Numbering loss in the version which came after the r3602
Hi Gerd, Since yesterday we (a group that compiles maps in Brazil) begin to see some loss of numbers in some of our maps. As we had never had this problem before, I suspected that the error could have been introduced in some recent version of mkgmap. So, I restored some old releases for testing and I found that the numbering were generated correctly until release 3602 as seen in the image below in the white-box ("390 Rua Porto Alegre"): *Snapshot taken form MapSource of a map compiled wiht mkgmap r3602* [image: Imagem inline 1] Using the same source osm file, the numbering is lost in the street "Rua Porto Alegre" (for example), if the map is compiled with any newer release of mkgmap than 3602. The snapshot bellow was taken from a map compiled with release 3616:: *Snapshot taken form MapSource of a map compiled wiht mkgmap r3616:* [image: Imagem inline 2] These compiled maps img are available for tests at the file sharing area: http://files.mkgmap.org.uk/download/270/03205200_r3602.img http://files.mkgmap.org.uk/download/271/03205200_r3616.img You can search for street "Rua Porto Alegre" to see the problem, but it problem can be found in other places too. My guess is that the problem arose when housenumber2 brahch was merged to trunk. If you need more information, please let me know. Thank you. Regards, Alexandre
Hi Alexandre,
You can search for street "Rua Porto Alegre" to see the problem
what data are you using for compilation? I have tried to find this place on OSM, I think this is actually Rua Manaus in OSM, with no house numbers: http://www.openstreetmap.org/way/93121424 -- Best regards, Andrzej
Hi Andrzej, Our group (http://www.tracksource.org.br) doesn't use OSM data. We draw our maps with other free tools and own development ones and convert the maps to osm format to compile with mkgmap. So you won't find the source data on OSM. However, I prepared a very short map containing only the essential data to simulate the problem. If you compile the attached osm file with r3602 you will see the numbering in street "Rua Porto Alegre". If you compile the same file with r3620 (for example), the street numbers for "Rua Porto Alegre" will be lost. I hope this osm file help you debug and and find the problem. Thanks. Alexandre 2015-06-17 9:11 GMT-03:00 Andrzej Popowski <popej@poczta.onet.pl>:
Hi Alexandre,
You can search for street "Rua Porto Alegre" to see the problem
what data are you using for compilation? I have tried to find this place on OSM, I think this is actually Rua Manaus in OSM, with no house numbers: http://www.openstreetmap.org/way/93121424
-- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Alexandre, I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion. -- Best regards, Andrzej
Hi Andrzej, Ok, thank you for your time and analyses. Lets wait for Gerd. regards, Alexandre 2015-06-17 13:13 GMT-03:00 Andrzej Popowski <popej@poczta.onet.pl>:
Hi Alexandre,
I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion.
-- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Alexandre, I think the problem here is that you have two ways named "RUA PORTO ALEGRE", one with id 2, the other with id 46, and both are in the same city. So far no problem, but the addr:interpolation ways on those two roads also produce a bunch of duplicate numbers. As a result, the new algo decides to ignore them. Unfortunately, the log FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 2(0) 2(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 3(0) 3(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/13 800..2, step=2 generated even interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/14 801..3, step=2 generated odd interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 419..205, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 203..3, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 418..204, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 202..2, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: found problems with interpolated numbers from addr:interpolations ways for roads with name RUA PORTO ALEGRE is not very clear about the reason and the final action, but I think the data is not okay and the algo is correct to ignore it. What do you think? Gerd Date: Wed, 17 Jun 2015 15:28:04 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602 Hi Andrzej, Ok, thank you for your time and analyses.Lets wait for Gerd. regards,Alexandre 2015-06-17 13:13 GMT-03:00 Andrzej Popowski <popej@poczta.onet.pl>: Hi Alexandre, I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion. -- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd and Steve, Thanks by your attention. I'm in vocation now and without access to my computer, so I can't provide more information if you need till my return in beginning of August. But I think that the data is correct because in despite the streets have the same name, they are different road since they aren't connect (there is a gap / a block between them). So I believe that the algo couldn't consider the name of street. But I understand your point of view, since it looks that can have a number overlapping/shadow of both streets, what would be a logical error. Unfortunately, I can make more test these days but as soon I come back I will. Thanks, Alexandre (Enviado via iPad)
Em 21/07/2015, às 06:21, Gerd Petermann <gpetermann_muenchen@hotmail.com> escreveu:
Hi Alexandre,
I think the problem here is that you have two ways named "RUA PORTO ALEGRE", one with id 2, the other with id 46, and both are in the same city. So far no problem, but the addr:interpolation ways on those two roads also produce a bunch of duplicate numbers. As a result, the new algo decides to ignore them. Unfortunately, the log FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 2(0) 2(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 3(0) 3(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/13 800..2, step=2 generated even interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/14 801..3, step=2 generated odd interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 419..205, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 203..3, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 418..204, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 202..2, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: found problems with interpolated numbers from addr:interpolations ways for roads with name RUA PORTO ALEGRE
is not very clear about the reason and the final action, but I think the data is not okay and the algo is correct to ignore it.
What do you think? Gerd
Date: Wed, 17 Jun 2015 15:28:04 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602
Hi Andrzej,
Ok, thank you for your time and analyses. Lets wait for Gerd.
regards, Alexandre
2015-06-17 13:13 GMT-03:00 Andrzej Popowski <popej@poczta.onet.pl>: Hi Alexandre,
I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion.
-- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd, I'm sorry the delay to answer, but I came on vacation and had many pending issues waiting for me. Regard this specific issue, your interpretation of the problem of duplication/overlap in the numbering interpolation is correct. In fact the data doesn't make sense and I agree with you that this case is an error in the data. To prove this, I got the short example sent before and correctly input the numbers eliminating the overlapping as shown below: [image: Imagem inline 3] And then I compiled the map with mkgmap versions 3612 and 3629 (the last one I found) and the numbers were not "missed" this time, proving you theory. *Snapshot taken form MapSource of a map compiled wiht mkgmap r3612* [image: Imagem inline 1] *Snapshot taken form MapSource of a map compiled wiht mkgmap r3629* [image: Imagem inline 2] So I think we can close the case and I have some work do clean the maps of my group. Thanks again for your attention and analysis. Best regards, Alexandre Loss 2015-07-21 9:06 GMT-03:00 Alexandre Loss <alexandre.loss@gmail.com>:
Hi Gerd and Steve,
Thanks by your attention. I'm in vocation now and without access to my computer, so I can't provide more information if you need till my return in beginning of August. But I think that the data is correct because in despite the streets have the same name, they are different road since they aren't connect (there is a gap / a block between them). So I believe that the algo couldn't consider the name of street. But I understand your point of view, since it looks that can have a number overlapping/shadow of both streets, what would be a logical error.
Unfortunately, I can make more test these days but as soon I come back I will.
Thanks,
Alexandre
(Enviado via iPad)
Em 21/07/2015, às 06:21, Gerd Petermann <gpetermann_muenchen@hotmail.com> escreveu:
Hi Alexandre,
I think the problem here is that you have two ways named "RUA PORTO ALEGRE", one with id 2, the other with id 46, and both are in the same city. So far no problem, but the addr:interpolation ways on those two roads also produce a bunch of duplicate numbers. As a result, the new algo decides to ignore them. Unfortunately, the log FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 2(0) 2(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 3(0) 3(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/13 800..2, step=2 generated even interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/14 801..3, step=2 generated odd interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 419..205, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 203..3, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 418..204, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 202..2, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: found problems with interpolated numbers from addr:interpolations ways for roads with name RUA PORTO ALEGRE
is not very clear about the reason and the final action, but I think the data is not okay and the algo is correct to ignore it.
What do you think? Gerd
------------------------------ Date: Wed, 17 Jun 2015 15:28:04 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602
Hi Andrzej,
Ok, thank you for your time and analyses. Lets wait for Gerd.
regards, Alexandre
2015-06-17 13:13 GMT-03:00 Andrzej Popowski <popej@poczta.onet.pl>:
Hi Alexandre,
I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion.
-- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Alexandre, okay, good to hear that. I decided to make mkgmap tests regarding addr:interpolation ways rather strict because in some areas (mostly Canada) these errors appear extremely often, caused by bad imports. Seems to be a generel problem of OSM generators ;-) Gerd Date: Tue, 18 Aug 2015 15:40:40 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602 Hi Gerd, I'm sorry the delay to answer, but I came on vacation and had many pending issues waiting for me.Regard this specific issue, your interpretation of the problem of duplication/overlap in the numbering interpolation is correct. In fact the data doesn't make sense and I agree with you that this case is an error in the data. To prove this, I got the short example sent before and correctly input the numbers eliminating the overlapping as shown below: And then I compiled the map with mkgmap versions 3612 and 3629 (the last one I found) and the numbers were not "missed" this time, proving you theory. Snapshot taken form MapSource of a map compiled wiht mkgmap r3612 Snapshot taken form MapSource of a map compiled wiht mkgmap r3629 So I think we can close the case and I have some work do clean the maps of my group. Thanks again for your attention and analysis. Best regards, Alexandre Loss 2015-07-21 9:06 GMT-03:00 Alexandre Loss <alexandre.loss@gmail.com>: Hi Gerd and Steve, Thanks by your attention.I'm in vocation now and without access to my computer, so I can't provide more information if you need till my return in beginning of August.But I think that the data is correct because in despite the streets have the same name, they are different road since they aren't connect (there is a gap / a block between them).So I believe that the algo couldn't consider the name of street.But I understand your point of view, since it looks that can have a number overlapping/shadow of both streets, what would be a logical error. Unfortunately, I can make more test these days but as soon I come back I will. Thanks, Alexandre (Enviado via iPad) Em 21/07/2015, às 06:21, Gerd Petermann <gpetermann_muenchen@hotmail.com> escreveu: Hi Alexandre, I think the problem here is that you have two ways named "RUA PORTO ALEGRE", one with id 2, the other with id 46, and both are in the same city. So far no problem, but the addr:interpolation ways on those two roads also produce a bunch of duplicate numbers. As a result, the new algo decides to ignore them. Unfortunately, the log FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 2(0) 2(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 3(0) 3(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/13 800..2, step=2 generated even interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/14 801..3, step=2 generated odd interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 419..205, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 203..3, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 418..204, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 202..2, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: found problems with interpolated numbers from addr:interpolations ways for roads with name RUA PORTO ALEGRE is not very clear about the reason and the final action, but I think the data is not okay and the algo is correct to ignore it. What do you think? Gerd Date: Wed, 17 Jun 2015 15:28:04 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602 Hi Andrzej, Ok, thank you for your time and analyses.Lets wait for Gerd. regards,Alexandre 2015-06-17 13:13 GMT-03:00 Andrzej Popowski <popej@poczta.onet.pl>: Hi Alexandre, I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion. -- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
So... Does this means that you fix it? It's good to hear this, because in that example I sent, in fact there was an error in data. But as soon I share this in my group, I receive a storm of examples were the data are similar that, but aren't problem in data. Alexandre (Enviado via iPad)
Em 19/08/2015, às 02:49, Gerd Petermann <gpetermann_muenchen@hotmail.com> escreveu:
Hi Alexandre,
okay, good to hear that. I decided to make mkgmap tests regarding addr:interpolation ways rather strict because in some areas (mostly Canada) these errors appear extremely often, caused by bad imports. Seems to be a generel problem of OSM generators ;-)
Gerd Date: Tue, 18 Aug 2015 15:40:40 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602
Hi Gerd,
I'm sorry the delay to answer, but I came on vacation and had many pending issues waiting for me. Regard this specific issue, your interpretation of the problem of duplication/overlap in the numbering interpolation is correct. In fact the data doesn't make sense and I agree with you that this case is an error in the data.
To prove this, I got the short example sent before and correctly input the numbers eliminating the overlapping as shown below:
And then I compiled the map with mkgmap versions 3612 and 3629 (the last one I found) and the numbers were not "missed" this time, proving you theory.
Snapshot taken form MapSource of a map compiled wiht mkgmap r3612
Snapshot taken form MapSource of a map compiled wiht mkgmap r3629 <image.png>
So I think we can close the case and I have some work do clean the maps of my group.
Thanks again for your attention and analysis.
Best regards,
Alexandre Loss
2015-07-21 9:06 GMT-03:00 Alexandre Loss <alexandre.loss@gmail.com>: Hi Gerd and Steve,
Thanks by your attention. I'm in vocation now and without access to my computer, so I can't provide more information if you need till my return in beginning of August. But I think that the data is correct because in despite the streets have the same name, they are different road since they aren't connect (there is a gap / a block between them). So I believe that the algo couldn't consider the name of street. But I understand your point of view, since it looks that can have a number overlapping/shadow of both streets, what would be a logical error.
Unfortunately, I can make more test these days but as soon I come back I will.
Thanks,
Alexandre
(Enviado via iPad)
Em 21/07/2015, às 06:21, Gerd Petermann <gpetermann_muenchen@hotmail.com> escreveu:
Hi Alexandre,
I think the problem here is that you have two ways named "RUA PORTO ALEGRE", one with id 2, the other with id 46, and both are in the same city. So far no problem, but the addr:interpolation ways on those two roads also produce a bunch of duplicate numbers. As a result, the new algo decides to ignore them. Unfortunately, the log FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 2(0) 2(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 3(0) 3(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/13 800..2, step=2 generated even interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/14 801..3, step=2 generated odd interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 419..205, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 203..3, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 418..204, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 202..2, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: found problems with interpolated numbers from addr:interpolations ways for roads with name RUA PORTO ALEGRE
is not very clear about the reason and the final action, but I think the data is not okay and the algo is correct to ignore it.
What do you think? Gerd
Date: Wed, 17 Jun 2015 15:28:04 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602
Hi Andrzej,
Ok, thank you for your time and analyses. Lets wait for Gerd.
regards, Alexandre
2015-06-17 13:13 GMT-03:00 Andrzej Popowski <popej@poczta.onet.pl>: Hi Alexandre,
I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion.
-- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Alexandre, whenever you think that mkgmap can do better please post the corresponding data. Gerd Alexandre Loss wrote
So... Does this means that you fix it? It's good to hear this, because in that example I sent, in fact there was an error in data. But as soon I share this in my group, I receive a storm of examples were the data are similar that, but aren't problem in data.
Alexandre
(Enviado via iPad)
Em 19/08/2015, às 02:49, Gerd Petermann <
gpetermann_muenchen@
> escreveu:
Hi Alexandre,
okay, good to hear that. I decided to make mkgmap tests regarding addr:interpolation ways rather strict because in some areas (mostly Canada) these errors appear extremely often, caused by bad imports. Seems to be a generel problem of OSM generators ;-)
Gerd Date: Tue, 18 Aug 2015 15:40:40 -0300 From:
alexandre.loss@
To:
mkgmap-dev@.org
CC:
Adm_Tec_Tracksource@
Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602
Hi Gerd,
I'm sorry the delay to answer, but I came on vacation and had many pending issues waiting for me. Regard this specific issue, your interpretation of the problem of duplication/overlap in the numbering interpolation is correct. In fact the data doesn't make sense and I agree with you that this case is an error in the data.
To prove this, I got the short example sent before and correctly input the numbers eliminating the overlapping as shown below:
And then I compiled the map with mkgmap versions 3612 and 3629 (the last one I found) and the numbers were not "missed" this time, proving you theory.
Snapshot taken form MapSource of a map compiled wiht mkgmap r3612
Snapshot taken form MapSource of a map compiled wiht mkgmap r3629
<image.png>
So I think we can close the case and I have some work do clean the maps of my group.
Thanks again for your attention and analysis.
Best regards,
Alexandre Loss
2015-07-21 9:06 GMT-03:00 Alexandre Loss <
alexandre.loss@
>:
Hi Gerd and Steve,
Thanks by your attention. I'm in vocation now and without access to my computer, so I can't provide more information if you need till my return in beginning of August. But I think that the data is correct because in despite the streets have the same name, they are different road since they aren't connect (there is a gap / a block between them). So I believe that the algo couldn't consider the name of street. But I understand your point of view, since it looks that can have a number overlapping/shadow of both streets, what would be a logical error.
Unfortunately, I can make more test these days but as soon I come back I will.
Thanks,
Alexandre
(Enviado via iPad)
Em 21/07/2015, às 06:21, Gerd Petermann <
gpetermann_muenchen@
> escreveu:
Hi Alexandre,
I think the problem here is that you have two ways named "RUA PORTO ALEGRE", one with id 2, the other with id 46, and both are in the same city. So far no problem, but the addr:interpolation ways on those two roads also produce a bunch of duplicate numbers. As a result, the new algo decides to ignore them. Unfortunately, the log FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 2(0) 2(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: keeping duplicate numbers assigned to different roads in cluster RUA PORTO ALEGRE 3(0) 3(1) FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/13 800..2, step=2 generated even interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/14 801..3, step=2 generated odd interpolated number(s) for id=2, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 419..205, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/65 203..3, step=2 generated odd interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 418..204, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberIvl e:\testdata\03205200-vila_velha.osm: http://www.openstreetmap.org/way/66 202..2, step=2 generated even interpolated number(s) for id=46, RUA PORTO ALEGRE FEIN: uk.me.parabola.mkgmap.osmstyle.housenumber.HousenumberGenerator e:\testdata\03205200-vila_velha.osm: found problems with interpolated numbers from addr:interpolations ways for roads with name RUA PORTO ALEGRE
is not very clear about the reason and the final action, but I think the data is not okay and the algo is correct to ignore it.
What do you think? Gerd
Date: Wed, 17 Jun 2015 15:28:04 -0300 From:
alexandre.loss@
To:
mkgmap-dev@.org
CC:
Adm_Tec_Tracksource@
Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602
Hi Andrzej,
Ok, thank you for your time and analyses. Lets wait for Gerd.
regards, Alexandre
2015-06-17 13:13 GMT-03:00 Andrzej Popowski <
popej@.onet
>:
Hi Alexandre,
I confirm, that there is no address on your map with current mkgmap. Data seems to be OK, so lets wait for Gerd opinion.
-- Best regards, Andrzej _______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
_______________________________________________ mkgmap-dev mailing list
mkgmap-dev@.org
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
image.png (47K) <http://gis.19327.n5.nabble.com/attachment/5852681/0/image.png>; image.png (36K) <http://gis.19327.n5.nabble.com/attachment/5852681/1/image.png>;
-- View this message in context: http://gis.19327.n5.nabble.com/Numbering-loss-in-the-version-which-came-afte... Sent from the Mkgmap Development mailing list archive at Nabble.com.
Hi Alexandre, if the problem is still open, please post the files again. Gerd Date: Tue, 16 Jun 2015 15:20:26 -0300 From: alexandre.loss@gmail.com To: mkgmap-dev@lists.mkgmap.org.uk CC: Adm_Tec_Tracksource@googlegroups.com Subject: [mkgmap-dev] Numbering loss in the version which came after the r3602 Hi Gerd, Since yesterday we (a group that compiles maps in Brazil) begin to see some loss of numbers in some of our maps.As we had never had this problem before, I suspected that the error could have been introduced in some recent version of mkgmap. So, I restored some old releases for testing and I found that the numbering were generated correctly until release 3602 as seen in the image below in the white-box ("390 Rua Porto Alegre"): Snapshot taken form MapSource of a map compiled wiht mkgmap r3602 Using the same source osm file, the numbering is lost in the street "Rua Porto Alegre" (for example), if the map is compiled with any newer release of mkgmap than 3602. The snapshot bellow was taken from a map compiled with release 3616:: Snapshot taken form MapSource of a map compiled wiht mkgmap r3616: These compiled maps img are available for tests at the file sharing area:http://files.mkgmap.org.uk/download/270/03205200_r3602.img http://files.mkgmap.org.uk/download/271/03205200_r3616.img You can search for street "Rua Porto Alegre" to see the problem, but it problem can be found in other places too. My guess is that the problem arose when housenumber2 brahch was merged to trunk. If you need more information, please let me know. Thank you. Regards, Alexandre _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Steve, thanks, just noticed that Alexandre posted another file that contains all that I need. Gerd
To: mkgmap-dev@lists.mkgmap.org.uk From: steve@parabola.me.uk Date: Tue, 21 Jul 2015 09:41:09 +0100 Subject: Re: [mkgmap-dev] Numbering loss in the version which came after the r3602
Hi
if the problem is still open, please post the files again.
There is no need to re-post, I've just made the files visible again.
..Steve _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
participants (5)
-
Alexandre Loss -
Andrzej Popowski -
Gerd Petermann -
GerdP
-
Steve Ratcliffe