I think it might be a mistake to suggest that we don't get to have a say in how (eg) the main osm map is rendered just because we can personally render our own. Firstly, setting up to do that rendering is not trivial, nor is reconfiguring it for a different view. But mainly because we contribute to the map to make it useful for other people.
Now, i don't see anyone sponsoring the necessary server and bandwidth capabilities (maybe i am wrong?) here in Oz. So i think we should assume that the 'other people' we want to help will be looking at osm.org or at an app that has been written by someone guided by what they see at osm.org. I agree that putting incorrect data into osm just because that makes it look right is evil. But we do have an obligation to make sure 'our' data is presented so its useful. In this case, if that means redefining terms like city, that's what we need do. We have a similar problem with connecting roads and even how frequently the name appears on a long road. Sigh... Incidentally, the documentation on how to set up a local rendering system is not great, i might do a bit of updating. And, importantly, there does not seem to be any "user guide" for osm.xml, quite a lot goes on in there ! Should we be documenting our views in the Australian page of the wiki ? David . Richard Weait <rich...@weait.com> wrote: >On Tue, Dec 18, 2012 at 11:16 PM, Nick Hocking <nick.hock...@gmail.com>wrote: > >> Richard wrote >> >> "Take a shot at creating rendering rules that fit your use case! :-)" >> >> I'm with John on this one - especially for the case of Australia. Maybe >> we need a special renderer for Australia. Just recently I managed to get >> some new mappers interested in mapping in rural NSW. >> >> They have started on the Wyndham area (which was just about completely >> unmapped, and they are having the same problem as John. I,E when they want >> to see Wyndham and it's neighbouurs (Candelo and Cathcart) at the same time >> on a map, OpenStreetMap shows almost nothing since they are too far apart. >> >> We really need a smart renderer that determines (for each tile in each >> zoom level) what are the most significant objects/ways in that tile/zoom >> and makes sure that they are rendered (even if they happen to be hundred >> mile long dirt tracks). Also if the major places in the tile are only >> localities,hamlets or villages then they should be rendered. That way we >> would not be tempted to elevate a village to a town just to make the map >> usable. >> > >Largely, I think that removing the temptation to elevate a village to a >town is an education problem, not a rendering problem. "Don't tag for the >renderer" is part of it and "look at this awesome transit map" is another >part. Back to creating a specialty rendering that is smarter about sparse >areas. > >I don't think that the smart sparse renderer is impossible. In fact, a new >feature was discussed on the mapnik list this week, transformation plugins, >that may be helpful. Transformation plugins allow you to analyze and >transform the data before rendering, so that might just be the place to >decide which place= to render at which size / logo / prominence. So if it >interests you, have a go at it. > >_______________________________________________ >Talk-au mailing list >Talk-au@openstreetmap.org >http://lists.openstreetmap.org/listinfo/talk-au
_______________________________________________ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
