the int_name? or the name rewritten in latin letters? Could make a difference ... Ideas? [1] http://www.openstreetbrowser.org greetings, Stephan -- Seid unbequem, seid Sand, nicht Öl im Getriebe der Welt! - Günther Eich ,-. | Stephan Plepelits, | | Technische Universität Wien -Studium Informatik Raumplanung | | openstreetbrowser.org couchsurfing.com tubasis.at bl.mud.at | | sk...@xover.htu.tuwien.ac.at - My Blog: http://plepe.at | `-' ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
Hi, Richard Fairhurst wrote: Don't forget we have _expressly_ asked Google, in the form of Ed Parsons at SOTM, and he has _expressly_ said, sorry, no, we don't have those rights to give away. Of course Russ's argument is that you do not have to be given those rights, by Ed Parsons or his upstream providers, and so the fact that Google doesn't have those rights assigned by contract (or is unwilling to assign them to you by contract) is irrelevant. Wikipedia also recommends you do a web search for the city name together with latitude and longitude so, hey, why stop at Google? You can infringe on lots of other people's content, too! Which brings me to an interesting question. We currently have a long-running theft investigation in Germany where the suspects are twin brothers. The fact that it is impossible to accuse one or the other seems to be a major complication for lawyers (it seems you cannot have a legal case against one of you two). Now I wonder what happens if: * If I google for the coordinates of something * find the same coordinates on 5 web pages * use them * it later turns out all these 5 pages have copied from google and the coordinates are an easter egg Who, then, has a legal case against me? I haven't lifted anything off Google, so it cannot be them; and the other 5 will have a hard time to prove I took something from them? And is something still illegal if it is impossible to bring a case against it? But evidently I'm being an armchair lawyer: Welcome to the club. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Tue, May 05, 2009 at 11:03:56PM -0700, Richard Fairhurst wrote: Actually, I think an OSMer said it best on Twitter. Openstreetmap is about gathering map data and sharing it. Some people seem desperate to import data from anywhere. GATHER IT YOURSELF. Well, actually its not. It might have been once, but now there is more data imported from somewhere (mostly TIGER) than there is collected data. I see the OSM projects goal as providing free geodata, doesn't matter where that came from. Lots of interesting data is not collectable. You dont' see border lines painted on the ground. While I love the aspect of collecting data myself and it is good for public relations to talk about moms pushing a pram and collecting geodata, thats only part of the project. Going out and collecting data was absolutely necessary to get this project jumpstarted. But now that OSM is accepted people are standing in line wanting to get their data imported. For better or worse, the project has changed. Jochen -- Jochen Topf joc...@remote.org http://www.remote.org/jochen/ +49-721-388298 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
Jochen Topf wrote: I don't think we have to worry about that. Google hasn't sued Wikipedia yet. And Wikipedia has been distributing all those points in bulk for years. It isn't about Google, it's about their data providers. Wikipedia is not a competitor to TeleAtlas. OpenStreetMap is. cheers Richard -- View this message in context: http://www.nabble.com/Wikipedia-POI-import--tp23392791p23401272.html Sent from the OpenStreetMap - General mailing list archive at Nabble.com. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On 05/05/2009 20:36, Russ Nelson wrote: On May 5, 2009, at 3:11 PM, Richard Fairhurst wrote: Adam Schreiber wrote: We don't know where the wikipedia users sourced their cooridinates from. Oh yes we do: Google Maps. http://en.wikipedia.org/wiki/Wikipedia:Obtaining_geographic_coordinates#Google_tools There is absolutely no way that Wikipedia-derived co-ordinates are suitable for mass import into OSM. In fact, we don't know this. And since Google didn't create those lat/ lon pairs, the Wikipedia editor did, Google had no participation in the act of creation, and thus no copyright claim. You guys have some really weird ideas about copyright. You think it's silly, we think it is silly, but it's not us, it's people like Ordnance Survey who claim this is a derived work. But when you say what effort did the provider put in in making a particular lat/lon, if you're getting that from their maps, you aren't doing it in a vacuum, you are choosing the point based on its relationship to what else you an see on the map (otherwise you wouldn't be using the map in the first place). And they've put lots of investment into preparing the map (in OS case, _our_ money as taxpayers!) so they are understandably protective of it. Copyright isn't just about making verbatim copies. If you scan a photo, clearly you are infringing the photographer's copyright. But if you recreate a set, hire a similar model, and take your own picture then you are _still_ infringing copyright even though you took the picture. There's a grey area when similar is really similar enough and there's a grey area with maps too as no one has really tested in court the map companies assertion that reading a coordinate off a map really is a derived work. But the fact they make the assertion is largely why we exist in the first place. Importing potentially tainted data undermines the whole reason for our existence. Why bother importing - why not just use Google maps and save the effort? David ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
2009/5/6 Jochen Topf joc...@remote.org: On Tue, May 05, 2009 at 09:14:30PM +0200, andrzej zaborowski wrote: Some notes: There's already the wikipedia=NAME tag (wikipedia=LA:NAME for non-english wikipedias, where LA=en,de...) in use in some places, so I'd recommend using that. Shouldn't that be wikipedia:LA=NAME ? I think wikipedia=XX:NAME was choosen because the other way you can add many wikipedia:XX= tags and the parsers will not have a single tag name to look at, also this would duplicate wikipedia's own mechanism for linking to different language pages (and what if the translation in OSM is inconsistent wikipedia's own?). By linking to only one wikipedia article in any language, you actually link to a whole group of articles in all wikipedias that are connected with their interlinks. I don't remember whether this was discussed here or on IRC. Cheers ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
andrzej zaborowski wrote: I think wikipedia=XX:NAME was choosen because the other way you can Chosen? Where? As far as I can see the only discussion is at http://wiki.openstreetmap.org/wiki/Proposed_features/External_links#Wikipedia and that says wikipedia:XX=article name in that language Besides, as present you can't have multiple values for a single tag in OSM, so wikipedia=XX:name wouldn't work. -- Jonathan (Jonobennett) ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
2009/5/6 Jonathan Bennett openstreet...@jonno.cix.co.uk: andrzej zaborowski wrote: I think wikipedia=XX:NAME was choosen because the other way you can Chosen? Where? On this list I think (or was that irc). Besides, as present you can't have multiple values for a single tag in OSM, so wikipedia=XX:name wouldn't work. You want only *one* value, you don't want to allow multiple values for the reasons I just gave. Cheers ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Wed, May 06, 2009 at 11:04:12AM +0100, Jonathan Bennett wrote: Besides, as present you can't have multiple values for a single tag in OSM, so wikipedia=XX:name wouldn't work. You don't need multiple values. Other languages are linked in Wikipedia, no need to duplicate this in OSM. If a place is described in 20 national Wikipedias do we really want 20 wikipedia=XX:name tags in OSM when only single wikipedia=XX:name links to all the pages? And more. You need no special handling for wikipedia=XX: links. E.g lets assume we have a client which opens a browser with Wikipedia page for a clicked POI using the wikipedia=* tag. And it has no language support (doesn't read wikipedia:XX=* tags, does not understand wikipedia=XX:*) and it gets: wikipedia=pl:Radiostacja_gliwicka It will open http://en.wikipedia.org/wiki/pl:Radiostacja_gliwicka Try that link yourself. You will get information about that place. Then click English on the sidebar. You will get English description about the same place. All with only one tag. With no special language support in the client. And it will work for other languages if localized Wikipedia pages are created. wikipedia=Gliwice_Radio_Tower Would also work, but would show the English page (Polish page is accessible with the Polski link on the left). But I guess, this POI is mostly interesting for Polish users, so polish page is a sane default. Greets, Jacek ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Tue, May 05, 2009 at 02:00:42PM -0400, Russ Nelson wrote: Any reason not to go through Wikipedia and import everything with a coordinate as a POI, with a url=http://wikipedia.org/NAME link, and name=NAME where NAME is the name of the Wikipedia entry? This would add a lot of data with no review. A lot of duplicates will be created making confusion about which one is more correct. Now, if a place is missing in OSM anybody can look it up in Wikipedia anyway and put it in OSM. That would be well-thought-out decision and even if wrong, that is a single mistake. Much less probable someone alive would add a duplicate or outdated information. And what is a use of Wikipedia POIs in areas where we have no other map data in OSM and no mappers to verify the import? IMHO it is much better to have good data then to have a lot of data. I see no reason for an automated import of that data. Greets, Jacek ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Wed, May 06, 2009 at 11:48:02AM +0200, andrzej zaborowski wrote: 2009/5/6 Jochen Topf joc...@remote.org: On Tue, May 05, 2009 at 09:14:30PM +0200, andrzej zaborowski wrote: Some notes: There's already the wikipedia=NAME tag (wikipedia=LA:NAME for non-english wikipedias, where LA=en,de...) in use in some places, so I'd recommend using that. Shouldn't that be wikipedia:LA=NAME ? I think wikipedia=XX:NAME was choosen because the other way you can add many wikipedia:XX= tags and the parsers will not have a single tag name to look at, also this would duplicate wikipedia's own mechanism for linking to different language pages (and what if the translation I am not sure I understand your arguments. Using wikipedia:language as key is used in OSM today in many cases. It is consistent with name:language and other uses of :language in keys, so it makes immediate sense to OSMers. Also Wikipedia uses something: as their way of doing namespacing. But generally not for languages, but for User pages, Special pages, Meta pages etc. Its in the OSM wiki that we use language names in this way, not in Wikipedia. in OSM is inconsistent wikipedia's own?). By linking to only one wikipedia article in any language, you actually link to a whole group of articles in all wikipedias that are connected with their interlinks. I don't remember whether this was discussed here or on IRC. This is true, but those interwiki links don't always exist and if they do, they might not refer to the right page. I'd rather have the option of adding explicit links to the pages I want to link to. Of course the interwiki links would work as fallback. So I think the language must be in the OSM key, not the value. Translating this into a link is straightforward then: wikipedia:en=London = http://en.wikipedia.org/wiki/London With your way wikipedia=en:London could be misunderstood as meaning http://en.wikipedia.org/wiki/en:London which could exist. Jochen -- Jochen Topf joc...@remote.org http://www.remote.org/jochen/ +49-721-388298 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
Jacek Konieczny wrote: You don't need multiple values. Other languages are linked in Wikipedia, no need to duplicate this in OSM. If a place is described in 20 national Wikipedias do we really want 20 wikipedia=XX:name tags in OSM when only single wikipedia=XX:name links to all the pages? If you want to be able to render maps for a specific language or territory, yes. Having the WP shortcut version means that if I wanted to link to the English version of a WP page when we have only the Polish version, I have to retrieve the Polish page from WP and parse it to find the link to the English version. This means I can't do this rendering offline, and would slow the job down somewhat. And more. You need no special handling for wikipedia=XX: links. Except you do -- special handling from an OSM point of view. The convention of tag:lang = value is already well-established. -- Jonathan (Jonobennett) ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [OSM-newbies] avoid repeating the name tag twice
On Tuesday 05 May 2009, Emilie Laffray wrote: What might be interesting and worth a discussion: A tag to describe the default language of this object, e.g. language=en. This could als be several tags, e.g. name=België - Belgique - Belgien name:nl=België name:fr=Belgique name:da=Belgien (and some more) language=nl;fr;da- this would be new (I hope I got the languages right, sorry if not) It should be de for German, but anyway. I like the idea of the language element. I would like to add an extra precision in this case. I think the order of language should be by importance. Unfortunately, in this case namely the country, it is something highly political to even consider an order. But, in the case of a town located in the French speaking part, it would be logical to put something like fr;nl;da instead of nl;fr;da. The reverse would be true. It's not really like that. In most of the country there's only one single official language. It's only a problem if you take the country as a whole, or look at Brussels: For the country name, it's just ordered according to number of people in each language community currently. Not exactly a controversial way of tagging. Brussels, which is officially bilingual French and Dutch, is a different beast. While it's uncontested that French is spoken more there, it's a real political minefield and one has to be careful to keep the two languages on par. Speaking of one language being more important than the other could easily provoke lively discussions or even edit-wars. But luckily we didn't have anything like that in OSM in Belgium yet and everyone is just tagging the French and Dutch names in the order they like in the name tag. But if we're going to add a tag that defines an order of importance, that can easily change. If it's possible at all to define an order, because while overall French is spoken more in Brussels, there are certainly areas with more Dutch speakers. So, in short: I'm not sure a language tag would work here. The languages should be known from the political entity it's located in, and since the places in areas with more than one official language are tagged with name:nl, name:fr etc, that's more than enough already, and a language tag gives unnecessary information. And it's especially a bad idea if it defines a certain order of importance. Ben ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [OSM-newbies] avoid repeating the name tag twice
On Wed, May 06, 2009 at 01:53:57PM +0200, Ben Laenen wrote: So, in short: I'm not sure a language tag would work here. The languages should be known from the political entity it's located in, and since the places in areas with more than one official language are tagged with name:nl, name:fr etc, that's more than enough already, and a language tag gives unnecessary information. And it's especially a bad idea if it defines a certain order of importance. But how does the system know which is the official language for that region? As far as I know, there's no function yet, so a language-tag would be useful (and could for example be applied to the administrative boundary as to save redundant information). greetings, Stephan -- Seid unbequem, seid Sand, nicht Öl im Getriebe der Welt! - Günther Eich ,-. | Stephan Plepelits, | | Technische Universität Wien -Studium Informatik Raumplanung | | openstreetbrowser.org couchsurfing.com tubasis.at bl.mud.at | | sk...@xover.htu.tuwien.ac.at - My Blog: http://plepe.at | `-' ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
http://en.wikipedia.org/wiki/London http://en.wikipedia.org/wiki/en:London Both links go to the same Wikipedia page. Ed ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Wed, May 6, 2009 at 11:11 AM, Jonathan Bennett openstreet...@jonno.cix.co.uk wrote: Jacek Konieczny wrote: You don't need multiple values. Other languages are linked in Wikipedia, no need to duplicate this in OSM. If a place is described in 20 national Wikipedias do we really want 20 wikipedia=XX:name tags in OSM when only single wikipedia=XX:name links to all the pages? If you want to be able to render maps for a specific language or territory, yes. Having the WP shortcut version means that if I wanted to link to the English version of a WP page when we have only the Polish version, I have to retrieve the Polish page from WP and parse it to find the link to the English version. This means I can't do this rendering offline, and would slow the job down somewhat. You can get a dump of Wikipedia and process that offline (just as you got a dump of OSM). There's also a current effort to get interwiki links into MediaWiki's database[1]. This allows for database dumps of the interwiki table[2] meaning you wouldn't have to maintain your own parser for grabbing these links. The current implementation of interwiki links in MediaWIki is the biggest cause of spam in article histories, it's not uncommon to have a small article with one real edit and 20-30 automated edits that amend the interwiki links. If we were going to list all of Wikipedia's interwiki links for applicable POIs in OSM that would eventually mean that we'd need our own interwiki bots connected to Wikipedia. That would suck. 1. http://www.mediawiki.org/wiki/WMDE_contract_offers/Store_interwiki-links_in_the_database 2. http://download.wikimedia.org/ ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [OSM-newbies] avoid repeating the name tag twice
On Wed, May 6, 2009 at 3:02 PM, Stephan Plepelits sk...@xover.htu.tuwien.ac.at wrote: On Wed, May 06, 2009 at 01:53:57PM +0200, Ben Laenen wrote: So, in short: I'm not sure a language tag would work here. The languages should be known from the political entity it's located in, and since the places in areas with more than one official language are tagged with name:nl, name:fr etc, that's more than enough already, and a language tag gives unnecessary information. And it's especially a bad idea if it defines a certain order of importance. But how does the system know which is the official language for that region? As far as I know, there's no function yet, so a language-tag would be useful (and could for example be applied to the administrative boundary as to save redundant information). Regarding the official language, or more precisely, which of the available languages to use, I've always felt that this is a rendering issue, sort of. I mean, that this is a higher level knowledge that should be an input to the rendering software, in addition to the osm db, much like the rule file. This new knowledge, which might reside in the rule file, should not be a part of the osm db. Everyone with sufficient knowledge and skills can draw themselves a nice map in their favorite language using a custom rule file: Dutch people in Brussels can leave out French names, Hebrew speakers can omit the Arab names, and vice versa. Why do you even need an official language? I guess you want to draw a single map, here on this site, which will be good for almost everyone. My solution to this situation would be to add polygon constraints support to the renderers rule files. For example, drop this line in the mapnik/t...@hrules file: - for Brussels (=polygon cordinates) replace names with :fr - :nl I placed france first. Any angry Duch want to start an edit war? Go ahead, fight over the poor rule file, not over the db. I hope it would be easier to continue the collaboration efforts on the db while disagreeing on the rule file. This would also allow other forms of localizations per country. Maybe it is custom in a certain country to draw motorways in yellow - they can have that on the osm site just for them. Sorry for barging in, just wanted to throw in my ideal solution. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On May 6, 2009, at 1:14 AM, Rob Reid wrote: Excellent, so there is nothing to stop me tracing my entire town off Google Imagery into osm, since all I would be doing is choosing points off their aerial photograghs and they are not contributing in any way to me doing that? WHERE do you guys get these weird ideas about copyright from? Do you think the garbage man has copyright interest in your OSM work because he's contributing to your efforts by taking out your trash? What about the company that built your computer? SURELY you could not edit OSM without their contribution. What about the ISP that carried the imagery to your computer? You can bet that THEY think they share copyright in your online works. How could you have produced them without your ISP, so your ISP owns everything you edit online. Copyright protects creativity (or sweat of the brow). It doesn't protect contributions to creativity. Think of the lawsuits if it did! Why, I'm contributing to your OSM efforts by clarifying copyright for you, so I would have a copyright interest in your OSM edits if your theory was correct. -- Russ Nelson - http://community.cloudmade.com/blog - http://wiki.openstreetmap.org/wiki/User:RussNelson r...@cloudmade.com - Twitter: Russ_OSM - http://openstreetmap.org/user/RussNelson ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On May 6, 2009, at 2:03 AM, Richard Fairhurst wrote: Russ Nelson wrote: What work or creativity did Google do towards the existence of that particular point? Google's imagery suppliers collected and rectified the imagery. For over a hundred years, English courts have held that a significant expenditure of labour is sufficient - that's, er, Wikipedia saying that. If they'd rectified them differently, your 14 digits would be different. True, but now you're putting yourself into the ridiculous situation of claiming that every possible set of coordinates infringes Google's copyright. Openstreetmap is about gathering map data and sharing it. Some people seem desperate to import data from anywhere. GATHER IT YOURSELF. The problem is that people say Why should I have to repeat this work? It's already been done. Why can't we just import it? -- Russ Nelson - http://community.cloudmade.com/blog - http://wiki.openstreetmap.org/wiki/User:RussNelson r...@cloudmade.com - Twitter: Russ_OSM - http://openstreetmap.org/user/RussNelson ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Wed, May 6, 2009 at 5:06 PM, Russ Nelson r...@cloudmade.com wrote: So, if I understand this discussion, I cannot create a POI based on Google aerial photography directly in OSM. But if I create my POI first in Wikipedia, then import it in OSM, it is permitted. Is that correct ? Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
At freemap.sk we have choosen not to import wikipedia POIs into OSM database to prevent long discussions about unclear copyright and other rights. We have Wikipedia as transparent click-able overlay with popups with short description of WIKIPEDIA POI and also with direct link to Wikipedia article. http://dev.freemap.sk/?zoom=10lat=48.47825459688335lon=18.094623150449117; maplayers=Wikipediamaplayer=Standard Dodi -Original Message- From: talk-boun...@openstreetmap.org [mailto:talk- boun...@openstreetmap.org] On Behalf Of Russ Nelson Sent: Wednesday, May 06, 2009 4:54 PM To: Talk Openstreetmap Subject: Re: [OSM-talk] Wikipedia POI import? On May 6, 2009, at 1:14 AM, Rob Reid wrote: Excellent, so there is nothing to stop me tracing my entire town off Google Imagery into osm, since all I would be doing is choosing points off their aerial photograghs and they are not contributing in any way to me doing that? WHERE do you guys get these weird ideas about copyright from? Do you think the garbage man has copyright interest in your OSM work because he's contributing to your efforts by taking out your trash? What about the company that built your computer? SURELY you could not edit OSM without their contribution. What about the ISP that carried the imagery to your computer? You can bet that THEY think they share copyright in your online works. How could you have produced them without your ISP, so your ISP owns everything you edit online. Copyright protects creativity (or sweat of the brow). It doesn't protect contributions to creativity. Think of the lawsuits if it did! Why, I'm contributing to your OSM efforts by clarifying copyright for you, so I would have a copyright interest in your OSM edits if your theory was correct. -- Russ Nelson - http://community.cloudmade.com/blog - http://wiki.openstreetmap.org/wiki/User:RussNelson r...@cloudmade.com - Twitter: Russ_OSM - http://openstreetmap.org/user/RussNelson ___ 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] Wikipedia POI import?
Russ Nelson wrote: WHERE do you guys get these weird ideas about copyright from? Tell you what. You work for CloudMade, right? I suggest you ask your bosses. Show them what you're proposing to import. Show them the Wikipedia page that explains how it's been gathered. Ask them if they'd be happy with that in their dataset and are prepared to run the legal risk. I'd be interested to know their response. (This would, of course, be better on legal-talk.) cheers Richard -- View this message in context: http://www.nabble.com/Wikipedia-POI-import--tp23392791p23408966.html Sent from the OpenStreetMap - General mailing list archive at Nabble.com. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
El Miércoles, 6 de Mayo de 2009, Russ Nelson escribió: The problem is that people say Why should I have to repeat this work? It's already been done. Why can't we just import it? Why should I have to repeat this work? It's already been done by TeleAtlas. Why can't we just import it? -- -- Iván Sánchez Ortega i...@sanchezortega.es Aviso: Este e-mail es confidencial y no debería ser usado por nadie que no sea el destinatario original. No se permite la reproducción mediante fotocopia, walkie-talkie, emisora de radioaficionado, satélite, televisión por cable, proyector, señales de humo, código morse, braille, lenguaje de signos, taquigrafía o cualquier otro medio. Bajo ningún concepto debe traducirse al francés este e-mail. Este e-mail no puede ser ridiculizado, parodiado, juzgado en una competición, o leído en voz alta con un acento gracioso llevando un bigote falso y/o cualquier tipo de sombrero, incluyendo pero no limitándose a pañuelos. No inciten ni provoquen a este e-mail. Si está medicándose, puede experimentar nauseas, desorientación, histeria, vómitos, pérdida temporal de la memoria a corto plazo y malestar general al leer este e-mail. Consulte a su médico o farmacéutico antes de leer este e-mail. Todas las modelos descritas en este e-mail son mayores de 18 años. Si ha recibido este e-mail por error es probablemente porque estaba borracho cuando escribí la dirección del destinatario. 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] Wikipedia POI import?
From: Pieren pier...@gmail.com To: Russ Nelson r...@cloudmade.com Cc: Talk Openstreetmap talk@openstreetmap.org Sent: Wednesday, 6 May, 2009 16:17:51 Subject: Re: [OSM-talk] Wikipedia POI import? On Wed, May 6, 2009 at 5:06 PM, Russ Nelson r...@cloudmade.com wrote: So, if I understand this discussion, I cannot create a POI based on Google aerial photography directly in OSM. But if I create my POI first in Wikipedia, then import it in OSM, it is permitted. Is that correct ? I think there are two issues here, which are only partially related. Firstly, there is what the relevant laws actually say on the matter. So for example, the laws say that if Wikipedia licenses its data as CC-BY-SA we can import it without worrying about it. If they didn't have the right to license the data as such, that is primarily their problem and we only have to respond if a court decides that this was improper and we should therefore remove the data. I am fairly happy with doing a mass import of Wikipedia data on this basis. Secondly, there is the risk that a company, irrespective of what the laws say, will start throwing lawsuits around. I am not convinced that Google would do this in this case, but those from whom they license the data might do. It doesn't matter whether they have a case or not, their claims could be completely and utterly bogus. But once they've thrown their lawsuit, you are then tied up in an expensive legal process from which it can be very hard to escape unless your lawyers are bigger than theirs (or more accurately, the pockets that pay your lawyers are deeper than the pockets that pay theirs). Think of this as the Microsoft approach to the law. It is this risk that we are more concerned about here, and this is what would make me wary of doing a mass import. If our product is a threat to their profitability then they will throw morals out of the Windows and do whatever they can to halt their competitors, claiming that they have a moral duty to look after the interests of their shareholders. Much as I hate the fact that this is how it works, it's a sad fact of this world that there are many large organisations who don't care about right or wrong who also happen to have pretty deep pockets. Cheers, Donald ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
Russ Nelson wrote: On May 6, 2009, at 2:03 AM, Richard Fairhurst wrote: Russ Nelson wrote: What work or creativity did Google do towards the existence of that particular point? Google's imagery suppliers collected and rectified the imagery. "For over a hundred years, English courts have held that a significant expenditure of labour is sufficient" - that's, er, Wikipedia saying that. If they'd rectified them differently, your 14 digits would be different. True, but now you're putting yourself into the ridiculous situation of claiming that every possible set of coordinates infringes Google's copyright. "Openstreetmap is about gathering map data and sharing it. Some people seem desperate to import data from anywhere. GATHER IT YOURSELF." The problem is that people say "Why should I have to repeat this work? It's already been done. Why can't we just import it?" I wrote the tweet Richard posted. The problem is that much of the data that is imported is poor quality. If we import data from other sources, the best we can ever be is a me-too map no better than the data we import. I don't want to create an also-ran map that happens to be free to use, I want to create a best-ever map and by-the -way it's free, and I can't do that by just copying someone else's stuff. At the weekend I heard from someone visiting Yorkshire who now lives in Arizona. He thinks OSM is poor because the TIGER import is incomplete and off by about 100 yards. I tried to persuade him to help, and he can best do that by gathering data. There are good imports of course and some imported data can be better than nothing, but just importing every thing and anything is not going to build that earth-moving map. So lets get back to creating the very best map of the world, part of which involves GATHERING DATA. Cheers, Chris ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Fwd: Gaza WMS server unavailable?
Aevar, The server changed, I'll send you the details directly. Note that this server is not public, because of restrictions on the imagery. Mikel From: Ævar Arnfjörð Bjarmason ava...@gmail.com To: openstreetmap talk@openstreetmap.org; Simone Cortesi sim...@cortesi.com Sent: Tuesday, May 5, 2009 8:02:53 AM Subject: [OSM-talk] Fwd: Gaza WMS server unavailable? Forwarding this to talk@ since the talk-ps@ moderator isn't allowing this through. -- Forwarded message -- From: Ævar Arnfjörð Bjarmason ava...@gmail.com Date: Wed, Apr 29, 2009 at 11:38 PM Subject: Gaza WMS server unavailable? To: talk...@openstreetmap.org Is the secret 2m WMS for Gaza unavailable?. I'm getting connection refused when trying to contact the usual server. ___ 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] Wikipedia POI import?
On May 6, 2009, at 11:06 AM, Russ Nelson wrote: On May 6, 2009, at 2:03 AM, Richard Fairhurst wrote: Openstreetmap is about gathering map data and sharing it. Some people seem desperate to import data from anywhere. GATHER IT YOURSELF. The problem is that people say Why should I have to repeat this work? It's already been done. Why can't we just import it? There's also the problem that there is some data which should be on the map which cannot be gathered in the field. For example, the NYS DEC Lands dataset is not discernable in the field. You can look at the signs, but the signs are not as accurate as the maintained shapefile. So should this data NOT be in the map? Someone certainly count gather it, but it is KNOWN to be less accurate than the shapefile we can simply import. THAT is why we import rather than gathering data. -- Russ Nelson - http://community.cloudmade.com/blog - http://wiki.openstreetmap.org/wiki/User:RussNelson r...@cloudmade.com - Twitter: Russ_OSM - http://openstreetmap.org/user/RussNelson ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On May 6, 2009, at 11:29 AM, Iván Sánchez Ortega wrote: El Miércoles, 6 de Mayo de 2009, Russ Nelson escribió: The problem is that people say Why should I have to repeat this work? It's already been done. Why can't we just import it? Why should I have to repeat this work? It's already been done by TeleAtlas. Why can't we just import it? WHERE do you guys get your ideas about copyright from?? Sheesh, you're being a poopy-head, Iván. TeleAtlas data is copyrighted, and when licensed is licensed under an incompatible copyright. Other data is either in the public domain or not copyrightable or licensed under a compatible copyright. I'm like, well duh! -- Russ Nelson - http://community.cloudmade.com/blog - http://wiki.openstreetmap.org/wiki/User:RussNelson r...@cloudmade.com - Twitter: Russ_OSM - http://openstreetmap.org/user/RussNelson ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Relation/Routes and Hikes in open Country
I enter a bit late in this discussion. But this was the only place I found which discussed the issued of the nonoffical routes, because I would like to do the same, so a kind of sharing of unoffical cycling routes. with OSM. I have a kind of mixed feeling about this. One one side I understand that the fundamental data of the OSM data base should be based on verifiable facts on the ground but on the other side there are some use cases which the current OSM data base does not (officialy) address. The first one is exactly the question of unoffical route proposals. I find it very covinient and logical to create my private routes as relations in JOSM by assmembling them from existing ways, such as any normal relation. But I agree that we should use special tagging for it, so that by default they do not appear on an offical map. But I would like to be able to link to them somehow or enable them. And it is true that the questions about when does a route changes from subjective to objective is quite tricky to decide. Are the requirements for making a route objective are really needed and a clear criteria of adding a route to the OSM database, such as: - marked on the ground - maintained and supported by an organisation This factors can change over time, so maybe the province stops maintaing a certain route, so the signs get invisble, the maps disappear. Should we then remove it from OSM ? (Or think about the future in which everyboddy has a GPS with OSM maps, so marking of routes is not needed any more ...) Maybe an individual person can maintain a route better then an organisation (even by just maintaining a web site with up-to-date information) I am not sure if the existing tags such as operated by are a good mechanism or not to diffenciate this. Maybe it would be just easier to simply use a tag such as marked=no which simply says if or if not we consider a route to follow the rules of verifiable on the ground or not and the renderes by default ignore them. Regards, Carsten I'd define it slightly differently - its do we want *subjective* routes in OSM? I don't think anyone is arguing that notable *objective* routes, like the Pennine Way in the UK or the Appalachian Way in the US can certainly be included as a route. Hi all, I'm going to be a bit provocative here, please bear with me 1) At what point does a route change from being 'subjective' to 'objective'? 2) 'Open Street Map is a map of everything'. 3) 'Your map, your way'. I understand the concern that we don't want the official map to be saturated with additional non-official (whatever that means) routes, however I don't think that it is a reason to prevent people/organisations adding there own relation/routes to the data base. At present the offical map does not render the relation/routes, when it does it can limit the ones it shows by using the operator and/or network tags. In the case of Bob Spirko, there is a huge resource of write-up and photos on his website. I believe that it is a benefit to add this information to the OSM database and (in my opinion) relation routes are the best way to do this. The first batch of trails are actually based around those published in a book, does this make them 'objective'? For OSM this get us additional ways on the ground, showing footpaths and tracks on the ground. The use of relations removes the surplus naming of ways (ie. a footpath would not have to be tagged 'Anderson Peak Trail' for example) and other marking of non-physical things. For Bob Spirko (or whomever) it gives the ability to render maps showing his routes (which can be done offline with osmarender or some other scheme) or to make GPS compilations for navigation. I've not got much problem with notable subjective routes being stored in the database. Obviously notable is a vague moving target, but what I mean here is that it's not a completely arbitrary concoction that no-one else has ever heard of. If we have a well known person writing books on walks etc then I see no problem in adding these to the DB. The problem is how you add them. A route relation, as in type=route, route=foot, has so far implied a signed, official route. This is what is being referred to as an objective route -- the key feature is generally that it is maintained by someone (the operator). For the subjective routes where someone has gone and looked for a nice walk, wrote it down and published it, this probably isn't the case. I'd rather the former weren't diluted by the latter, with no easy way to distinguish them. What I'd suggest is that you just change some of the tag names: type=suggested_route route=foot suggested_by=Bob Spirko Then everyone knows where they stand, and what this represents. I'm sure there are people convinced this is a bad idea... and if the goal is to avoid conversations about what is considered notable then I'd probably take their point. Dave Carsten Behring Software
Re: [OSM-talk] gdalwarp question
Hello Jukka, no, thanks for your help, any hint and discussion is really appreciated. Sorry, I misunderstood a bit what you were going to do. It may well be that for Mapnik you'll need to reproject raster image first. I do not much about Mapnik and while a have been using gdal utilities very much I don't believe I have ever needed to create output in Google projection. By first look what you have done does make sense. You can make a couple of further tests: I solved the issue now in a different way. Based on information on wikipedia about mercaator projection i did the reprojection of the blue marble myself. It took quite a while to execute, but the reprojection itself worked quite fine. Now the borders (shoreline_300) are exactly on the borders of the continents and islands on the blue marble. The relevant part of the script i use is attached below, if anybody is interested please write me an email (i use an own library to handle the images themselves, that part needs to be adapted to e.g. PyImage). Thank you all for providing me help, i hope the script below is of value to anybody. Best regards, Torsten. #! /usr/bin/python #coding: latin-1 import ri from math import * import gc import sys #he = 21600 he = 43200 np = 8 skp_a = 90.0 - (atan(sinh(pi)) * 180.0 / pi) print skp_a skp_p = int(he * skp_a / 90.0) print skp_p #sys.exit(0) pat = '/local/vid/earth/strs_%i.tif' max_phi = atan(sinh(pi)) print max_phi, max_phi * 180 / pi def do_img(img, i): wi = img.width() hi = img.height() yi0 = hi / 2 wo = wi ho = he yo0 = ho / 2 print i, i, wo, wo, ho, ho out = ri.Ri(wo, ho, 1, 0) for x in range(wo): for ydo in range(ho / 2): yn = float(ydo) / (ho / 2) * pi phi = asin(tanh(yn)) ydi = phi / max_phi * (hi / 2) r, g, b, a = img.get(x, int(yi0 + ydi)) out.set(x, int(yo0 + ydo), r, g, b) r, g, b, a = img.get(x, int(yi0 - ydi)) out.set(x, int(yo0 - ydo), r, g, b) out.saveAsPng(/local/vid/earth/ta_%i.png % (i)) del out for i in range(4, np): img = ri.Ri(pat % (i), ri.TIFF) do_img(img, i) del img gc.collect() ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Tue, May 05, 2009 at 09:41:27PM +0200, andrzej zaborowski wrote: I very much agree about OpenAerialMap -- if we can't trust the OpenAerialMap contributors about the licensing why should any person in OSM trust any other OSM contributor rather than start redrawing everything they can from scratch and only trust their own contributions. Hmm, trust. So, it has a messaging system, a blogging system, and a little bit of a “friends” network… OSM could build it’s own web of trust system too! *runs away giggling* Actually, seriously, I would sign my changesets using OpenPGP if I could easily do so. Simon -- A complex system that works is invariably found to have evolved from a simple system that works.—John Gall signature.asc Description: Digital signature ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] shoreline_300, world borders _without_ rectangles
Hello, i'm glad i got the blu marble working nw and can create a map of the whole world with the satellite images of the earth as background. It looks really great. But what would be great if i could also overlay the world / continents / islands / states borders as outline. I can't use the shoreline_300 file as it is split into several polygons and adds heaps of plates to the earth and does not look good. Does anybody know a way to get the world outline (as in shoreline_300) but without the plates (interruptions of the polygons)? Is it maybe possible to get these data from PostGIS with a clever SQL request? Thanks for any hints, Torsten. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
Russ Nelson wrote: TeleAtlas data is copyrighted, and when licensed is licensed under an incompatible copyright. The data you're proposing taking from Wikipedia is probably derived, via Google, from that same TeleAtlas (or Navteq) data. It doesn't seem plausible that deriving information from TA data is fine, provided it's done at arm's length via a Wikipedian who didn't read the Google Maps terms of use. These explicitly say that you must not make derivative works of the Content or any part thereof, or give access to mass downloads or bulk feeds of any Content, including but not limited to numerical latitude or longitude coordinates. -dair ___ d...@refnum.com http://www.refnum.com/ ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Borders of the federal states of Germany (again)
Hello, this issue is not related to another post that i did here regarding the outline of the world (shoreline_300). I want to create a map of Germany with many details, especially the federal states should have different background colors. I got a hint already for a ShapeFile that contains these data. That file looked great, but the borderlines defined in there are _much_ more coarse than the data in the PostGIS, so creating the background from the rough, coarse data does not look that good. It does not match the borders that are drawn. Is there a way to get the border information from the PostGIS server that runs locally on my machine? Things i tried so far: - Google for it - Get the data from PostGIS with SQL commands - Download the single federal states from www.geofabrik.de as ShapeFiles and then find out they don't contain the borders. - Download Germany.shp from www.geofabrik.de as ShapeFiles and then find out they don't contain the borders. - Install QGIS and try to get the data from PostGIS (failed as my PostGIS installation does not seem to support GEOS). With the last topic there could still be a chance to get a result, i still need to try that. But it means to re-install PostGIS and add some steps that i'm not yet sure about (GEOS support). It would be great if anybody had a hint on this. Thanks for any hints, Torsten. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Relation/Routes and Hikes in open Country
This is complicated - but in principle I would support mapping something that is there - and used - on the ground, regardless of whether or not it is official. Indeed the definition of official will vary greatly from country to country and may not even be obvious. I agree that non-official organisations may be much better than official ones at maintaining (and mapping!) off-road routes ... whether cycle or on foot. In the UK we have a very elaborate and complex legal system around the whole issue of public rights of way (whether pedestrian, bicycle or equestrian) and the options for tagging these (in the UK) have been widely discussed in this newsgroup. More importantly, perhaps, in relation to your message ... We have a very large number of routes that are actually named as long-distance walking or cycling routes and are often promoted by (and sometimes even in part maintained by the official bodies) but include segments that are not public rights of way at all but are of a permissive or customary nature. I have a major and much-used medium-distance walking route a few km from home (the sg. Sandstone Trail) that is a mixture of public rights of way of various kinds, permissive and customary segments over which the public has no right of passage, plus sections of public roads. Yet it is widely promoted by the authorities to promote tourism in the region. It would seem perverse not to map this - and it would also seem appropriate to map it is a route relation. The individual sections can then be tagged with whatever seems appropriate for each. Hope this is pragmatic and a little bit helpful! Mike Harris -Original Message- From: BEHRING Carsten [mailto:carsten.behr...@efsa.europa.eu] Sent: 06 May 2009 18:44 To: talk@openstreetmap.org Subject: Re: [OSM-talk] Relation/Routes and Hikes in open Country I enter a bit late in this discussion. But this was the only place I found which discussed the issued of the nonoffical routes, because I would like to do the same, so a kind of sharing of unoffical cycling routes. with OSM. I have a kind of mixed feeling about this. One one side I understand that the fundamental data of the OSM data base should be based on verifiable facts on the ground but on the other side there are some use cases which the current OSM data base does not (officialy) address. The first one is exactly the question of unoffical route proposals. I find it very covinient and logical to create my private routes as relations in JOSM by assmembling them from existing ways, such as any normal relation. But I agree that we should use special tagging for it, so that by default they do not appear on an offical map. But I would like to be able to link to them somehow or enable them. And it is true that the questions about when does a route changes from subjective to objective is quite tricky to decide. Are the requirements for making a route objective are really needed and a clear criteria of adding a route to the OSM database, such as: - marked on the ground - maintained and supported by an organisation This factors can change over time, so maybe the province stops maintaing a certain route, so the signs get invisble, the maps disappear. Should we then remove it from OSM ? (Or think about the future in which everyboddy has a GPS with OSM maps, so marking of routes is not needed any more ...) Maybe an individual person can maintain a route better then an organisation (even by just maintaining a web site with up-to-date information) I am not sure if the existing tags such as operated by are a good mechanism or not to diffenciate this. Maybe it would be just easier to simply use a tag such as marked=no which simply says if or if not we consider a route to follow the rules of verifiable on the ground or not and the renderes by default ignore them. Regards, Carsten I'd define it slightly differently - its do we want *subjective* routes in OSM? I don't think anyone is arguing that notable *objective* routes, like the Pennine Way in the UK or the Appalachian Way in the US can certainly be included as a route. Hi all, I'm going to be a bit provocative here, please bear with me 1) At what point does a route change from being 'subjective' to 'objective'? 2) 'Open Street Map is a map of everything'. 3) 'Your map, your way'. I understand the concern that we don't want the official map to be saturated with additional non-official (whatever that means) routes, however I don't think that it is a reason to prevent people/organisations adding there own relation/routes to the data base. At present the offical map does not render the relation/routes, when it does it can limit the ones it shows by using the operator and/or network tags. In the case of Bob Spirko, there is a huge resource of write-up and photos on his website. I believe that it is a benefit to add this information to the OSM database and (in my opinion) relation routes are the best way to
[OSM-talk] RSS Feed for OSM API Changesets
Hello, I'm proud to announce my first version of a script to generate RSS 2.0 newsfeeds out of the new OSM changesets API. Detailed information could be found on my blog: http://www.steffenvogel.de/2009/05/06/osm-changesets-als-rss-feed/ greeting from Germany Steffen -- Steffen Vogel Karlstraße 5 64347 Griesheim Tel: +49 (6155) 78877 Handy: +49 (176) 96978528 EMail: i...@steffenvogel.de Web: http://www.steffenvogel.de ICQ: 176824121 MSN: i...@steffenvogel.de signature.asc Description: Dies ist ein digital signierter Nachrichtenteil ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On Wed, May 6, 2009 at 8:03 AM, Richard Fairhurst rich...@systemed.netwrote: For over a hundred years, English courts have held that a significant expenditure of labour is sufficient - that's, er, Wikipedia saying that. Has there been any sweat of the brow cases after the database directive has been implemented? In the Scandinavian countries a, somewhat, similar right exists (anyone who gathers a large number of of facts). I have seen legal arguments that this is invalid after the database directive, but has not been able to find any court cases that are relevant. - Gustav ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Xapi and version attribute
Is it possible that in the Xapi servers, the version attribute is only present in nodes that have been changed after the 0.6 transition? If I download data from Xapi, it is missing in most nodes, except for those edited after 2009-04-28 (in my dataset). I think this is quite an important issue as the 0.6 api will not accept uploads without a version. Regards, Maarten ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
On May 6, 2009, at 2:02 PM, Dair Grant wrote: Russ Nelson wrote: TeleAtlas data is copyrighted, and when licensed is licensed under an incompatible copyright. The data you're proposing taking from Wikipedia is probably derived, via Google, from that same TeleAtlas (or Navteq) data. Or OpenStreetMap data. How would you know? Perhaps TA and N have easter eggs. Just to make things interesting, there are many potential sources of the Wikipedia coordinates. It's quite possible that you would have both TA and N claiming that they own all of Wikipedia coordinates? But the use of easter eggs is to prove that a majority of the data is infringing because it can be proven that a minority of the data is without doubt infringing. What happens if two parties attempt to claim that they own all the data? But y'all are STILL focussing on the WRONG PROBLEM. Okay, here's what we have for objections: o Wikipedia editors are instructed to use Google Maps thus their geodata is potentially infringing. o We should be gathering our data from the field (so that means that the data we currently have is reliable enough, modulo any currently-known copyright problems). o But some of the Wikipedia POIs are already in OSM. Can you see how this points a way forward? We look at the Wikipedia lat/lons and POI names. We look in OSM for nearby POIs. We *replace* the Wikipedia lat/lons with OSM lat/lons. In fact, we turn this into a continuous process. When somebody enters a POI, we look in Wikipedia for that entity, and we link to the Wikipedia page and replace its lat/lon with our own. -- Russ Nelson - http://community.cloudmade.com/blog - http://wiki.openstreetmap.org/wiki/User:RussNelson r...@cloudmade.com - Twitter: Russ_OSM - http://openstreetmap.org/user/RussNelson ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
Russ Nelson wrote: On May 6, 2009, at 2:02 PM, Dair Grant wrote: Russ Nelson wrote: TeleAtlas data is copyrighted, and when licensed is licensed under an incompatible copyright. The data you're proposing taking from Wikipedia is probably derived, via Google, from that same TeleAtlas (or Navteq) data. Or OpenStreetMap data. How would you know? Perhaps TA and N have easter eggs. I actually did a paper on this last term (map easter eggs): Both TA and N are known to contain easter eggs (though I don't recall which of the two denies this publicly). But y'all are STILL focussing on the WRONG PROBLEM. Okay, here's what we have for objections: o Wikipedia editors are instructed to use Google Maps thus their geodata is potentially infringing. o We should be gathering our data from the field (so that means that the data we currently have is reliable enough, modulo any currently-known copyright problems). o But some of the Wikipedia POIs are already in OSM. Can you see how this points a way forward? We look at the Wikipedia lat/lons and POI names. We look in OSM for nearby POIs. We *replace* the Wikipedia lat/lons with OSM lat/lons. In fact, we turn this into a continuous process. When somebody enters a POI, we look in Wikipedia for that entity, and we link to the Wikipedia page and replace its lat/lon with our own. I guess I missed something...how is this not the obvious answer? I took it for granted that this was happening already. signature.asc Description: OpenPGP digital signature ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
N are known to contain easter eggs (though I don't recall which of the two denies this publicly). signature.asc Description: OpenPGP digital signature ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
Hi, Russ Nelson wrote: When somebody enters a POI, we look in Wikipedia for that entity, and we link to the Wikipedia page and replace its lat/lon with our own. * Possible if both Wikipedia and OSM are CC-BY-SA. * Impossible if you assume that the lat/lon is subejct to copyright and Wikipedia is GFDL (unsure what their current status is). * Very interesting if OSM should go ODbL. ODbL would allow the re-use of insubstantial amounts of data (one lat/lon pair = cleary insubstantial) but request that the repeated extraction of insubstantial amounts of data into another database would at some point constitute a derivative process and force you to put the other database under ODbL, something that would not be possible with Wikipedia. The latest drafts have some wording about compatible licenses but it is unclear to me if that could be used to declare CC-BY-SA compatible. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [OSM-newbies] avoid repeating the name tag twice
I like the idea of the language element. I would like to add an extra precision in this case. I think the order of language should be by importance. Unfortunately, in this case namely the country, it is something highly political to even consider an order. But, in the case of a town located in the French speaking part, it would be logical to put something like fr;nl;da instead of nl;fr;da. The reverse would be true. Belgium is one interesting place to look at; Spain might be an other place to look at. The area around Barcelona is likely to be named in Catalan nowadays rather than Castellano. The order of language is likely to be a minefield because we are talking about something highly political, but it would allow to support the concept of native language, which I believe is very important. Things like different alphabet might be also interesting to look at even if it is likely that it can be subsumed under translation. I don't think the relatively peaceful and stable country of Belgium is at all the only example where ranking the languages would be quite political and/or controversial. What about in New Zealand where a treaty between the Queen and the tribes of New Zealand established the Maori tribes as equals, and their language, culture and courts as equals - yet only a small percentage of the population speaks it? What order do you put the languages? There would be a political uproar if the government of NZ tried to suggest a ranking. Or in Tibet, where before the 1950's invasion by the communists everyone speaks Tibetan, but now gradually more and more people are shipped in who speak Mandarin Chinese and will probably vastly outnumber the Tibetans in their homeland within our lifetime? People in the Basque country might be a little horrified if you put Castillian in front of Basque (and some might even object to it being in the list). Regions of India have many languages and subdialects, sometimes switching in a borderless fashion within a small region/space. Lastly what actual value would ranking the languages spoken in region by importance give the project - ie could it be shown on a map, or interpreted in any device that would be meaningful? I'm just curious and playing devil's advocate - happy to be convinced that there is true benefit to a proposal like this... ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [OSM-newbies] avoid repeating the name tag twice
Hello, Yes, it is potentially a minefield. I clearly indicated so previously. Well, I guess what matters is what you are going to do with the data. Someone earlier raised the very valid point that most of the problem could actually be solved by using the rendering in Mapnik by specifying the language you want the results in. Maybe I am worrying for no reason but I could see cases where it would be useful for geocoding and routing. I could be wrong, but if someone sends you an address, it is going to be most likely in the native language of the person. The way I see things is that the order would only apply to one named field. I don't see it as a replacement for an entire country. In the initial email by Nick Black, he asked what was the convention for the street in Kiev. He was using both addr:en and addr:ua apparently omitting the default addr. Personally, I would prefer to have the local language than English. I also saw that problem with the problem we had with Lithuania where people realized it was an import of Teleatlas because the language was using the old Russian names. As for your example for the Basque people, I think it is the same example as the Catalan one. I have seen how Castellano has been removed from signs over the past few years. Anyway, to some extent, the fact that we are using only name:Rue du moulin rouge is an implicit admission that we are using a native language. I just thought that in some countries it could result in having clearer result. Emilie Laffray Joe Richards wrote: I like the idea of the language element. I would like to add an extra precision in this case. I think the order of language should be by importance. Unfortunately, in this case namely the country, it is something highly political to even consider an order. But, in the case of a town located in the French speaking part, it would be logical to put something like fr;nl;da instead of nl;fr;da. The reverse would be true. Belgium is one interesting place to look at; Spain might be an other place to look at. The area around Barcelona is likely to be named in Catalan nowadays rather than Castellano. The order of language is likely to be a minefield because we are talking about something highly political, but it would allow to support the concept of native language, which I believe is very important. Things like different alphabet might be also interesting to look at even if it is likely that it can be subsumed under translation. I don't think the relatively peaceful and stable country of Belgium is at all the only example where ranking the languages would be quite political and/or controversial. What about in New Zealand where a treaty between the Queen and the tribes of New Zealand established the Maori tribes as equals, and their language, culture and courts as equals - yet only a small percentage of the population speaks it? What order do you put the languages? There would be a political uproar if the government of NZ tried to suggest a ranking. Or in Tibet, where before the 1950's invasion by the communists everyone speaks Tibetan, but now gradually more and more people are shipped in who speak Mandarin Chinese and will probably vastly outnumber the Tibetans in their homeland within our lifetime? People in the Basque country might be a little horrified if you put Castillian in front of Basque (and some might even object to it being in the list). Regions of India have many languages and subdialects, sometimes switching in a borderless fashion within a small region/space. Lastly what actual value would ranking the languages spoken in region by importance give the project - ie could it be shown on a map, or interpreted in any device that would be meaningful? I'm just curious and playing devil's advocate - happy to be convinced that there is true benefit to a proposal like this... signature.asc Description: OpenPGP digital signature ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Denmark contact
Is there anyone here from the CPH area that can be a contact for organising an OSM event? We've been offered some space to do something near a conference, reboot 11. Please drop me a line. PS the Denmark wiki page needs love, doesn't even link to the DK mailing list. Well, now it does. http://wiki.openstreetmap.org/wiki/Denmark Best Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia POI import?
El día Wednesday 06 May 2009 21:32:49, Gustav Foseid dijo: Has there been any sweat of the brow cases after the database directive has been implemented? I've got JUR 2007, 166551 (Sentencia de la Audiencia Provincial de Madrid, sección 13ª, del 16 de octubre de 2006 or Madrid provincial courthouse (13rd section) ruling, 16th oct 2006), regarding Ibercarta. (FYI, the EU DB directive was implemented into spanish law circa 1998) The ruling is brief on details regarding the DB directive per se, but acknowledges that the very delicate labour of data interpretation and recopilation is what gives the work originality, and that originality grants the copyright protection. Geez, I shouldn't be posting this on t...@. Is somebody ever going to follow-up this to legal@ ?? -- Iván Sánchez Ortega i...@sanchezortega.es Un ordenador no es una televisión ni un microondas: es una herramienta compleja. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] RSS Feed for OSM API Changesets
On Wed, May 6, 2009 at 2:27 PM, Steffen Vogel i...@steffenvogel.de wrote: I'm proud to announce my first version of a script to generate RSS 2.0 newsfeeds out of the new OSM changesets API. Detailed information could be found on my blog: http://www.steffenvogel.de/2009/05/06/osm-changesets-als-rss-feed/ Well, that's way cool... I don't read German so I'm not sure what you said in your blog entry but the example URLs are clear enough. Now if there were just some way to better visualize the changes being made... -- Jeff Ollie ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Nominations for State of the Map Travel Scholarships
The OpenStreetMap Foundation is excited to announce a program to cover full travel and accomodations costs for 15 mappers to attend State of the Map. We're seeking nominations from the community for potential mappers. Generally, we are seeking people from places where costs would prohibit attendance, developing countries, and places that are interesting geopolitically. The ideal candidates for funding are from countries with a small OSM community, perhaps just a few mappers in total. They have made a significant start at mapping their city, either through Yahoo imagery or with their own GPS, and are directly familiar with the process of OSM. They may have started communicating among themselves, and made plans and scoped out the process for their local district. But, the community is nowhere near critical mass, and they need the inspiration and support to take OSM to the next level. We need to act fast. State of the Map is just over two months away, tickets and visas need to be arranged. In order to allow enough time for all the arrangements, the nomination period will be short, one week only, ending next Wednesday, May 13. From the nominations received, we'll review the list and choose 15 mappers to approach with the offer. Depending on their availability to attend, we'll work our way through the list. We only recently secured funding for this program, so the process has to be quick. Please send your nominations to sotm.scholars...@gmail.com. For each nomination, include the mappers name, OSM user name, email address, location, and a paragraph or two on why they'd be great to have at SOTM. And also, please forward this message to other relevant local OSM lists. As for regions, here are a few regions that seem to fit the bill, but nominations are not limited to these places at all. * Eastern Europe: the Caucasus, Russia, Bulgaria, Romania. * Arab States: Lebanon, Jordan, Egypt. Egypt is particularly interesting, as the ban on GPS units there was recently lifted. * South Asia: India, Pakistan. While both countries have seen significant activities, relative to size and population they are in the very early stages. * Southeast Asia: Phillipines, Vietnam, Thailand. * South America: Colombia, Bolivia, Peru. Promising leads in government for the release of data for use in OSM. * East Africa: Kenya is a hotspot for mapping right now (Ushahidi, AgCommons, MapMaker..) Many thanks to the Open Society Institute for helping make this happen. -Mikel ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk-nl] Landduinen taggen
hehe, klinkt goed ;) ik denk dat we onze vlieghoogte dan wel moeten opschroeven naar 100m pakken we dan ook meteen even het strand mee ? Op 6 mei 2009 00:41 heeft Milo van der Linden m...@opengeo.nl het volgende geschreven: 1. Stefan of Rob benaderen 2. Een zonnige windstille dag uitkiezen 3. Een aantal referentie-paspunten uitzetten 4. de QuadCopters zoals beschreven op blog.opengeo.nl de lucht in jagen 5. de luchtfoto's georefereren 6. de duinen overtekenen in josm of merkaartor Klinkt dat als iets leuks? ;-) Leon Vrancken wrote: Hallo, Weet iemand hoe je landduinen kan taggen? Groet, Leon ___ Talk-nl mailing list Talk-nl@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-nl ___ Talk-nl mailing list Talk-nl@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-nl ___ Talk-nl mailing list Talk-nl@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-nl
[OSM-talk-nl] openstreetphoto
Hi! I am a french engineer very interested in the openstreetphoto project. I have already worked in image processing and matching (using sift, surf, cross correlation), and bundle adjustment. I've made research on Structure From Motion for car navigation. These are topics that seem to be very important in this project and I think I could help. There are very few information on the web but I think I am at the right place to talk. If I am wrong please tell me where to go. Persons from the project (if interested) can contact me by email. Regards, Etienne ___ Talk-nl mailing list Talk-nl@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-nl
Re: [OSM-talk-nl] openstreetphoto
Hi Etiene, On Wed, 6 May 2009, echienne m wrote: I am a french engineer very interested in the openstreetphoto project. I have already worked in image processing and matching (using sift, surf, cross correlation), and bundle adjustment. I've made research on Structure From Motion for car navigation. Sounds like an academic mind we need ;) These are topics that seem to be very important in this project and I think I could help. There are very few information on the web but I think I am at the right place to talk. If I am wrong please tell me where to go. Persons from the project (if interested) can contact me by email. Currently one student from GSOC has applied to do recognition on streetsigns and trafficsigns. I am currently doing some stuff in the direction of orthocorrection, and stitching. Specifically doing this not only on a point cloud but constraining on lines. (Next to the flying part of the project.) As you might have seen there is a lot of material available, even more stored locally. So if you would like to help, brainstorm or something like it #osp on irc.oftc.net would be the place. Stefan ___ Talk-nl mailing list Talk-nl@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-nl
Re: [Talk-de] Garminkarte jetzt mit maxspeed
Hallo, Frederik Ramm schrieb: Was genau muesste osmcut denn tun, um diese Luecken zu vermeiden? Dann Um die Lücken zu vermeiden muß in die Kacheln, in der ein Weg auftaucht alle nodes hinein, die der Weg benutzt. Das ist zwar manchmal richtig viel (daher meine gelegentlichen Riesenkacheln) aber nur den ersten Punkt in der Nachbarkachel zu nehmen reicht nicht. Mein Cutter merkt sich für alle Nodes die Zielkachel und die Koordinaten, danach kann er beim Bearbeiten der Ways die Nodes heraussuchen, die er alle braucht und ebenfalls in die richtige Zielkachel kopieren. Die Ways erscheinen immer nur einmal, in der Kacheln in der der erste Node des ways ist. Das große Problem ist im Moment die Behandlung der Relationen, da muß auf jeden Fall was abgeschnitten werden, sonst macht jemand eine Relation ist in der BRD und schon kann ich meinen cutter vergessen. :-) (Wie die Zielkachelnummern ausgerechnet werden steht übrigens auf meiner Wiki-Seite/Diskussionsseite) kann ich das ja mal reparieren. - Aber osmcut teilt ja ganz rigide an einem geometrischen Muster, ist das ueberhaupt noch zeitgemaess, ist der Mechanismus des Splitters nicht viel besser (wenn auch vielleicht noch etwas buggy)? Warum nicht mit festen Kacheln erstmal bleiben, die Elementdichte sollte noch eine Weile ausreichen, und man hat ein festes Raster um Kacheln zusammenzustellen, die Nummern bleiben ja auch immer gleich (zumindest bei mir). -- Viele Gruesse Computerteddy ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Garminkarte jetzt mit maxspeed
Hallo, Christoph Wagner schrieb: Kachelgrenzen, deswegen nehm ich erstmal nicht osmcut. Und das tool, was computerteddy benutzt scheint ja leider nicht veröffentlich zu sein. Ich bin immer noch daran den Programmierer von CutTheOsmPlanet zu überzeugen, das endlich zu veröffentlichen. Wer übrigens hindert Dich daran einfach meine geschnittenen OSM-Kacheln zu verwenden? OK, das Routing funktioniert dann erstmal nicht über die Kachelgrenzen, aber vielleicht findet sich mal jemand dem Steve zu erklären, wie meine Karten aufgebaut sind, damit er das Routing so in mkgmap einbauen kann, das dann auch mit meinen Kacheln funktioniert. -- Viele Gruesse Computerteddy ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Treppenrouting fuer Radfahrer
Am 5. Mai 2009 17:50 schrieb Sven Sommerkamp s_sommerk...@gmx.de: Bei 500m Umweg würde ich niemals die Treppen nehmen und ich denke mehr wird es sehr, sehr selten sein. Innerstädtisch magst Du recht haben, aber wenn man an Flussbrücken ausserorts denkt, kann das schon sehr schnell passieren. z.B. spaltet sich der Radfernweg R3 bei Okriftel wegen einer Treppe in zwei Alternativen auf: http://openstreetmap.org/?lat=50.0522lon=8.4958zoom=14layers=00B0FTF Grüße, jens ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Ref tag bei ways erforderlich wenn eine relation existiert?
Werner Hoch schrieb: ich frage mich ob ich an jeden einzelnen Wegabschnitt ein ref= tag anbringen muss obwohl viele Wege ja bereits über eine Relation eine ref= tag haben. Beispiel1: Jeder Weg mit Referenz: http://www.openstreetmap.org/browse/relation/3420 operator = Bodenseekreis ref = K 7735 route = road type = route Beispiel2: Wege ohne Referenz: http://www.openstreetmap.org/browse/relation/10949 operator = Bodenseekreis ref = K 7719 route = road type = route Im Beispiel2 werden die Referenzen auf den Straßen in Mapnik und TAH nicht dargestellt. Momentan ist es wohl besser, die Tags doppelt zu führen, da die Relation mit type=route und route=road von keiner mir bekannten Anwendung ausgewertet wird. Zukünftig ist es natürlich sinnvoller, gemäß Beispiel 2 zu taggen. Gruß, Stefan ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
[Talk-de] abkickende vorfahrt
Hi, gibt es eine anerkannte Loesung um abknickende vorfahrten zu taggen? Im moment wird das ja mit unserem datenmodell nicht funktionieren das das navi entweder sagt Der abbiegende Vorfahrt nach Rechts folgen oder auch nur einfach die klappe haelt wie mein festeinbau. Im moment probiere ich bei sowas immer aus der abkickenden straße eine Kurve zu modellieren so das quasi visuell die straße eine Kurve beschreibt und in der Kurve eine Straße abbiegt. Das entspricht ja aber desoefteren mal nicht der realitaet ... Eine idee die ich hatte war das mit den turn restriction relation zu verknoten um eine designated direction mit einzufuegen - allerdings waere das sprachlich ja falsch weil keine restriction im engeren Sinne. Ideen? Flo -- Florian Lohoff f...@rfc822.org +49-171-2280134 Those who would give up a little freedom to get a little security shall soon have neither - Benjamin Franklin signature.asc Description: Digital signature ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Garminkarte jetzt mit maxspeed
On 09-05-06 01:48:53 CEST, Colin Marquardt wrote: Am 5. Mai 2009 23:28 schrieb Christoph Wagner freemaps@googlemail.com: Okay okay, im Moment läuft ein erneuter Versuch. Könnte sein, dass ich die Karten auf dem Server wohl doch nicht einfach so automatisch per ftp überschreiben konnte. Hab sie erstmal von Hand gelöscht und nach ca. ner Stunde sollten sie wieder aktuell da sein. Wenn nicht, müsst ihr bis morgen früh warten. Da guck ich dann wieder rein. Es ist eine Karte dort, aber der Inhalt scheint mir genau wieder der alte Stand zu sein, wenn ich mich nicht selber vertan habe. ja, mit 342128906 bytes länge ist sie exakt genauso lang wie die vom 2009-04-24, und ein herunterladen von ein paar kilobytes bestätigt auch dass sie gleich sind. offenbar ist durch den upload per ftp nur der originale zeitstempel verloren gegangen, aber es handelt sich wieder um die alte version. irgendeine kleinigkeit muss da noch übersehen worden sein... ;-) rj ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Garminkarte jetzt mit maxspeed
On Wed, May 06, 2009 at 08:03:05AM +0200, Carsten Schwede wrote: Um die Lücken zu vermeiden muß in die Kacheln, in der ein Weg auftaucht alle nodes hinein, die der Weg benutzt. Das ist zwar manchmal richtig viel (daher meine gelegentlichen Riesenkacheln) aber nur den ersten Punkt in der Nachbarkachel zu nehmen reicht nicht. Waere es nicht eigentlich ausreichend aus dem letzten punkt in der kachel und dem ersten punkt ausserhalb der kachel einen punkt direkt auf der kachelgrenze zu berechnen und die restlichen Punkten ausserhalb der Kachel aus dem Weg zu entfernen? Flo -- Florian Lohoff f...@rfc822.org +49-171-2280134 Those who would give up a little freedom to get a little security shall soon have neither - Benjamin Franklin signature.asc Description: Digital signature ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
Original-Nachricht Datum: Wed, 6 May 2009 08:28:22 +0200 Von: Florian Lohoff f...@rfc822.org An: talk-de@openstreetmap.org Betreff: [Talk-de] abkickende vorfahrt Hi, gibt es eine anerkannte Loesung um abknickende vorfahrten zu taggen? Eigentlich kenne ich gar keine Moeglichkeit in OSM die Vorfahrt zu taggen. Bei der Auswertung gehe ich davon aus, dass die hoehere Strassenklasse Vorfahrt vor der niedrigeren hat und beruecksichtige das auch beim Taggen, was aber nicht immer konfliktfrei geht. Das wechseln von niedrig auf hoeher ist dann fuer den Router 'teurer' als der Wechseln von hoeher auf niedriger, egal ob abknickend oder nicht. Wenns die Realitaet hergibt, also die Vorfahrststrasse in der Realitaet baulich hervorgehoben ist, versuche ich das auch aehnlich wie von dir beschrieben abzubilden, sonst nicht. Gruesse Hubert -- Pt! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger01 ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Ref tag bei ways erforderlich wenn eine relation existiert?
Hallo! Stefan Dettenhofer (StefanDausR) schrieb: Werner Hoch schrieb: ich frage mich ob ich an jeden einzelnen Wegabschnitt ein ref= tag anbringen muss obwohl viele Wege ja bereits über eine Relation eine ref= tag haben. Beispiel1: Jeder Weg mit Referenz: http://www.openstreetmap.org/browse/relation/3420 operator = Bodenseekreis ref = K 7735 route = road type = route Momentan ist es wohl besser, die Tags doppelt zu führen, da die Relation mit type=route und route=road von keiner mir bekannten Anwendung ausgewertet wird. Zukünftig ist es natürlich sinnvoller, gemäß Beispiel 2 zu taggen. Auch wenn die Aussage grundsätzlich nicht falsch ist, möchte ich wie immer bei diesem Thema darauf hinweisen, daß vor einer technischen Umsetzung auch noch einige grundsätzliche logische Probleme zu klären und Festlegung zu treffen wären. Z.B. was überhaupt passieren soll, wenn der way keine ref hat, aber er gehört zu einer road-Relation, zu einer bus-Relation und zu zwei hiking-Relationen und alle 4 Relationen haben eine ref. Das mit den Relationen sieht nur auf den ersten Blick einfach aus, ist aber ziemlich knifflig. Stichwort für Informatiker: Mehrfachvererbung. bye Nop ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
Florian Lohoff schrieb: Hi, gibt es eine anerkannte Loesung um abknickende vorfahrten zu taggen? Mir nicht bekannt. Im moment wird das ja mit unserem datenmodell nicht funktionieren das das navi entweder sagt Der abbiegende Vorfahrt nach Rechts folgen oder auch nur einfach die klappe haelt wie mein festeinbau. Oder wie bei meinem noch schlimmer: manchmal meint er einen Abbiegehinweis zu geben und manchmal nicht. Da kann man sich überhaupt nicht mehr drauf verlassen. Ich fürchte, daß läßt sich algorithmisch nicht in den Griff bekommen ... Im moment probiere ich bei sowas immer aus der abkickenden straße eine Kurve zu modellieren so das quasi visuell die straße eine Kurve beschreibt und in der Kurve eine Straße abbiegt. Das entspricht ja aber desoefteren mal nicht der realitaet ... Die Straße sollte ja auf der Karte schon eher der geometrischen Realität folgen und weniger der Vorfahrtsregelung ;-) Eine idee die ich hatte war das mit den turn restriction relation zu verknoten um eine designated direction mit einzufuegen - allerdings waere das sprachlich ja falsch weil keine restriction im engeren Sinne. Ist zwar ähnlich, aber halt keine restriction. Spontane Ideen: Für einfach Fälle (wo die Vorfahrt der höchsten Straßenklasse folgt): Node mit: traffic_sign=right_of_way Für schwierige Fälle: Eine Relation (ähnlich wie turn_restrictions) mit to, via und from und type=right_of_way Gruß, ULFL ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Treppenrouting fuer Radfahrer
Morgen Martin! Das geht vielleicht hier unten im Thread etwas unter. Wenn der Vorschlag ausgereift ist, könntest Du ihn ggf. als eigenen Thread beginnen. Martin Koppenhoefer schrieb: 1. 2 Spuren für Kinderwägen (oder Fahrräder) 2. 1 Rille (aus meist Stahlprofil) für Fahrräder 3. parallele Transportbänder für Gepäck (selten, aber gibt's z.B. in Tübingen am Hauptbahnhof). Es gibt auch eine (etwas breitere) Beton- oder Steinspur am Rand, wo das Rad-Hochschieben IMHO noch bequemer ist als mit einer Stahlrille. Gesehen am Bahnhof Ratingen Ost in der Unterführung am Busbahnhof: http://www.openstreetmap.org/?mlat=51.295548mlon=6.863969zoom=14 ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Ref tag bei ways erforderlich wenn eine relation existiert?
Hallo Nop, Nop schrieb: Hallo! Stefan Dettenhofer (StefanDausR) schrieb: Werner Hoch schrieb: ich frage mich ob ich an jeden einzelnen Wegabschnitt ein ref= tag anbringen muss obwohl viele Wege ja bereits über eine Relation eine ref= tag haben. Beispiel1: Jeder Weg mit Referenz: http://www.openstreetmap.org/browse/relation/3420 operator = Bodenseekreis ref = K 7735 route = road type = route Momentan ist es wohl besser, die Tags doppelt zu führen, da die Relation mit type=route und route=road von keiner mir bekannten Anwendung ausgewertet wird. Zukünftig ist es natürlich sinnvoller, gemäß Beispiel 2 zu taggen. Auch wenn die Aussage grundsätzlich nicht falsch ist, möchte ich wie immer bei diesem Thema darauf hinweisen, daß vor einer technischen Umsetzung auch noch einige grundsätzliche logische Probleme zu klären und Festlegung zu treffen wären. Z.B. was überhaupt passieren soll, wenn der way keine ref hat, aber er gehört zu einer road-Relation, zu einer bus-Relation und zu zwei hiking-Relationen und alle 4 Relationen haben eine ref. Das mit den Relationen sieht nur auf den ersten Blick einfach aus, ist aber ziemlich knifflig. Stichwort für Informatiker: Mehrfachvererbung. eigentlich ist es doch ganz einfach (oder sehe ich da was falsch): 1) Es werden erst einmal nur die Relationen gerendert bzw. ausgewertet, also nicht mehr auf der Ebene von ways gearbeitet sondern eine Ebenen höher. 2) Dann ist das mit der Vererbung auch kein Problem, denn die Tags in der Relation stellen den Defaultwert dar, der durch einen speziellen am way überschrieben werden kann. 3) Zuletzt werden noch alle ways, die keiner road-Relation angehören, klassisch rerendert. Um bei Deinem Beispiel zu bleiben: - Mit Hilfe der road-Relation und vererbtem ref-Tag (ggf. auch name-Tag) plus aller way ohne road-Relation wird die normale Straßenkarte erzeugt. - Mit Hilfe der bus-Relation und vererbtem (anderem) ref-Tag werden die Buslinien darüber gezeichnet - Mit Hilfe der hiking-Relation und vererbtem (anderem) ref-Tag (und weiteren vererbten Tags) werden die Wanderwege generiert. Aber das ist doch jetzt im Prinzip auch schon nicht anders, oder? Gruß, Stefan P.S.: So können auch die Superrelationen behandelt werden: Erst alle Superrelationen eines typs, dann alle Relationen dieses typs, die keinen Superrelation angehören und zum Schluss alle ways, die keiner Relation angehören. ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
Original-Nachricht Datum: Wed, 06 May 2009 09:18:01 +0200 Von: Ulf Lamping ulf.lamp...@googlemail.com An: Openstreetmap allgemeines in Deutsch talk-de@openstreetmap.org Betreff: Re: [Talk-de] abkickende vorfahrt Für einfach Fälle (wo die Vorfahrt der höchsten Straßenklasse folgt): Node mit: traffic_sign=right_of_way Mein Loesungsansatz waere, die zwei Schilder fuer die beiden Richtungen an der Vorfahrtsstrasse zu erfassen und in eine relation mit der Strecke dazwischen zu geben (siehe auch mein Schilderthema oben). Vollstaendig beschrieben ist die Situation, wenn man auch noch die 'Vorfahrt achten'-Schilder der anderen Strasse(n) reinnimmt. Die Auswertung waere zwar nicht trivial, aber eindeutig loesbar. Gruesse Hubert -- Pt! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger01 ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Garminkarte jetzt mit maxspeed
Hallo, Florian Lohoff schrieb: Waere es nicht eigentlich ausreichend aus dem letzten punkt in der kachel und dem ersten punkt ausserhalb der kachel einen punkt direkt auf der kachelgrenze zu berechnen und die restlichen Punkten ausserhalb der Kachel aus dem Weg zu entfernen? Leider reicht das nicht, zumindest für Polygone, es kann dann wieder zu Fehlern kommen, wenn die Punkte so verteilt sind: *+ * ''' * ' * *+'** ' ' Die durch die ' angedeutete Kachel wird gerechnet. Die + bezeichnen die Nodes, die die ersten außerhalb der Kachel sind. Wenn man hier jetzt entweder die Kreuuzung des Polygons mit der Kachegrenze oder auch die ersten außenliegenden Punkte betrachtet, dann gibt das Endergebnis ein falsches Polygon bei dem die Ecke abgeschnitten ist. Ich denke auch, daß diese Berechnung richtig viel Zeit in Anspruch nimmt. -- Viele Gruesse Computerteddy ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
Florian Lohoff schrieb: Im moment wird das ja mit unserem datenmodell nicht funktionieren das das navi entweder sagt Der abbiegende Vorfahrt nach Rechts folgen oder auch nur einfach die klappe haelt wie mein festeinbau. Ich denke dass man das ohne spezieller turn_restriction o.ä. nur von der höheren Straßenklasse ableiten kann und ggf. bei gleicher Straßenklasse auswertet, ob sich der ref-Tag bzw. der Name ändert. Bei meinem Navi ist das auch oft so, dass es sagt: in 300m geradeaus fahren, obwohl nur links eine Straße der selben Klasse abzweigt. Gruß, Stefan ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
Hallo. Am Mittwoch 06 Mai 2009 09:39:08 schrieb Stefan Dettenhofer (StefanDausR): Ich denke dass man das ohne spezieller turn_restriction o.ä. nur von der höheren Straßenklasse ableiten kann und ggf. bei gleicher Straßenklasse auswertet, ob sich der ref-Tag bzw. der Name ändert. Bei meinem Navi ist das auch oft so, dass es sagt: in 300m geradeaus fahren, obwohl nur links eine Straße der selben Klasse abzweigt. Es wird oft genug ohne eine besondere Angabe der Vorfahrtsregelung nicht möglich sein, die Situation algorithmisch zu erkennen. Beispiel: http://www.openstreetmap.org/?lat=49.005005lon=9.502024zoom=18 Die Bundesstraße zweigt ab, in einem Winkel 90°, grade aus geht nur eine Landesstraße. Eigentlich spricht da alles für eine abknickende Vorfahrtstraße. Wie man hier sieht... http://maps.google.de/?ll=49.004951,9.501194z=18t=k ...ist das aber nicht so. Gruß, Bernd -- Am Ende sind alle Probleme der Wirtschaft Personalprobleme. - Alfred Herrhausen (dt. Bankier) signature.asc Description: This is a digitally signed message part. ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Ref tag bei ways erforderlich wenn eine relation existiert?
Hi! Stefan Dettenhofer (StefanDausR) schrieb: eigentlich ist es doch ganz einfach (oder sehe ich da was falsch): Ja. :-) Du machst einige Annahmen, für die es noch weder eine Einigung noch eine dokumentierte Festlegung gibt. 1) Es werden erst einmal nur die Relationen gerendert bzw. ausgewertet, also nicht mehr auf der Ebene von ways gearbeitet sondern eine Ebenen höher. 2) Dann ist das mit der Vererbung auch kein Problem, denn die Tags in der Relation stellen den Defaultwert dar, der durch einen speziellen am way überschrieben werden kann. 3) Zuletzt werden noch alle ways, die keiner road-Relation angehören, klassisch rerendert. Funktioniert nur bei einer einzigen Relation, nicht mehr bei mehreren. Um bei Deinem Beispiel zu bleiben: - Mit Hilfe der road-Relation und vererbtem ref-Tag (ggf. auch name-Tag) plus aller way ohne road-Relation wird die normale Straßenkarte erzeugt. - Mit Hilfe der bus-Relation und vererbtem (anderem) ref-Tag werden die Buslinien darüber gezeichnet - Mit Hilfe der hiking-Relation und vererbtem (anderem) ref-Tag (und weiteren vererbten Tags) werden die Wanderwege generiert. Es gibt derzeit keine definierte Unterscheidung von Relationen, es gibt nur einen kurzen Kommentar im Wiki daß _alle_ Tags weitervererbt werden sollen. Es gibt keine Festlegung über die Art- und Weise welche oder wie diese Relationen ausgewertet werden sollen und auch keine definierte Reihenfolge. Das Ergebnis wäre nach heutigem Stand der Definition absolut zufällig. Deine Lösung nimmt an, daß Relationen gezielt unterschieden werden können, daß es extra Layers für unterschiedliche Themen gibt und Du gehst nur auf die speziellen Relationen in meinem Beispiel ein. Die Software könnte mangels Festlegungen und Regeln die Relationen nicht unterscheiden, die sind von der derzeitigen Definition her alle gleich. Und ich kann jederzeit eine Relation mit neuem Typ eintragen und wieder alles durcheinanderbringen. Allein für Wanderrouten gibt es z.B. 3 Typen und 2 weitere in Diskussion. Wie gesagt: Der Fall mit einer Relation sieht einfach aus. Für Fälle mit mehreren, konkret angegebenen Relationen kann ein Mensch noch eine Lösung austüfteln, die aber evtl. für unterschiedliche Karten unterschiedlich aussieht. Aber davon, daß wir es so gut beschrieben haben, daß es in allen Renderern das gleiche Ergebnis liefert und daß ein Mapper irgendeinen Plan hat, was seine Einträge bedeuten werden, sind wir noch sehr weit entfernt. bye Nop ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
Bernd Wurst schrieb: Hallo. Am Mittwoch 06 Mai 2009 09:39:08 schrieb Stefan Dettenhofer (StefanDausR): Ich denke dass man das ohne spezieller turn_restriction o.ä. nur von der höheren Straßenklasse ableiten kann und ggf. bei gleicher Straßenklasse auswertet, ob sich der ref-Tag bzw. der Name ändert. Bei meinem Navi ist das auch oft so, dass es sagt: in 300m geradeaus fahren, obwohl nur links eine Straße der selben Klasse abzweigt. Es wird oft genug ohne eine besondere Angabe der Vorfahrtsregelung nicht möglich sein, die Situation algorithmisch zu erkennen. Beispiel: http://www.openstreetmap.org/?lat=49.005005lon=9.502024zoom=18 Die Bundesstraße zweigt ab, in einem Winkel 90°, grade aus geht nur eine Landesstraße. Eigentlich spricht da alles für eine abknickende Vorfahrtstraße. Wie man hier sieht... http://maps.google.de/?ll=49.004951,9.501194z=18t=k ...ist das aber nicht so. Gruß, Bernd Ja das wollte ich ja sagen: 1) Auswertung nach Straßenklasse - Normalfall 2) Spezielle relation analog der turn_restriction - Sonderfall Der Sonderfall kann dann natürlich auch eine Vorfahrtsregel sein, die geradeaus weist. Stefan ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
wir haben doch dieses maechtige werkzeug namens relationen. ich kenn mich zwar nicht wirklich damit aus aber waere folgendes nicht relativ einfach machbar: eine abknickende vorfahrtstraße gibt's eigentlich nur an einer kreuzung oder einmuendung mit drei oder mehr ways, wobei exakt zwei davon die vorfahrtstraße bilden. da sollte es doch ein einfaches zu sein, eine relation vom type priority oder so zu erstellen, die diese beiden ways enthaelt. optional kann man ja auch noch den verbindungsnode und die entsprechenden schilder mit reinnehmen. signature.asc Description: This is a digitally signed message part. ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
Original-Nachricht Datum: Wed, 6 May 2009 09:46:41 +0200 Von: Bernd Wurst be...@bwurst.org An: Openstreetmap allgemeines in Deutsch talk-de@openstreetmap.org Betreff: Re: [Talk-de] abkickende vorfahrt Es wird oft genug ohne eine besondere Angabe der Vorfahrtsregelung nicht möglich sein, die Situation algorithmisch zu erkennen. Stimmt natuerlich, darum hatte ich oben geschrieben, dass das nicht konfliktfrei geht. Selbst wenn die Vorfahrt ein herausgehobenes Kriterium bei der Klassifizierung waere, bin ich schon ueber Situationen gestolpert, bei denen die Vorfahrt so verflochten ist, dass man in eine Kreisabhaengigkeit kommt, die nicht loesbar ist. Solange man nichts anderes hat, ist die Klasse noch das beste Kriterium. Die refs sind noch problematischer. Landes-, Kreis und Gemeindestrassen gehen deutlich oefter schlecht nachvollziehbar ineinander ueber als Bundesstrassen mit den anderen Klassen. Gruesse Hubert -- Pt! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger01 ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Ref tag bei ways erforderlich wenn eine relation existiert?
Hi! Bernd Wurst schrieb: Die Mehrfachvererbung wird aber nur dann zum Problem, wenn man keine mehreren Werte haben kann. Da es in der Praxis durchaus passiert, dass eine Straße gleichzeitig zwei ref- Nummern trägt, ist das Problem also dadurch zu lösen, dass mehrere Werte einfach okay sind. Busrouten und Straßen sollten natürlich anhand des Taggings unterschieden werden, eine Straße sollte nicht durch Mehrfachvererbung plötzlich B 123 und Stadtbus 53 parallel als ref erhalten. Genau darum geht es: Wie willst Du das verhindern? Und zwar so allgmein formuliert, daß es in jedem Renderer zuverlässig umsetzbar ist? Jetzt hast schon in dem einfachen Beispiel Relationen, von denen Du mehrfache Werte zulassen würdest und andere, von denen Du keine oder nur einfache Vererbung haben willst - und keine definierte Regel, wie man das unterscheiden könnte und auch keine Möglchkeit ein die beiden oder auch eine andere-Ref zu erzeugen. Nehmen wir noch eine neue, unbekannte Relation hinzu (ich darf ja jederzeit neue Typen eintragen oder den Way weiteren Relationen zuordnen), und das Chaos ist komplett. bye Nop ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Garminkarte jetzt mit maxspeed
Am 6. Mai 2009 08:28 schrieb Robert Joop 5313501608656...@rainbow.in-berlin.de: On 09-05-06 01:48:53 CEST, Colin Marquardt wrote: Am 5. Mai 2009 23:28 schrieb Christoph Wagner freemaps@googlemail.com: Okay okay, im Moment läuft ein erneuter Versuch. Könnte sein, dass ich die Karten auf dem Server wohl doch nicht einfach so automatisch per ftp überschreiben konnte. Hab sie erstmal von Hand gelöscht und nach ca. ner Stunde sollten sie wieder aktuell da sein. Wenn nicht, müsst ihr bis morgen früh warten. Da guck ich dann wieder rein. Es ist eine Karte dort, aber der Inhalt scheint mir genau wieder der alte Stand zu sein, wenn ich mich nicht selber vertan habe. ja, mit 342128906 bytes länge ist sie exakt genauso lang wie die vom 2009-04-24, und ein herunterladen von ein paar kilobytes bestätigt auch dass sie gleich sind. offenbar ist durch den upload per ftp nur der originale zeitstempel verloren gegangen, aber es handelt sich wieder um die alte version. irgendeine kleinigkeit muss da noch übersehen worden sein... ;-) Tatsächlich! Schon wieder eine Kleinigkeit übersehen! Es ist unglaublich wieviel man eigentlich falsch machen kann bei so nem blöden Shellskript. Schaut bitte jetzt nochmal, ob sich das Problem erledigt hat. Jetzt hab ich den Fehler aber wirklich gefunden... Nochmals sorry... Grüße Christoph ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] abkickende vorfahrt
On Wed, May 06, 2009 at 09:39:08AM +0200, Stefan Dettenhofer (StefanDausR) wrote: Ich denke dass man das ohne spezieller turn_restriction o.ä. nur von der höheren Straßenklasse ableiten kann und ggf. bei gleicher Straßenklasse auswertet, ob sich der ref-Tag bzw. der Name ändert. Bei meinem Navi ist das auch oft so, dass es sagt: in 300m geradeaus fahren, obwohl nur links eine Straße der selben Klasse abzweigt. Ich habe hier mehrere Faelle wo das alles die gleiche Klasse ist - Selbst der Name aendert sich an der Stelle - Die abknickende Vorfahrt existiert da meiner Meinung nach um dem typischen Verkehrsfluss Rechnung zu tragen. Aber selbst wenn Klasse, Name und Ref gleich bleibt die Straße aber nach rechts abknickt erwarte ich vom Navi das es mir sagt Der Bundesstraße nach rechts folgen - Denn es mag durchaus Situationen geben - gerade Innerorts - wo fuer den Ortsunkundigen bei Dunkelheit nicht ersichtlich ist das die Land/Bundes/Kreisstraße abknickt. Hingegen wenn der Verkehrsfluss durch eine abknickende Vorfahrt in eine Richtung gelenkt wird, kann das Navi die Klappe halten. D.h. ein explizites taggen einer abknickenden Vorfahrt ist noetig. Da es sich um from/to/via geschichten handelt ist es ein identisches problem wie die turn restrictions. Flo -- Florian Lohoff f...@rfc822.org +49-171-2280134 Those who would give up a little freedom to get a little security shall soon have neither - Benjamin Franklin signature.asc Description: Digital signature ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de
Re: [Talk-de] Ref tag bei ways erforderlich wenn eine relation existiert?
Hallo Nop, Nop schrieb: (...) Ja. :-) Du machst einige Annahmen, für die es noch weder eine Einigung noch eine dokumentierte Festlegung gibt. (...) Du hast natürlich recht, dass das momentan so noch nicht läuft (daher hatte ich ja weiter oberen geschrieben, dass man derzeit die ref-tags besser auch noch an die ways schreibt. Wir diskutieren hier aber doch eher über die zukünftige Entwicklung! Aber auch nach momentanem Stand ist es doch so (egal ob beschrieben oder nicht): Eine Relation mit route=road und type=route wird gar nicht ausgewertet, stellt aber eine logische Zusammenfassung einer Straße dar, die aus vielen ways besteht (mit unterschiedlichen maxspeed=..., bridge=... etc.), aber die selbe ref und highway-typ hat. Also hier mein verbesserter *Vorschlag*, wie man das lösen könnte: Relation: operator = Bodenseekreis ref = K 7735 route = road type = route highway = secondary Member 1: cycleway = lane name = Meistershofener Straße Member 2: junction = roundabout oneway = yes ... Es wird auf Relationsebene gerendert, wenn dort auch der highway-typ angegeben ist. Die ways selber dürfen dann keine highway-Tag mehr enthalten und die tags werden vererbt. Ways ohne road/highway-Relation werden normal behandelt. Es läuft doch jetzt bei der Bus- und Wanderrouten genauso: Man benutzt doch für die Darstellung der Route nur die Geometrie der Wege und die Tags, die einem interessieren, der Rest wird aus der Relation genommen. Ich denke, dass langfristig ein Umdenken erfolgen muss, was die Bedeutung der Grundtypen angeht: früher gab es: node - segment - way jetzt gibt es: node - way - relation ich sehe das eigentlich so, dass ein way eher einem segment entspricht, das aus mehreren geordneten Stützpunkten (nodes) besteht und mehrere ways zu einer geordneten relation zusammengefasst werden. Bei den Multipolygonen ist das ja auch schon so. Je mehr Attribute erfasst werden (maxspeed=, ...) um so zerstückelter und kürzer werden die einzelnen ways, so dass es durchaus Sinn macht, diese Teile wieder logisch zusammenzufassen. Stefan ___ Talk-de mailing list Talk-de@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-de