Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
On Sun, May 24, 2020 at 19:37 Paul Norman via dev wrote: ... out. Mapbox GL styles are written in JSON and most of the tools > overwrite any formatting. This means there's no way to add comments to Have you investigated HJSON? See: https://hjson.github.io/ The Raku language (https://raku.org) has a module that handles it. Blessings, -Tom ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
On Tuesday 26 May 2020, nd...@redhazel.co.uk wrote: > > Rasterization is a data generalization technique too. No it isn't. Generalization is the most frequently abused term in cartography. Calling ST_Simplify and way_area filtering on OSM landcover polygon data generalization is a bit like calling JPEG compression noise reduction. But we are drifting off topic here. > Yes, in some > cases it leaves no artifacts at all Who would say that? In fact sampling artefacts (a.k.a. aliasing) are the mother of all artefacts in computer graphics. The point is that sampling artefacts are very well understood and it is known how to minimize them. > I've been looking into the code and I couldn't find anything related > to polygon generalization. Where is it done? Lossy geometry compression is happening in ST_AsMVTGeom - which is not just the inevitable coordinate rounding to the vector tile coordinate grid but also the subjective choice of per feature Douglas-Peucker line compression. See https://github.com/postgis/postgis/pull/463 for the well argued suggestion not to do that. The way_area filtering is explicit in the SQL code. -- Christoph Hormann http://www.imagico.de/ ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
On 26/05/2020 13:26, Christoph Hormann wrote: As Joseph has hinted when it comes to accurate rendering for precise mapper feedback without unintentional incentives at the lower zoom levels there are other options based on pre-rasterizing the data which would still allow using the benefits of client side styling. I have demonstrated that in the past. Rasterization is a data generalization technique too. Yes, in some cases it leaves no artifacts at all but that's not true in a general case and almost never true on mobile platforms. It is a viable solution but to make it work well the style should not use borders or bake in antialiasing (i.e. leave that to the client code). I am concerned about something else - even with polygon generalization in place, the performance is not great. Zooming/panning is rather slow on my not so weak PC - the map would be unusable on a mobile device or at best it would be draining the battery quickly. I've been looking into the code and I couldn't find anything related to polygon generalization. Where is it done? If it happens automatically in the Mapbox GL renderer that's too late - all the cost of preparing tiles, transmitting them, processing them in the client has already been incurred. Displaying data via WebGL is probably the cheapest part of the processing chain. ndrw ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
On Tuesday 26 May 2020, Tom Hughes via dev wrote: > More specifically what you are seeing is that it will let you zoom in > beyond zoom 8 but you are seeing data that was simplified on the > basis that it would be shown at zoom 8 where those artefacts would > not be visible. As said that is not quite true - but of course attributing artefacts, noise and rendering distortions that are visible to specific data processing steps is not easy. As Joseph has hinted when it comes to accurate rendering for precise mapper feedback without unintentional incentives at the lower zoom levels there are other options based on pre-rasterizing the data which would still allow using the benefits of client side styling. I have demonstrated that in the past. -- Christoph Hormann http://www.imagico.de/ ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
On 2020-05-26 10:59, Simon Poole wrote: I think you are missing "Only zoom 0 to 8 has been implemented so far. I started at zoom 0 and am working my way down." Oh zoom 8 is way further out than I was, my bad. The simplifications are not visible at that zoomlevel. Maarten Simon Am 26.05.2020 um 10:42 schrieb Maarten Deen: On 2020-05-26 10:15, Paul Norman via dev wrote: On 2020-05-25 1:15 a.m., Maarten Deen wrote: Still, it looks to be a very simplified subset of the complete data. So I'd be interested to see how this works on a full dataset. No, it has the full data except for admin boundaries and bay/straight names on zoom 7 and 8. When you compare it to OpenStreetMap Carto they're pretty similar. There are some labeling differences with sizes of labels and when labels come in, but those don't have anything to do with performance or size. No, that's not what I'm seeing. I am seeing very basic and rectangular landuses, obviously simplified, along with waterways, railways, motorways, trunk, primary secondary roads and they too have been simplified, to the extent that parallel ways on motorways are now crossing eachother. This is in my area: https://pnorman.dev.openstreetmap.org/cartographic/mapbox-gl.html#12.14/51.36196/6.08322 https://imgur.com/a/WcYlVUc Regards, Maarten ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
More specifically what you are seeing is that it will let you zoom in beyond zoom 8 but you are seeing data that was simplified on the basis that it would be shown at zoom 8 where those artefacts would not be visible. Tom On 26/05/2020 09:59, Simon Poole wrote: I think you are missing "Only zoom 0 to 8 has been implemented so far. I started at zoom 0 and am working my way down." Simon Am 26.05.2020 um 10:42 schrieb Maarten Deen: On 2020-05-26 10:15, Paul Norman via dev wrote: On 2020-05-25 1:15 a.m., Maarten Deen wrote: Still, it looks to be a very simplified subset of the complete data. So I'd be interested to see how this works on a full dataset. No, it has the full data except for admin boundaries and bay/straight names on zoom 7 and 8. When you compare it to OpenStreetMap Carto they're pretty similar. There are some labeling differences with sizes of labels and when labels come in, but those don't have anything to do with performance or size. No, that's not what I'm seeing. I am seeing very basic and rectangular landuses, obviously simplified, along with waterways, railways, motorways, trunk, primary secondary roads and they too have been simplified, to the extent that parallel ways on motorways are now crossing eachother. This is in my area: https://pnorman.dev.openstreetmap.org/cartographic/mapbox-gl.html#12.14/51.36196/6.08322 https://imgur.com/a/WcYlVUc Regards, Maarten ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev -- Tom Hughes (t...@compton.nu) http://compton.nu/ ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
I think you are missing "Only zoom 0 to 8 has been implemented so far. I started at zoom 0 and am working my way down." Simon Am 26.05.2020 um 10:42 schrieb Maarten Deen: > On 2020-05-26 10:15, Paul Norman via dev wrote: >> On 2020-05-25 1:15 a.m., Maarten Deen wrote: >>> Still, it looks to be a very simplified subset of the complete data. >>> So I'd be interested to see how this works on a full dataset. >> >> No, it has the full data except for admin boundaries and bay/straight >> names on zoom 7 and 8. When you compare it to OpenStreetMap Carto >> they're pretty similar. There are some labeling differences with sizes >> of labels and when labels come in, but those don't have anything to do >> with performance or size. > > No, that's not what I'm seeing. I am seeing very basic and rectangular > landuses, obviously simplified, along with waterways, railways, > motorways, trunk, primary secondary roads and they too have been > simplified, to the extent that parallel ways on motorways are now > crossing eachother. > > This is in my area: > https://pnorman.dev.openstreetmap.org/cartographic/mapbox-gl.html#12.14/51.36196/6.08322 > https://imgur.com/a/WcYlVUc > > Regards, > Maarten > > ___ > dev mailing list > dev@openstreetmap.org > https://lists.openstreetmap.org/listinfo/dev signature.asc Description: OpenPGP digital signature ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
On 2020-05-26 10:15, Paul Norman via dev wrote: On 2020-05-25 1:15 a.m., Maarten Deen wrote: Still, it looks to be a very simplified subset of the complete data. So I'd be interested to see how this works on a full dataset. No, it has the full data except for admin boundaries and bay/straight names on zoom 7 and 8. When you compare it to OpenStreetMap Carto they're pretty similar. There are some labeling differences with sizes of labels and when labels come in, but those don't have anything to do with performance or size. No, that's not what I'm seeing. I am seeing very basic and rectangular landuses, obviously simplified, along with waterways, railways, motorways, trunk, primary secondary roads and they too have been simplified, to the extent that parallel ways on motorways are now crossing eachother. This is in my area: https://pnorman.dev.openstreetmap.org/cartographic/mapbox-gl.html#12.14/51.36196/6.08322 https://imgur.com/a/WcYlVUc Regards, Maarten ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] OpenStreetMap Cartographic: A client-side rendered OpenStreetMap Carto
On 2020-05-25 1:15 a.m., Maarten Deen wrote: Still, it looks to be a very simplified subset of the complete data. So I'd be interested to see how this works on a full dataset. No, it has the full data except for admin boundaries and bay/straight names on zoom 7 and 8. When you compare it to OpenStreetMap Carto they're pretty similar. There are some labeling differences with sizes of labels and when labels come in, but those don't have anything to do with performance or size. ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev