A couple of comments, these are sort of related but not so scientific as Martin's work:
First when you validate JOSM validation catches unclosed ways, (buildings for example.) crossing ways, ie two highways that cross but don't have a node so routing software won't know you can change from one to the other and highways that end near another highway. Again if there isn't a node routing software will be unable to make the connection. I don't think any other software does these checks. highway=road is a fairly common problem its defined as a temporary highway type and was used to turn a GPS track into a highway but without any other information you don't know if its a footpath with steps or a motorway. If you're mapping from imagery in Africa I think the African highway wiki suggests unclassified or tertiary are better than highway=road. The problem with highway=road is a number of the routing software programs ignore them. This can be problematic if a bridge is so tagged with unclassified or tertiary either side. I usually check by a search for highway=road. Some oddities I've come across, first a tile invalidated for not mapping buildings when buildings were not part of the instructions. Second a tile invalidated because some buildings had been mapped previously, the note was buildings are not required, the tile itself was fine and everything requested in the instructions had been mapped. Third a tile validated by someone who had mapped one tile previously, I added in the six settlements that had been missed. The fourth was a bit odd. Projects that have lots of green tiles often attract more mappers. I was validating on one tile and brought in a bit of the next to check a highway connection and noticed an unmapped settlement. I added it in then looked at the tile. It had been validated by the project manager. I looked over several other tiles that had been validated by the same person on the same project and on more than half there were things missing. I make no comment other than it looked a little odd. If the tags are rich ie lots of different ones Maperitive has an export-tags command which outputs a .csv file. Load it up into a spreadsheet and look for spelling mistakes in the tags. Most HOT projects use a very limited set of tags so this isn't of much use but if you want to check a large town etc it can be useful. Many tags are predefined so misspelt tags happen more rarely these days. Cheerio John On 7 February 2017 at 19:34, Martin Dittus <[email protected]> wrote: > After my talk at State of the Map in Brussels, Nick Allen asked: are > newcomers more likely to be retained if we give them positive validation > feedback? I had no answer at the time, but Tyler kindly gave me permission > to look at the data to find out. The resulting findings are now under peer > review, and I will share the full research once the process has concluded. > > In the meantime, I just posted a diary entry with the key findings: > https://www.openstreetmap.org/user/dekstop/diary/40421 > > Do post a comment in case you have any questions. Or maybe you have made > your own observations that can contribute to a better understanding of the > impact of validation? > > Greetings from London, > > m. > _______________________________________________ > HOT mailing list > [email protected] > https://lists.openstreetmap.org/listinfo/hot >
_______________________________________________ HOT mailing list [email protected] https://lists.openstreetmap.org/listinfo/hot
