Re: [Tagging] In defence of OSM Carto (was: Re: Irrigation: ditches, canals and drains)
On Fri, 31 May 2019 at 13:14, Christoph Hormann wrote: the whole waterway line with stepping across zoom levels is full of > fairly strange historic artefacts and not really well thought through. > Combined with removing minor waterways from z13 waterways are quite a > mess now. > You have my sympathies. Fixing all of that sounds very hard. However, given Andy's technique, I'm only asking you to fix a small part of it. The way ditches are handled relative to streams and rivers may not be perfect, but it's not obviously wrong (to many consumers). The way mill races and leats are treated as full-blown navigation canals (akin to rendering a ditch the same way as a river) is obviously wrong (to most data consumers). I suspect that if you ever do fix all the the waterway code to be cleaner and consistent at all zoom levels, you'd still end up with mill races being handled identically to ditches and drains. In most (technical, not tagging) ways, a mill race/leat/whatever is a subset of the class drain anyway, it's just a clarification of the purpose of the drain, in the same way that tagging its use as irrigation or flood control would be. -- Paul ___ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
Re: [Tagging] In defence of OSM Carto (was: Re: Irrigation: ditches, canals and drains)
On Fri, 31 May 2019 at 12:14, Andy Townsend wrote: > On 31/05/2019 11:26, Paul Allen wrote: > > > > Example of the horrors of using canal for a leat with current carto: > https://www.openstreetmap.org/way/609805692#map=16/52.0804/-4.6799 > At z=19 it's actually close to the true width of the leat. > > A bigger problem is the lack of granularity of rendering width at various > zoom levels (see for example > https://www.openstreetmap.org/#map=13/54.1856/-0.8334 , > https://www.openstreetmap.org/#map=14/54.1850/-0.8258 and compare with > https://map.atownsend.org.uk/maps/map/map.html#zoom=14=54.18504=-0.80956 > ). > That's the point, actually. Sub-types of canals are treated with far more importance than they deserve. Things like mill races/leats ought to disappear at lower zooms in the same way that ditches do. Not a perfect solution, and specialist carto styles may want to handle them differently, but your mapping of mill race and leat to ditch is a very good compromise. Ditches disappear at other than the highest zooms, streams disappear at even lower zooms. At the moment mill races are being treated like canals, which is as bad as if we treated ditches and streams like we do rivers. I'm aware of the discussions over on github. They're mainly focused on how to handle things if canal width is specified (as the wiki suggests), and that it's very hard and complex to make that work. The merit of your solution is that the difficult code to handle ditches is already there and the additional code to render a leat the same way as a ditch is trivial. It's clear (to me) that your solution would be acceptable to at least those people who have tagged for the renderer by tagging leats as ditches rather than canals. Even if others say it's not perfect, it's a LOT better than we have now by rendering leats the same way as navigation canals. It strikes me that it's a good interim solution pending something better, and if something better never happens I can live with that. -- Paul ___ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
Re: [Tagging] In defence of OSM Carto (was: Re: Irrigation: ditches, canals and drains)
On Friday 31 May 2019, Andy Townsend wrote: > > I suspect that the OSM Carto style would be open to pull requests > that looked at the sub-tags of canals etc. if it could be done in a > way that wasn't over-complicated - look at OSM Carto's handling of > leaf type for a possible way forward. Indeed. There is discussion on this happening in: https://github.com/gravitystorm/openstreetmap-carto/issues/3354 The important thing is to look at the data and to do it world wide and to avoid wishful thinking along the lines of "this tag looks like it could be useful to differentiate rendering so let's just assume is is actually used in the a way it would be helpful". leaf_type is easy because it represents a simple and well defined biological fact. Characterizing canals as human built structures in a similarly clear way is much harder. > A bigger problem is the lack of granularity of rendering width at > various zoom levels (see for example > https://www.openstreetmap.org/#map=13/54.1856/-0.8334 , > https://www.openstreetmap.org/#map=14/54.1850/-0.8258 and compare > with > https://map.atownsend.org.uk/maps/map/map.html#zoom=14=54.18504 >on=-0.80956 ). Yes. As mentioned in https://github.com/gravitystorm/openstreetmap-carto/issues/3354#issuecomment-496449087 the whole waterway line with stepping across zoom levels is full of fairly strange historic artefacts and not really well thought through. Combined with removing minor waterways from z13 waterways are quite a mess now. And more generally speaking creating a map style that does an equally decent job at representing all kinds of geographic settings around the world as it is the stated aim of OSM-Carto is inevitably a constant uphill battle because the vast majority of mappers and developers in OSM simply are from urban environments in Europe and North America which brings an inherent bias with it. How well OSM-Carto manages to fulfill its function to create a map for the whole OSM community to a large extent depends on how well we manage to compensate for this inherent bias. -- Christoph Hormann http://www.imagico.de/ ___ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
[Tagging] In defence of OSM Carto (was: Re: Irrigation: ditches, canals and drains)
On 31/05/2019 11:26, Paul Allen wrote: Example of the horrors of using canal for a leat with current carto: https://www.openstreetmap.org/way/609805692#map=16/52.0804/-4.6799 At z=19 it's actually close to the true width of the leat. I suspect that the OSM Carto style would be open to pull requests that looked at the sub-tags of canals etc. if it could be done in a way that wasn't over-complicated - look at OSM Carto's handling of leaf type for a possible way forward. A bigger problem is the lack of granularity of rendering width at various zoom levels (see for example https://www.openstreetmap.org/#map=13/54.1856/-0.8334 , https://www.openstreetmap.org/#map=14/54.1850/-0.8258 and compare with https://map.atownsend.org.uk/maps/map/map.html#zoom=14=54.18504=-0.80956 ). To cut the OSM Carto folks some slack, they're trying to implement a global rendering style that has to cope with (in this case) all of the different sorts of waterways everywhere on the planet. There are always going to be places where a certain feature is densely mapped and where it is important but isn't - look at the way that highway=footway becomes essentially invisible at zoom levels where it would be really useful (in rural areas) because it would overwhelm the map elsewhere (central European cities). I'm sure that they'd be open to a pull request that addressed the stream width issue above, but it'd need to be tested elsewhere on the planet - and I'm sure that there are places where the presence of a stream is "literally the most important thing on the map" at z14. I therefore wouldn't use OSM Carto as an example of "here's what you get when you tag as X". Often there's a specialist map somewhere designed to show , and that's probably the better option where it exists. Best Regards, Andy ___ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging