Re: [OSM-talk] Area mapping density gap - Was: Wikipedia article
On Saturday 26 October 2013, Florian Lohoff wrote: > > But isnt the widening gap a very natural thing to happen for a geo > database? In the end your mappers are distributed unevenly so your > pace is distributed unevenly. Not everything can be done with > armchair mapping so we as the one living in the very good mapped > areas can't help to create a complete map of very sparse mapped > areas. Different levels of completeness are natural and as i said at the beginning they will continue to exist. Having a widening range in completeness and quality however is not i think. Note i am not primarily talking about differences between areas far away from each other, like between Madagaskar and Germany. This is fully to be expected and i also don't think these differences are generally increasing. Also it would be counterproductive to try reducing this mainly through remote mapping from the distance by European mappers. I am more talking about differences at close range, take for example http://www.openstreetmap.org/#map=11/61.5554/8.4735 where one feature (the lakes) has been mapped to a high level of detail while another (the glaciers) is very crude. Again this is fully normal, whoever mapped the lakes might have been focussed on those and is not interested in the glaciers or might lack the necessary information or skills. But it seems to me there is very little communication on such matters. Partly this is a matter of having the right tools (both map notes and fixme tags are not optimal here) but it is also a matter of mapping culture i think. It bothers me when i see such things because they are strongly visible quality issues which could be solved with relatively little work. -- Christoph Hormann http://www.imagico.de/ ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Area mapping density gap - Was: Wikipedia article
Hi, On Sat, Oct 26, 2013 at 02:25:37PM +0200, Christoph Hormann wrote: > I think this is an important point - OSM does and will for the > forseeable future contain both extremely well and extremely sparsely > mapped areas ('areas' being meant here both spatially and > thematically). One of the major tasks will be to keep both the well > mapped parts up-to-date and improve the sparsely mapped parts. > > Although this is difficult to back up with numbers i have the impression > the gap between well mapped and badly mapped areas in Openstreetmap is > widening even though you would think it is much easier to improve a > badly mapped area than a well mapped one. When during use of > Openstreetmap i look at some area (because i read about it in a news > report or whatever reason) i am frequently amazed by the detailed > information i find there but i am equally often appalled by the lack of > data. One of the motivations in Wikipedia for having notability rules > certainly is to address exactly this kind of problem and to focus > efforts on those parts considered important. Openstreetmap obviously > should not follow a similar path, especially considering how it proved > damaging in Wikipedia but just attracting additional contributors is > not enough. In my opinion there is need for a more active discourse on > gaps and uniform quality of the data. But isnt the widening gap a very natural thing to happen for a geo database? In the end your mappers are distributed unevenly so your pace is distributed unevenly. Not everything can be done with armchair mapping so we as the one living in the very good mapped areas can't help to create a complete map of very sparse mapped areas. I dont see this as a problem at all. I for example have an emotional link to Madagaskar. I typically explain the value of OpenStreetMap with it. I always tell people that there is no economic value for anyone commercial to map Villages where all of the 500 People dont own a car. Although there is a road, this road will most likely never appear in any SatNav. With OpenStreetMap we dont need an economic return of invest. We dont need an economic reason to map this street. We map because we want completeness, a beautiful map, fairness and equality of all the worlds people or whatever reason. So suddenly there is a map showing how to reach these villages. And for these areas we dont need the same completeness, level of detail or the same speed of updates. Even if the bend of the street has changed, or a bridge has flushed away and is replaced with a ford the map still shows how to reach these villages. Maybe i am to optimistic but i was a very early adopter of Linux and i have now used it for nearly 20 years and nobody could imaging its success in the early '90s. OSM will be THE source for geo data in the future. It will be the most up to date, most detailed, most used data world wide. There is no way around - we simply have to be patient, wait and probably develop more and better tools for processing and editing of OSM Data. One day - probably 20 Years from now we will discuss whether we want to map the gras middle strip of the road in some unknown Village in Madagaskar and whether it'll be cut by a cow or mower. The gap will probably exists as long as we have an economical gap, so fight the G8, Globalization and probably US Aid. Flo -- Florian Lohoff f...@zz.de signature.asc Description: Digital signature ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
On Sat, Oct 26, 2013 at 10:10:19PM +0800, Eugene Alvin Villar wrote: > Two good things about Wikipedia that I hope OSM would emulate are (1) how > easy it is to see what an edit has changed in an article, and (2) how easy > it is to revert an edit—especially good for obvious vandalism. IMHO Easy tools for reverting will promote local "warlords" who will simply revert every single change somebody else does as we have right now with Wikipedia. Flo -- Florian Lohoff f...@zz.de signature.asc Description: Digital signature ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
On Fri, Oct 25, 2013 at 05:21:36PM -0400, Tom MacWright wrote: > Perhaps something to note is that, beyond technical and policy issues, one > of the more common complaints about Wikipedia is that there's an > unfriendly, elitist attitude amongst the established editors. My article > asks for some relatively deep changes to infrastructure and user > experience, but the more actionable and immediately useful thing that > everyone can do is to be friendly. I have experienced the same elitist attitude with OSM especially with notes i opened lately in my non primary areas. Questions or even remarks have been responded to in a way that as a newbie that would have been my very last note. So i think OSM goes the same path as wikipedia but i think the revert first simply doesnt work for OSM as we dont have a single click button to revert changes. So before somebody asks on a list to let somebodys changes be reverted the revert gets discussed. So i think the ability to discuss changes in OSM is its technical inability to ease the revert. Flo -- Florian Lohoff f...@zz.de signature.asc Description: Digital signature ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Mind the Gap - how to see where OSM is lacking
In the Wikipedia topic Christoph Hormann said >Although this is difficult to back up with numbers i have the impression >the gap between well mapped and badly mapped areas in Openstreetmap >is widening even though you would think it is much easier to improve a >badly mapped area than a well mapped one. This is something that has been on my mind as I compare my two main areas of mapping. My home country of Iceland (mostly detailed and currently looking at imports from the national land registry, via proper channels) and Botswana, another sparsely populated country, one sorely lacking in data. My first task was to find the 2011 Botswana Census and make sure all of the towns and villages there were on the map. Most of them weren't so I enlisted the aid of Wikipedia (the Italian one actually has the most detail!) to find their GPS co-ordinates. Then I started to draw roads and buildings in Bokspits, the southernmost village, and its surroundings. Soon I figured that this would not be the most efficient use of time, rather I should focus on at least connecting each of the town/villages to the road network. That is currently my focus and I created a MySQL table to make it easier to have an overview over which are connected to road network and which are yet to be connected. Botswana was, and still is, pretty lacking in details outside of its largest cities. So it is a badly mapped area. Iceland is a well mapped area but still lacking details, so I created a taskboard in Trello to give us a to do list. This list shows on the leftmost column those town villages lacking imagery (Vantar loftmyndir), lacking buildings (Vantar byggingar), in progress (Í vinnslu) and finally completed (Búið), meaning roads and buildings have been mapped. URL: https://trello.com/b/dn0f5v5p I started to set up a similar list for Botswana but soon found that 498 towns spread over 7 such task boards was very unwieldy! So the Icelandic OSM community, which formed Hliðskjálf, a society for free and open GIS data, last Thursday, decided to start to work on some sort of Quality Inspector. Currently all discussions about it are only available in Icelandic but once our ideas are better formed we will switch to English for wider discussion. The basic idea is to be able to create automatically an overview of a locality within a region, listing for example number of emergency services within it, education facilities, roads, buildings etc. Then human input gives ratings for various things like how much of the road network is done, how many buildings, cycling and pedestrian network etc. It will probably require substantial work to set it up but once it is up and running it should make it easier to notice sore thumbs, areas completely or mostly lacking in data, wether on a global scale or a more local scale. Another idea is to run differentials on it every month or so, making it easier to notice if a place is suddenly shooting up in services or buildings or whatever metric, meaning a mapper is working on it, one we can perhaps support, and if needed, gently and tactfully help him/her improve his/her work. We plan to run Iceland and Botswana into it for initial testing and hopefully, if it proves to work, to open it up for others to import their own areas. As said, currently all we have on it is in Icelandic and it is still on the drawing board but we are working on prototyping and converting to English. best wishes, Jóhannes ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
On Sat, Oct 26, 2013 at 10:53 PM, Jason Remillard wrote: > Eugene - Obviously, I think it is OK right now that it is hard to diff > and revert changes. We are not under assault by spammers. > But you do agree that it's something that needs to be improved eventually? It's not spammers alone that are the problem. For example, I think the criticism about iD showing a prominent trash icon for deleting objects would be lessened if people have an easy way of reverting such mistakes. > However, check this link out. It shows that Wikipedia has about > 36,000+ active editors (90 day average) > > > http://www-users.cs.umn.edu/~halfak/publications/The_Rise_and_Decline/images/decline.png > > OSM, currently have about 18,000 active editors (30 day average) > > http://osmstats.altogetherlost.com/ > > We know that 80% of the edits are done by the active editors. Using > this important metric, we are about half the size of Wikipedia, which > is amazing. > The 36,000 number is only for the English Wikipedia. If you get the edits for all Wikipedia languages, which makes the number more comparable to OSM, the number of active Wikipedians is around 71,000: http://stats.wikimedia.org/EN/TablesWikipediansEditsGt5.htm In addition, in Wikipedia, an active editor is defined as one who has edited at least 5 times in a calendar month. The OSMstats page you linked seems to count a user as active when he or she has contributed at least once (which seems correct when looking at the stats for my country). So, if we were to use the same definition of "active user", I'm sure the difference in counts would be even larger than the 71,000 vs. 18,000. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
Hi, Simon - If you look at the slashdot and hackernews links, I think you will see that many of the people that are upset probably don't have an ideological ax to grind. Eugene - Obviously, I think it is OK right now that it is hard to diff and revert changes. We are not under assault by spammers. One last thought. It is interesting to study Wikipedia because the project is so successful. It is a top 10 web site, everybody knows and uses it, they have a well funded foundation, etc, etc. Hardly anybody knows about OSM, and our registered user count is quite small compared to Wikipedia. However, check this link out. It shows that Wikipedia has about 36,000+ active editors (90 day average) http://www-users.cs.umn.edu/~halfak/publications/The_Rise_and_Decline/images/decline.png OSM, currently have about 18,000 active editors (30 day average) http://osmstats.altogetherlost.com/ We know that 80% of the edits are done by the active editors. Using this important metric, we are about half the size of Wikipedia, which is amazing. Jason On Sat, Oct 26, 2013 at 12:11 AM, Jason Remillard wrote: > Hi Tom > > Your blog post is very interesting. > > Just in case anybody thinks that the rapid growth of OSM is inevitable > at this point, this study shows how Wikipedia turned off its growth > like a switch when they starting clamping down on first time editors. > Since 2007 the number of active editors has actually decreased. > > http://www-users.cs.umn.edu/~halfak/publications/The_Rise_and_Decline/ > > Unless the map in your area is 100% perfect and complete, be extra > nice to those new editors! > > Jason > > > > > > > > On Fri, Oct 25, 2013 at 5:21 PM, Tom