Personally, I think using TIGER as an example of an import gone wrong is not accurate. Knowing what we know now, things certainly could have been done better. If nothing else, waiting for TIGER 2010 would have been prudent, as the accuracy was much improved. But that wasn't something that was knowable at the time, so it makes no sense to second guess based on that.
That said, without TIGER, OSM would have been useless (and still would be!) in large swaths of the US. Thanks to the import, routing clear across the US was possible many years before it otherwise would have been. I was able to do a lot more work in Tulsa than would have been possible from GPS traces alone thanks to the import. As a newbie at the time, correcting geometry was much more doable than mapping completely from scratch. If you want an example of a bad import, GNIS is a much better example, IMO. The data was largely 30 years out of date at the time of the import. I still find myself outright deleting many of those nodes when I come across them. (On the rare occasion I have the time and motivation to map these days, anyway) By no means is this to say that any data set one comes across is appropriate for OSM or that many of the controls or improved processes put in place in the last several years are in any way a bad thing. However, I feel like some people want to throw the baby out with the bathwater. -Nathan On March 14, 2016 11:57:19 PM EDT, Tod Fitch <t...@fitchdesign.com> wrote: >Ditto to Mike’s comments. > >I’ve been dealing with the clean up of bad imports, usually TIGER but >others too, where ever I map so I think I understand where people like >Frederick are coming from. > >But I also see the reality in the U.S. of huge geographical areas with >very few OSM mappers. An all volunteer map will always be years behind >other offerings here unless we allow and even encourage carefully >importing high quality data. > >The U.S. might be unique in that there are vast quantities of excellent >geographical data that are public domain. Unfortunately there is also a >vast quantity of public domain map data of, shall we say, lesser >quality. Had the original U.S. highway import data come from the USGS >rather than the census bureau, people probably would have a very >different opinion about imports. > >At least the experience with bad imports has shown there can be issues. >And there is now a lot better understanding of how the data and import >procedures need to be vetted. So we are in a better place to do imports >and we should not shy away from importing high quality data when the >stars line up (good data, appropriate copyright, competent OSM mappers >available, documented and tested work flows, etc.). > >Tod > >> On Mar 14, 2016, at 8:36 PM, Mike Thompson <miketh...@gmail.com> >wrote: >> >> I support the careful import of high quality data whose license is >compatible with OSM. Those appears to be one of those cases. I believe >the existence of high quality data will aid in the recruitment of new >mappers and will encourage high quality contributions from those >mappers. It is much easier, and less daunting, to add additional >detail from an on-the-ground survey to some high quality data than it >is to start from scratch. People also like to be associated with >successful projects, and the more high quality data we have the more >successful we will be in the eyes of potential new mappers. >> >> Mike >> >> >> On Mon, Mar 14, 2016 at 8:46 PM, Kevin Kenny <kken...@nycap.rr.com ><mailto:kken...@nycap.rr.com>> wrote: >> Since I received only a total of three comments about this idea, one >strongly negative (from Frederik Ramm) and two only lukewarm in >support, I'm forced to conclude that this proposal has no chance of >gaining a broad community support. Consider it withdrawn. >> >> I find myself somewhat frustrated about the question of how to >recruit mappers when it appears that the map has such a paucity of data >that it will never become useful solely through the effort of volunteer >mappers. I can demonstrate the map at >http://kbk.is-a-geek.net/catskills/test3.html ><http://kbk.is-a-geek.net/catskills/test3.html>, and state that OSM is >one of many data sources that go into it, but when people go to >openstreetmap.org <http://openstreetmap.org/> and look at it, my >experience is that they lose the connection entirely between the data >that OSM has and the map that OSM enables. The huge blank area is too >intimidating for my friends, it appears! >> >> The fact that we apparently cannot use data that are not our own in >presenting our public face, together with the fact that we do not wish >to import data for which OSM will not become the authoritative source, >leaves us with an impoverished public appearance outside the cities >where streets are sparse. Perhaps this is outside OSM's ambit. It is, >after all, Open STREET Map. It seems to leave, however, very limited >pathways for citizen mappers to build on what the government has done. >Few mappers can manage to produce such a map under their own steam, and >I certainly don't have the bandwidth - either personal or network - to >support that map as a public resource out of a solo project. >> >> I'm really at a loss where to go from here. >> >> Kevin Kenny >> >> On 02/28/2016 11:42 PM, Kevin Kenny wrote: >> Oops: Just realized I originally sent this reply privately: meant to >send to the list. >> >> On 02/27/2016 05:18 PM, Frederik Ramm wrote: >> An import is great if it enables a community to go further, or forms >the >> basis of solid work in the future. An import is great if it is one >> ingredient that makes OSM the best map of the region. But it sounds >to >> me as if your proposed import is hardly more than a small time saver >for >> people who want to make maps of the Adirondack - they *could* go to >the >> original source at any time, and the likelihood of OSM hydrography >being >> *better* than the official data is very low. >> >> In my view, a good import is a catalyst for future OSM data >improvement. >> But you seem to say quite clearly that such is unlikely to happen >with >> the data you are planning to import. Your main point is that it'll >look >> better on the map, which for me isn't good enough. >> >> Can you point to areas where your import would encourage mappers, >> including yourself, to add more knowledge and surveyed data to OSM? >> My personal interest is mostly from the standpoint of improving OSM >as a resource for hikers - and recruiting citizen mappers to the task. >Available databases of hiking trail alignments are pretty poor. The >USGS maps, once stellar, have not been updated since the first Bush >administration, and keeping them up to date is no longer in the USGS's >charter. They have neither the mission nor the funding to map hiking >trails, shelters, campsites, privies, viewpoints, and similar >amenities. Mapping them falls on the shoulders of private companies >such as National Geographic, and they are happy to sell us maps - even >ones in electronic format if we are extremely fortunate - of obsolete >data of the most popular areas. The less popular areas are entirely >neglected. If trail data are to be collected, it will have to be >citizen mappers that do it, and OSM is an obvious repository for it. >And none of that data is what I propose to import. >> >> Why, then, should I import what I don't plan to improve >substantially? When I've tried to recruit my contacts in the hiking >community to mapping for OSM, when they see the state of the tiles at >openstreetmap.org <http://openstreetmap.org/>, they are put off >immediately. "Why should I bother?" they say, "there's nothing there!" >Particularly before the import of lakes and ponds was done - an import >to which your argument equally applies - this entire area simply >appeared entirely featureless, with no hope of using OSM to produce a >map that could be helpful for anyone. >> >> When, on the other hand, I show them >https://kbk.is-a-geek.net/catskills/test3.html?la=44.1232&lo=-73.9804&z=15 ><https://kbk.is-a-geek.net/catskills/test3.html?la=44.1232&lo=-73.9804&z=15> >, they see a map that's already useful for navigating the region, >although deeply flawed in many ways. I can point out that trails shown >in magenta with their names in UPPER CASE are from a State data set >that is digitized at an inappropriately large scale (and for that >reason alone, even before license concerns, I wouldn't propose >importing it). I can point out that a good many of the trail shelters, >privies, parking areas, register kiosks, viewpoints and similar >amenities are missing. I can tell hikers that they can improve OSM by >capturing that information. I can point out that if enough of us do it >as a community, we'll have up-to-date maps that we can maintain as a >community. >> >> The approach has worked for me. For instance, I was able to persuade >a contact who was hiking the route shown with the overlay in >https://kbk.is-a-geek.net/catskills/test3.html?la=44.1232&lo=-73.9804&z=15 ><https://kbk.is-a-geek.net/catskills/test3.html?la=44.1232&lo=-73.9804&z=15> >to capture GPS data and contribute it. (The uploads show my ID because >I handled conflating it, simplifying the tracks, vetting alignment >against orthophotos, and similar administrative tasks.) >> >> OSM is really the only place where the data about trails and >associated amenities can be assembled properly, as far as I can tell. >The government agencies in the US have not had the funding or authority >to collate those data in over twenty years. Web sites like >alltrails.com <http://alltrails.com/> are great for sharing your >experience with a single route, but don't really make any effort at all >to assemble a map. And the companies like National Geographic and >DeLorme are more than happy to sell our own data back to us at a >premium price, burden it with usage restrictions, and make it available >in formats that we cannot annotate and improve. >> >> I don't have a good way to address your argument that data whose >authoritiative source is not OSM should not be imported >> into OSM - and frankly, I mostly agree with it. I tend to believe >that the underlying problem is not what we choose to import or not to >import, but what we show to newcomers. I believe that the maps we >present to the public would be improved if they included (at least >optionally) layers derived from government data sources that we >taxpayers have the right to use. You can see in the maps that I've >presented that I'm also using (and do NOT propose to import) National >Land Cover Database, National Elevation Dataset, USFWS National >Wetlands Inventory, and layers from the GIS departments of several >states. I'm also using National Hydrographic Dataset - which has been >imported with some degree of success in regions other than mine. All of >these data sources fall in your hated category of "stuff that OSM >mappers can't readily maintain, for which some other source will likely >be more authoritiative." >> >> Without these external layers, what we present in the tiles is so >sparse in some areas that I, at least, find it nearly impossible to >explain the value of OSM. >> >> I chose the idea of pursuing an import because I haven't very much >hope of convincing anyone that our public face might include non-OSM >data sources. At least there is precedent for importing government data >into OSM; there is none for non-OSM-derived layers on our tiles. >> >> About the best argument that I can make about the specific data is >that the import should be "mostly harmless", because physiography in a >wilderness area is so slow to change. With the exception that >settlements, roads, railroads, farms and mines have been reclaimed by >nature, bridges have fallen, and trails have been built and abandoned, >a topographic map of the region from 1916 would be nearly as useful as >one from 2016. This is an area where "Man is a visitor who does not >remain." >> >> >> >> -- >> 73 de ke9tv/2, Kevin >> >> >> _______________________________________________ >> Talk-us mailing list >> Talk-us@openstreetmap.org <mailto:Talk-us@openstreetmap.org> >> https://lists.openstreetmap.org/listinfo/talk-us ><https://lists.openstreetmap.org/listinfo/talk-us> >> >> _______________________________________________ >> Talk-us mailing list >> Talk-us@openstreetmap.org >> https://lists.openstreetmap.org/listinfo/talk-us > > > >------------------------------------------------------------------------ > >_______________________________________________ >Talk-us mailing list >Talk-us@openstreetmap.org >https://lists.openstreetmap.org/listinfo/talk-us _______________________________________________ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
