remove splitter option align-for-dem again or change mkgmap?
Hi all, with current code in mkgmap this option is of no use, as mkgmap rounds the dem-dist values (down) to multiples of 16. (9942 -> 9936 , 3314 ->3312) This means that points are always interpolated, no matter where the tile boundaries are. So, I see two possibilites: 1) remove the option from splitter (it is not yet documented anyway) 2) change the code in mkgmap to allow using the exact dem-dist values without rounding I think this works fine with MapSource and Basecamp, but maybe some devices will not be able to handle that, as Garmin maps always use the rounded values. A possible benefit of the exact values is that we may be able to avoid interpolation, at least in some areas of the world. What do you think? Gerd
Hi all, sorry, seems I forget what I already found before: there is probably no way to avoid interpolation, so I think it is better to remove the option from splitter again. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Gerd Petermann <GPetermann_muenchen@hotmail.com> Gesendet: Freitag, 26. Januar 2018 08:16 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: [mkgmap-dev] remove splitter option align-for-dem again or change mkgmap? Hi all, with current code in mkgmap this option is of no use, as mkgmap rounds the dem-dist values (down) to multiples of 16. (9942 -> 9936 , 3314 ->3312) This means that points are always interpolated, no matter where the tile boundaries are. So, I see two possibilites: 1) remove the option from splitter (it is not yet documented anyway) 2) change the code in mkgmap to allow using the exact dem-dist values without rounding I think this works fine with MapSource and Basecamp, but maybe some devices will not be able to handle that, as Garmin maps always use the rounded values. A possible benefit of the exact values is that we may be able to avoid interpolation, at least in some areas of the world. What do you think? Gerd _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd, option probably won't be useful. Maybe you could change it to some kind of alternative to --resolution, where value is set in degree, but it still wouldn't look useful to me. -- Best regards, Andrzej
Hi Andrzej, I've thought about removing --resolution as well in the past, but for different reasons: I'd prefer to let splitter decide what resolution is reasonable. I also wonder if there is a good reason to store coordinates in map units instead of degrees in splitter, maybe this was only done to be able to reuse code from mkgmap, maybe it was done to make sure that mkgmap can use the bounding boxes "as is", but this cannot work since the bboxes are stored in the osm format (as bounds statements). Anyway, I think I'll remove most changes in splitter introduced with r585 and r586. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Andrzej Popowski <popej@poczta.onet.pl> Gesendet: Freitag, 26. Januar 2018 12:29 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] remove splitter option align-for-dem again or change mkgmap? Hi Gerd, option probably won't be useful. Maybe you could change it to some kind of alternative to --resolution, where value is set in degree, but it still wouldn't look useful to me. -- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd, I think resolution in bits is a clever idea. This way we can get a tile, where border of all layers is exactly the same. Maybe this is not requirement for a map, but it seems to be a neat solution. -- Best regards, Andrzej
Hi Andrzej, I don't understand. What layers do you mean? How does it help? Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Andrzej Popowski <popej@poczta.onet.pl> Gesendet: Sonntag, 28. Januar 2018 00:48 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] remove splitter option align-for-dem again or change mkgmap? Hi Gerd, I think resolution in bits is a clever idea. This way we can get a tile, where border of all layers is exactly the same. Maybe this is not requirement for a map, but it seems to be a neat solution. -- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd, I say layers, but maybe you call them levels, like in option: levels = 0:24, 1:23, 2:22, 3:20, 4:18 If splitter uses --resolution=18, then for the case above, layer/level 4 can get exactly the same bounding box as layer/level 0. If you use arbitrary resolution, then there were some rounding errors. I understand mkgmap can cope with this somehow. -- Best regards, Andrzej
Hi Andrzej, sorry, I don't get the idea. What bounding box do you mean? I assume you talk about what is called SubDivision in mkgmp? What rounding errors do you fear? Do you think that it can cause trouble when a sub division from tile A overlaps another from tile B at level 1..7 ? I think this is what usually happens now, but I am not 100% sure. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Andrzej Popowski <popej@poczta.onet.pl> Gesendet: Sonntag, 28. Januar 2018 18:19 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] remove splitter option align-for-dem again or change mkgmap? Hi Gerd, I say layers, but maybe you call them levels, like in option: levels = 0:24, 1:23, 2:22, 3:20, 4:18 If splitter uses --resolution=18, then for the case above, layer/level 4 can get exactly the same bounding box as layer/level 0. If you use arbitrary resolution, then there were some rounding errors. I understand mkgmap can cope with this somehow. -- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd,
Do you think that it can cause trouble when a sub division from tile A overlaps another from tile B at level 1..7 ?
I don't think there is a problem with overlap. I'm only stating, that splitter idea of "resolution" is a clever one, since it can lead to no overlap at all. -- Best regards, Andrzej
Hi Andrzej, okay, and this is also true when you use the default resolution 13. I see no reason to use a much higher resolution, it just seems to need more resources. My idea about automatic selection of resolution might start with res 15 and - if memory is a problem - decrease it. The search for a solution would start with resolution 12 and try higher values if no nice split is found. Just an idea, not sure if it would help in those cases where splitter fails to find a split. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von Andrzej Popowski <popej@poczta.onet.pl> Gesendet: Montag, 29. Januar 2018 16:00 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] remove splitter option align-for-dem again or change mkgmap? Hi Gerd,
Do you think that it can cause trouble when a sub division from tile A overlaps another from tile B at level 1..7 ?
I don't think there is a problem with overlap. I'm only stating, that splitter idea of "resolution" is a clever one, since it can lead to no overlap at all. -- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
participants (3)
-
Andrzej Popowski -
Gerd Petermann -
Gerd Petermann