MacWright wrote: >> I wrote an article somewhat in the same vein: >> >>> http://macwright.org/2013/10/15/point-and-shoot.html >> >> Perhaps something to note is that, beyond technical and policy issues, one >> of the more common complaints about Wikipedia is that there's an unfriendly, >> elitist attitude amongst the established editors. My article asks for some >> relatively deep changes to infrastructure and user experience, but the more >> actionable and immediately useful thing that everyone can do is to be >> friendly. >> >> >> On Fri, Oct 25, 2013 at 5:06 PM, Jason Remillard >> wrote: >>> >>> Hi, >>> >>> The MIT technology review just published this article on Wikipedia. >>> >>> >>> http://www.technologyreview.com/featuredstory/520446/the-decline-of-wikipedia/ >>> >>> It is sport criticizing Wikipedia, but two things stuck out. >>> >>> Wikipedia is trying to get more editors. However, they seem to have >>> some additional problems that OSM does not have. >>> >>> Wikipedia failed to roll out the new GUI article editor. >>> >>> If you read the discussion on hacker news, and Slashdot. >>> >>> >>> http://news.slashdot.org/story/13/10/23/1643228/wikipedias-participation-problem >>> https://news.ycombinator.com/item?id=6612638 >>> >>> It seems like Wikipedia has revert first policy on questionable edits. >>> It makes it unpleasant to start with the project, since probably every >>> bodies first edits are questionable. >>> >>> OSM policy/culture of discussing a change *before* reverting is really >>> good thing. >>> >>> Jason >>> >>> ___ >>> talk mailing list >>> talk@openstreetmap.org >>> https://lists.openstreetmap.org/listinfo/talk >> >> ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
On Sat, Oct 26, 2013 at 5:06 AM, Jason Remillard wrote: > It seems like Wikipedia has revert first policy on questionable edits. > It makes it unpleasant to start with the project, since probably every > bodies first edits are questionable. > > OSM policy/culture of discussing a change *before* reverting is really > good thing. > Two good things about Wikipedia that I hope OSM would emulate are (1) how easy it is to see what an edit has changed in an article, and (2) how easy it is to revert an edit—especially good for obvious vandalism. In OSM, trying to figure out what exactly happened in all but the most simple changesets is quite hard. Changesets pages only show what objects were added/modified/deleted but we have no good "diff" tool unlike in Wikipedia. (Granted, diff-ing text is a well-known problem with lots of solutions; diff-ing geodata is relatively new.) The OSM History Viewer is, I think, the best tool we have for analyzing changesets, but it still lacks important features (for example, it can show you objects that have been deleted on a map but it doesn't tell you what those objects are and what tags they had). While we have tools for reverting changesets, they are not as easy to use as with Wikipedia and complex changes sometimes need to be referred to the DWG. In addition, it's hard to partially revert a changeset—reverting only the problematic objects and leaving the rest untouched (or improved). ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
On Saturday 26 October 2013, Simon Poole wrote: > > But then on the other hand it is a > fairly mature project and the easy stuff simply has been done, we > probably can show similar trends in extremely well mapped areas. I think this is an important point - OSM does and will for the forseeable future contain both extremely well and extremely sparsely mapped areas ('areas' being meant here both spatially and thematically). One of the major tasks will be to keep both the well mapped parts up-to-date and improve the sparsely mapped parts. Although this is difficult to back up with numbers i have the impression the gap between well mapped and badly mapped areas in Openstreetmap is widening even though you would think it is much easier to improve a badly mapped area than a well mapped one. When during use of Openstreetmap i look at some area (because i read about it in a news report or whatever reason) i am frequently amazed by the detailed information i find there but i am equally often appalled by the lack of data. One of the motivations in Wikipedia for having notability rules certainly is to address exactly this kind of problem and to focus efforts on those parts considered important. Openstreetmap obviously should not follow a similar path, especially considering how it proved damaging in Wikipedia but just attracting additional contributors is not enough. In my opinion there is need for a more active discourse on gaps and uniform quality of the data. Another important difference between Wikipedia and Openstreetmap is that OSM does not have a no-original-research-rule. In fact original research both in-the-field and from the armchair are preferred in comparison to second hand information (a.k.a. imports). This makes OSM potentially much more suited for professional contributors who in Wikipedia always risk being accused of lacking neutrality. There are however other barriers that discourage such people to become active contributors. -- Christoph Hormann http://www.imagico.de/ ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
iD is a leap forward for getting more people to contribute. Technical people I try to get involved do ask "so anyone can delete anything" with some incredulity but as Scroobius Pip says "some people are just nice" and so far the ratio I've seen in OSM is that 99,999% are. A troll passes by now and then but that is usually easily corrected and quickly due to obvious deletions or changes. What also hurt the English Wikipedia (which is by far the biggest and what people usually refer to when saying Wikipedia) was the notability "clamp-down". Deletionists had a field day in deeming locally important or well known things as non-notable and promptly deleted. This included football clubs in lower divisions who had played for decades or a century and even some villages or other localities "fell foul" of the global notibility that the deletionist movement, who must be thinking digital space is limited, demanded. I myself was an active contributor on the Icelandic and English Wikipedias. I am an admin on the Icelandic one (tiny but focuses more on local matters, has a niche and thrives in it, no sense in trying to emulate the scientific coverage the English one has) but have long since stopped trying to do anything beyond mere obvious corrections on the English one, the red tape there driving not only new editors but also experienced editors away. A couple of weeks ago I deleted boilerplates (another red-tape excess the English wikipedia has indulged in, slapping on the front-page comments that should belong on talk pages) from several Botswanan villages where they were under the threat of deletion due to being non-notable. Something that the notability guidelines themselves frown upon (a village being notable in it self is the rule) but nothing that has stopped the deletionism movement. Personally I try and keep an eye on mappers working in "my areas" which are Iceland and Botswana, and add them as friend on OSM and send them messages if they have done something superb or try to inform them of appropriate OSM-wiki pages if I notice something odd being done. Here I am fortunate, so to speak, as in the number of active editors in these regions is so far not very high. I dream of the day when the number grows though! --Jóhannes Þann 26.10.2013 04:11, skrifaði Jason Remillard: Hi Tom Your blog post is very interesting. Just in case anybody thinks that the rapid growth of OSM is inevitable at this point, this study shows how Wikipedia turned off its growth like a switch when they starting clamping down on first time editors. Since 2007 the number of active editors has actually decreased. http://www-users.cs.umn.edu/~halfak/publications/The_Rise_and_Decline/ Unless the map in your area is 100% perfect and complete, be extra nice to those new editors! Jason On Fri, Oct 25, 2013 at 5:21 PM, Tom MacWright wrote: I wrote an article somewhat in the same vein: http://macwright.org/2013/10/15/point-and-shoot.html Perhaps something to note is that, beyond technical and policy issues, one of the more common complaints about Wikipedia is that there's an unfriendly, elitist attitude amongst the established editors. My article asks for some relatively deep changes to infrastructure and user experience, but the more actionable and immediately useful thing that everyone can do is to be friendly. On Fri, Oct 25, 2013 at 5:06 PM, Jason Remillard wrote: Hi, The MIT technology review just published this article on Wikipedia. http://www.technologyreview.com/featuredstory/520446/the-decline-of-wikipedia/ It is sport criticizing Wikipedia, but two things stuck out. Wikipedia is trying to get more editors. However, they seem to have some additional problems that OSM does not have. Wikipedia failed to roll out the new GUI article editor. If you read the discussion on hacker news, and Slashdot. http://news.slashdot.org/story/13/10/23/1643228/wikipedias-participation-problem https://news.ycombinator.com/item?id=6612638 It seems like Wikipedia has revert first policy on questionable edits. It makes it unpleasant to start with the project, since probably every bodies first edits are questionable. OSM policy/culture of discussing a change *before* reverting is really good thing. Jason ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Wikipedia article
I think you are jumping to conclusions (just as the TR piece does in a big way). Sure in the Web 2.0 (isn't that really dated by now btw?) world any project that doesn't show exponential growth and the potential to have participant numbers in the billions is not really interesting and the fact that you can single out a number in Wikipedia that is actually declining with time, guarantees damnation. In reality Wikipedia is and continues to be a huge success, and there are aspects of that success that we would be happy if we could emulate them. Sure they have challenges and the TR article does touch on some of them a bit. Attracting contributors with knowledge outside of the mainstream is clearly one and that their barrier to entry is now rather high (editor, complexity of the expected article structure) is not really a secret. But then on the other hand it is a fairly mature project and the easy stuff simply has been done, we probably can show similar trends in extremely well mapped areas. I would question if Wikipedia really has a general issue with being nice to new editors (outside of turf wars that we have had in OSM too), a lot of the complaints seem to originate from fringe groups (creationists etc.) that thrive in the US of A, but are of little or no consequence outside. Luckily for us, our idealogical fights tend to be about cycleway tagging and tend not to get as much media coverage :-). Our main challenge is simply covering area and detail, there is no difference between a street name entered by somebody with a PhD in social sciences and one added by a 1st grader. We don't need anything outside of knowledge of your surroundings to contribute in the first place, and to become a regular contributor, it is really only necessary to have a certain love to detail and enough interest to dedicate a significant amount of time to OSM. The later is clearly the largest barrier to contributing to OSM (http://www.slideshare.net/manuelaschmidt1/poster-dresden-icc) and while we may be able to motivate more and more diverse groups to contribute, we shouldn't expect that limiter to go away. Simon Am 26.10.2013 06:11, schrieb Jason Remillard: > Hi Tom > > Your blog post is very interesting. > > Just in case anybody thinks that the rapid growth of OSM is inevitable > at this point, this study shows how Wikipedia turned off its growth > like a switch when they starting clamping down on first time editors. > Since 2007 the number of active editors has actually decreased. > > http://www-users.cs.umn.edu/~halfak/publications/The_Rise_and_Decline/ > > Unless the map in your area is 100% perfect and complete, be extra > nice to those new editors! > > Jason > > > > > > > > On Fri, Oct 25, 2013 at 5:21 PM, Tom MacWright wrote: >> I wrote an article somewhat in the same vein: >> >>> http://macwright.org/2013/10/15/point-and-shoot.html >> Perhaps something to note is that, beyond technical and policy issues, one >> of the more common complaints about Wikipedia is that there's an unfriendly, >> elitist attitude amongst the established editors. My article asks for some >> relatively deep changes to infrastructure and user experience, but the more >> actionable and immediately useful thing that everyone can do is to be >> friendly. >> >> >> On Fri, Oct 25, 2013 at 5:06 PM, Jason Remillard >> wrote: >>> Hi, >>> >>> The MIT technology review just published this article on Wikipedia. >>> >>> >>> http://www.technologyreview.com/featuredstory/520446/the-decline-of-wikipedia/ >>> >>> It is sport criticizing Wikipedia, but two things stuck out. >>> >>> Wikipedia is trying to get more editors. However, they seem to have >>> some additional problems that OSM does not have. >>> >>> Wikipedia failed to roll out the new GUI article editor. >>> >>> If you read the discussion on hacker news, and Slashdot. >>> >>> >>> http://news.slashdot.org/story/13/10/23/1643228/wikipedias-participation-problem >>> https://news.ycombinator.com/item?id=6612638 >>> >>> It seems like Wikipedia has revert first policy on questionable edits. >>> It makes it unpleasant to start with the project, since probably every >>> bodies first edits are questionable. >>> >>> OSM policy/culture of discussing a change *before* reverting is really >>> good thing. >>> >>> Jason >>> >>> ___ >>> talk mailing list >>> talk@openstreetmap.org >>> https://lists.openstreetmap.org/listinfo/talk >> > ___ > talk mailing list > talk@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk