Re: [talk-au] navit and bays/coastline around Sydney
I think putting the coastline (natural=coastline) across the mouth of a river (or a bay) is fine. This is quite common. - Ben. On 27 April 2011 19:16, Andrew Harvey wrote: > > Yep, and I added the coastline=imaginary tag to indicate that the line > tagged as coastline which crosses the mouth of the river is not a real > coastline, rather it is there to help existing software that doesn't > like breaks in the coastline. > > ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit and bays/coastline around Sydney
On Tue, Apr 26, 2011 at 10:40 PM, Franc Carter wrote: >Yep - I will look in to that but from looking in to the code it looks >hard(ish) as I don't see a generic approach for dealing with ways that >are concatenated to form closed way. Navit would then have problems rendering any multipolygon relations made up of multiple segments. Though not all the area bays are like this some are just regular closed ways at the moment. > Yes, there is an examples of this across Port Hacking with a coastline > tag runnnig across open water. Conceptually wrong I think, but needed > in practice fro rendering to have any chance. Yep, and I added the coastline=imaginary tag to indicate that the line tagged as coastline which crosses the mouth of the river is not a real coastline, rather it is there to help existing software that doesn't like breaks in the coastline. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit and bays/coastline around Sydney
[snip] > > I've been aiming to tag bays as areas rather than just a node in the > centre. As a consequence my initial thought was to tag the area as > natural=bay. Traditionally most renderers didn't render this as water. > the OSM Mapnik style now does, but many others still don't. The > problem was I couldn't tag as both natural=water to get the rendering > and natural=bay to indicate the type of feature. Yes agreed, tagging the bay area seems like the right sort of thing to do. I've had the problem with mulitvalued tags before - the ';' approach seems rather hacky to me so I can understand why went for bay. > > I've since realised that tagging as natural=water, water=bay could be > a solution to use, but as natural=bay already had widespread use with > nodes, I wanted to keep consistency. > > As Markus_g mentioned you could try to edit Navits stylesheets so it > renders natural=bay areas as water (some times as a multipolygon), > although... Yep - I will look in to that but from looking in to the code it looks hard(ish) as I don't see a generic approach for dealing with ways that are concatenated to form closed way. > > I'm not opposed to your suggestion of keeping a kind of coastline tag, > I'm just not sure the best way to implement it, please discuss it if > you like. > > * Perhaps the coastline tag should be reserved for the ocean facing > coast. If we want to tag anything say inside a bay or harbour maybe we > could use a shoreline tag. Yes, this where the definitions aren't entirely clear to me. At some points we can say it's clearly coastline, at some points it's clearly riverbank, but the transition is not so clear. I suspect that the ways that make up a bay should be tagged with something, they mark a transition between water and not-water and the bay is the area inside these. The relation does this nicely. > * A problem with tagging the area of a bay is that while the shoreline > is mostly well defined, the other edge is fuzzy. A possible solution > is to use a multi polygon relation to tag just the non-shoreline > segments of the bay outline as fuzzy=*. I suppose that using a > multipolygon relation you can keep jest the shoreline segments > together in another relation for say a larger harbour or river Yes, there is an examples of this across Port Hacking with a coastline tag runnnig across open water. Conceptually wrong I think, but needed in practice fro rendering to have any chance. Not that it helps, but it occurs to me that the rendering model may be inverted. If the default was everywhere is water and then we have closed areas of land then things may go smoother - but tool late now ;-) Because of the way coastline rendering works I can't see a way to make things render correctly without having at least one hack (i.e the fake coastline boundary). The least hacky approach that I can see if for the water/non-water transition to be tagged coastline or riverbank (and shoreline if we can define it sensibly). At the point of transition there will need to be some tagging-for-the-rendere, but that's unavoidable. cheers > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit and bays/coastline around Sydney
On Tue, Apr 26, 2011 at 8:28 PM, Franc Carter wrote: > Hi, > > I remember there was a discussion a while ago tagging of bays - which > I didn;t pay much attention too ;-( > > They now seem to be in a relation for each Bay, which is nice for > defining the area of the bay. However there are no tags on the ways > themselves and as a result Navit does nto render them as water, which > makes the map on Navit significantly less easy to use. > > My thought would be to put the coastline tag back on the ways and keep > the relations - any thoughts or suggestions on how to recover Navit > goodness ? I've been aiming to tag bays as areas rather than just a node in the centre. As a consequence my initial thought was to tag the area as natural=bay. Traditionally most renderers didn't render this as water. the OSM Mapnik style now does, but many others still don't. The problem was I couldn't tag as both natural=water to get the rendering and natural=bay to indicate the type of feature. I've since realised that tagging as natural=water, water=bay could be a solution to use, but as natural=bay already had widespread use with nodes, I wanted to keep consistency. As Markus_g mentioned you could try to edit Navits stylesheets so it renders natural=bay areas as water (some times as a multipolygon), although... I'm not opposed to your suggestion of keeping a kind of coastline tag, I'm just not sure the best way to implement it, please discuss it if you like. * Perhaps the coastline tag should be reserved for the ocean facing coast. If we want to tag anything say inside a bay or harbour maybe we could use a shoreline tag. * A problem with tagging the area of a bay is that while the shoreline is mostly well defined, the other edge is fuzzy. A possible solution is to use a multi polygon relation to tag just the non-shoreline segments of the bay outline as fuzzy=*. I suppose that using a multipolygon relation you can keep jest the shoreline segments together in another relation for say a larger harbour or river ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit and bays/coastline around Sydney
Hi Franc, I had a look and Navit doesn't have natural=bay as a supported tag even though it is a supported tag in OSM. It might be worth contacting the developers of Navit and ask them to include the tag natural=bay from a relation to render as water. Regards, Markus -Original Message- From: Franc Carter [mailto:franc.car...@gmail.com] Sent: Tuesday, 26 April 2011 7:59 PM To: OSM Australian Talk List Subject: [talk-au] navit and bays/coastline around Sydney Hi, I remember there was a discussion a while ago tagging of bays - which I didn;t pay much attention too ;-( They now seem to be in a relation for each Bay, which is nice for defining the area of the bay. However there are no tags on the ways themselves and as a result Navit does nto render them as water, which makes the map on Navit significantly less easy to use. My thought would be to put the coastline tag back on the ways and keep the relations - any thoughts or suggestions on how to recover Navit goodness ? cheers -- Franc ___ 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
[talk-au] navit and bays/coastline around Sydney
Hi, I remember there was a discussion a while ago tagging of bays - which I didn;t pay much attention too ;-( They now seem to be in a relation for each Bay, which is nice for defining the area of the bay. However there are no tags on the ways themselves and as a result Navit does nto render them as water, which makes the map on Navit significantly less easy to use. My thought would be to put the coastline tag back on the ways and keep the relations - any thoughts or suggestions on how to recover Navit goodness ? cheers -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
Some experimenting this morning indicates this is fixed in the version I run (svn 2453) cheers On Tue, Oct 13, 2009 at 6:20 PM, Liz wrote: > navit files built form osm data don't show a road if there is a reuse of > the > admin boundary for the road > > and so navigation around adelaide is going to be really interesting > as a lot of main roads have just disappeared > > please guys > don't reuse admin boundaries ( like ABS ) to replace perfectly good data > that > was already in the database. > > Just like i used to reuse the post office node for the place name and then > the > rendering rules changed for mapnik and none of those named places showed up > on > the map. > > :( > > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
Franc Carter wrote: > Hmm, I'm just in the early stages of browsing the code I got from svn in the > last 6-8 weeks, my initial > impression is that it does indeed return multiple match's for ways (up to > 10). I'll need to wait until the > weekend to confirm as only my carpc has an os release recent enough to > actually compile it ;-( It does return multiple matches, but only if the number of matching attributes is the same. The attributes to match come from the big attrmap string. -- Sam Couter | mailto:s...@couter.id.au OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 8530 03AE DE89 C75C signature.asc Description: Digital signature ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Thu, Oct 15, 2009 at 9:37 PM, Sam Couter wrote: > Elizabeth Dodd wrote: > > > http://www.openstreetmap.org/?lat=-34.9378&lon=138.6228&zoom=13&layers=B000FTF > > > http://www.openstreetmap.org/?lat=-34.2954&lon=138.9733&zoom=12&layers=B000FTF > > Using the latest SVN version of osm2navit, these two and the one near > Robertson NSW seem to turn out right. They're routable roads anyway, I > don't know how admin boundaries affect Navit. > Hmm, I'm just in the early stages of browsing the code I got from svn in the last 6-8 weeks, my initial impression is that it does indeed return multiple match's for ways (up to 10). I'll need to wait until the weekend to confirm as only my carpc has an os release recent enough to actually compile it ;-( cheers -- > Sam Couter | mailto:s...@couter.id.au > OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 8530 03AE DE89 C75C > > -BEGIN PGP SIGNATURE- > Version: GnuPG v1.4.10 (GNU/Linux) > > iEYEARECAAYFAkrW+3kACgkQhTADrt6Jx1zO/QCfazVmK6zvUUCkQUqIkKdC1Dex > LwEAn0uCKk33AB6a7ZdaRJYE6tb5I8+3 > =2d0d > -END PGP SIGNATURE- > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au > > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Thu, 15 Oct 2009, Sam Couter wrote: > Elizabeth Dodd wrote: > > http://www.openstreetmap.org/?lat=-34.9378&lon=138.6228&zoom=13&layers=B0 > >00FTF > > http://www.openstreetmap.org/?lat=-34.2954&lon=138.9733&zoom=12&layers=B0 > >00FTF > > Using the latest SVN version of osm2navit, these two and the one near > Robertson NSW seem to turn out right. They're routable roads anyway, I > don't know how admin boundaries affect Navit. I'm still downloading the planet.bin I've just compiled the latest svn The one near Robertson I redrew - the road has moved from the boundary Next to work out is why the long distance routing is crazy. It might be related to which roads have speed limits and which don't Liz ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
Elizabeth Dodd wrote: > http://www.openstreetmap.org/?lat=-34.9378&lon=138.6228&zoom=13&layers=B000FTF > http://www.openstreetmap.org/?lat=-34.2954&lon=138.9733&zoom=12&layers=B000FTF Using the latest SVN version of osm2navit, these two and the one near Robertson NSW seem to turn out right. They're routable roads anyway, I don't know how admin boundaries affect Navit. -- Sam Couter | mailto:s...@couter.id.au OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 8530 03AE DE89 C75C signature.asc Description: Digital signature ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Thu, 15 Oct 2009, Elizabeth Dodd wrote: > On Wed, 14 Oct 2009, Franc Carter wrote: > > Has anyone got a permalink to an area with this issue, so that I've got > > something to test > > with > > > > thanks > > > > On Tue, Oct 13, 2009 at 6:20 PM, Liz wrote: > > > navit files built form osm data don't show a road if there is a reuse > > > of the > > > admin boundary for the road > > > > > > and so navigation around adelaide is going to be really interesting > > > as a lot of main roads have just disappeared > > > > > > please guys > > > don't reuse admin boundaries ( like ABS ) to replace perfectly good > > > data that > > > was already in the database. > > > > > > Just like i used to reuse the post office node for the place name and > > > then the > > > rendering rules changed for mapnik and none of those named places > > > showed up on > > > the map. > > > > > > :( > > http://www.openstreetmap.org/?lat=-34.9378&lon=138.6228&zoom=13&layers=B000 >FTF > http://www.openstreetmap.org/?lat=-34.2954&lon=138.9733&zoom=12&layers=B000 >FTF > JS please what version of osm2navit are you uisng? ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
Franc Carter wrote: > On Tue, Oct 13, 2009 at 9:55 PM, John Smith wrote: > > > I wonder if it's possible to duplicate the object to break it up into > > 2 objects instead of one, and strip appropriate tags from the > > respective objects. > > > > Yep, that's the 'correct' solution that I am assuming, without having > actually delved > in to the code at the moment. The solution outlined by cp15 (see Liz's earlier message) is to modify attr_longest_match to return more than just one match when appropriate. The hard part will be deciding how to determine when more matches should be returned and what those matches should be. Returning too many matches will greatly increase the size of the Navit map file, which is obviously undesirable. -- Sam Couter | mailto:s...@couter.id.au OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 8530 03AE DE89 C75C signature.asc Description: Digital signature ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Wed, 14 Oct 2009, Franc Carter wrote: > Has anyone got a permalink to an area with this issue, so that I've got > something to test > with > > thanks > > On Tue, Oct 13, 2009 at 6:20 PM, Liz wrote: > > navit files built form osm data don't show a road if there is a reuse of > > the > > admin boundary for the road > > > > and so navigation around adelaide is going to be really interesting > > as a lot of main roads have just disappeared > > > > please guys > > don't reuse admin boundaries ( like ABS ) to replace perfectly good data > > that > > was already in the database. > > > > Just like i used to reuse the post office node for the place name and > > then the > > rendering rules changed for mapnik and none of those named places showed > > up on > > the map. > > > > :( http://www.openstreetmap.org/?lat=-34.9378&lon=138.6228&zoom=13&layers=B000FTF http://www.openstreetmap.org/?lat=-34.2954&lon=138.9733&zoom=12&layers=B000FTF ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
Has anyone got a permalink to an area with this issue, so that I've got something to test with thanks On Tue, Oct 13, 2009 at 6:20 PM, Liz wrote: > navit files built form osm data don't show a road if there is a reuse of > the > admin boundary for the road > > and so navigation around adelaide is going to be really interesting > as a lot of main roads have just disappeared > > please guys > don't reuse admin boundaries ( like ABS ) to replace perfectly good data > that > was already in the database. > > Just like i used to reuse the post office node for the place name and then > the > rendering rules changed for mapnik and none of those named places showed up > on > the map. > > :( > > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Tue, Oct 13, 2009 at 9:42 PM, Elizabeth Dodd wrote: > On Tue, 13 Oct 2009, Franc Carter wrote: > > I've been looking over the navit code and have made some minor tweaks to > > suite me. > > My first guess is that the issue might be able to be worked around by > > changing the order > > that osm2navit looks at properties - e.g making ti look at decide these > are > > roads instead > > of boundaries. > the part of the irc conversation below may help > Yep, interesting - my guess may have been wrong - but it gives me some leads - thanks > > > > > Of course this will mean boundaries will disappear in those places. > > > > Solving it properly will probably be trickier. I'll have a look at it on > > the weekend. > > > > cheers > > > > On Tue, Oct 13, 2009 at 7:25 PM, John Smith > wrote: > > > 2009/10/13 Elizabeth Dodd : > > > > On Tue, 13 Oct 2009, John Smith wrote: > > > >> This is really a bug in their software, specifically the osm2navit > > > >> binary, it would be better to file a bug and get them to fix the > issue > > > >> rather than trying to work around their bugs. > > > >> > > > >> Have you filed a bug report for this at all? > > > > > > > > I've sat on irc for nearly a week waiting for some action to discuss > it > > > > > > their > > > > > > > irc channel first > > > > as i also wanted to talk about the strange computed routes i was > > > > getting > > > > > > Pretty sure I had the same issue when I tried to report their routing > > > engine was screwy too. > > > > > > Alternatively there is plan B, there is a number of people on this > > > list that can code, and even more on the dev list, I'm sure between us > > > we can cook up a suitable patch. Is anyone familiar with the osm2navit > > > code at all? > > > > progress on IRC > nominated Thiele Highway, goes north from Kapunda in SA > > but trouble getting the OSM way ID > can get point identifiers instead > > > > 20:57] This is probably due to the "longest match" functionality in > osm2navit. osm items are converted to the navit items which are matching > the > most tags > [20:58] If there are several matches of the same length, the item > gets > duplicated, otherwise the shorter matches will be ignored > [20:58] Maybe we have to add an extra match to compensate for this > > > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Tue, Oct 13, 2009 at 9:55 PM, John Smith wrote: > 2009/10/13 Franc Carter : > > > > I've been looking over the navit code and have made some minor tweaks to > > suite me. > > My first guess is that the issue might be able to be worked around by > > changing the order > > that osm2navit looks at properties - e.g making ti look at decide these > are > > roads instead > > of boundaries. > > > > Of course this will mean boundaries will disappear in those places. > > I wonder if it's possible to duplicate the object to break it up into > 2 objects instead of one, and strip appropriate tags from the > respective objects. > Yep, that's the 'correct' solution that I am assuming, without having actually delved in to the code at the moment. It's been a *long* time seince I was a real developer, so most of my approaches tend to be aimed at getting stuff to work the way I want. So, unless it's easier than I am expecting I'll probably go for the 'it works' solution (obviously not expecting the patch to be accepted) -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
2009/10/13 Franc Carter : > > I've been looking over the navit code and have made some minor tweaks to > suite me. > My first guess is that the issue might be able to be worked around by > changing the order > that osm2navit looks at properties - e.g making ti look at decide these are > roads instead > of boundaries. > > Of course this will mean boundaries will disappear in those places. I wonder if it's possible to duplicate the object to break it up into 2 objects instead of one, and strip appropriate tags from the respective objects. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Tue, 13 Oct 2009, Franc Carter wrote: > I've been looking over the navit code and have made some minor tweaks to > suite me. > My first guess is that the issue might be able to be worked around by > changing the order > that osm2navit looks at properties - e.g making ti look at decide these are > roads instead > of boundaries. the part of the irc conversation below may help > > Of course this will mean boundaries will disappear in those places. > > Solving it properly will probably be trickier. I'll have a look at it on > the weekend. > > cheers > > On Tue, Oct 13, 2009 at 7:25 PM, John Smith wrote: > > 2009/10/13 Elizabeth Dodd : > > > On Tue, 13 Oct 2009, John Smith wrote: > > >> This is really a bug in their software, specifically the osm2navit > > >> binary, it would be better to file a bug and get them to fix the issue > > >> rather than trying to work around their bugs. > > >> > > >> Have you filed a bug report for this at all? > > > > > > I've sat on irc for nearly a week waiting for some action to discuss it > > > > their > > > > > irc channel first > > > as i also wanted to talk about the strange computed routes i was > > > getting > > > > Pretty sure I had the same issue when I tried to report their routing > > engine was screwy too. > > > > Alternatively there is plan B, there is a number of people on this > > list that can code, and even more on the dev list, I'm sure between us > > we can cook up a suitable patch. Is anyone familiar with the osm2navit > > code at all? > > progress on IRC nominated Thiele Highway, goes north from Kapunda in SA but trouble getting the OSM way ID can get point identifiers instead 20:57] This is probably due to the "longest match" functionality in osm2navit. osm items are converted to the navit items which are matching the most tags [20:58] If there are several matches of the same length, the item gets duplicated, otherwise the shorter matches will be ignored [20:58] Maybe we have to add an extra match to compensate for this ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
John Smith wrote: > Pretty sure I had the same issue when I tried to report their routing > engine was screwy too. The idlers and lurkers will read the backlog when they return. Conversations can go very slowly this way, but communication is possible. Bug tracker: http://trac.navit-project.org/ Forums (no idea how closely they're monitored): http://sourceforge.net/projects/navit/forums There doesn't appear to be a mailing list. > Alternatively there is plan B, there is a number of people on this > list that can code, and even more on the dev list, I'm sure between us > we can cook up a suitable patch. Is anyone familiar with the osm2navit > code at all? I've looked it over a few times and even modified it a little. It's 5500 lines of dense code with lots of global variables, nearly no comments and a highly abstracted attribute-driven metaschema so I'll make no claims of familiarity. -- Sam Couter | mailto:s...@couter.id.au OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 8530 03AE DE89 C75C signature.asc Description: Digital signature ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
I've been looking over the navit code and have made some minor tweaks to suite me. My first guess is that the issue might be able to be worked around by changing the order that osm2navit looks at properties - e.g making ti look at decide these are roads instead of boundaries. Of course this will mean boundaries will disappear in those places. Solving it properly will probably be trickier. I'll have a look at it on the weekend. cheers On Tue, Oct 13, 2009 at 7:25 PM, John Smith wrote: > 2009/10/13 Elizabeth Dodd : > > On Tue, 13 Oct 2009, John Smith wrote: > >> This is really a bug in their software, specifically the osm2navit > >> binary, it would be better to file a bug and get them to fix the issue > >> rather than trying to work around their bugs. > >> > >> Have you filed a bug report for this at all? > > I've sat on irc for nearly a week waiting for some action to discuss it > their > > irc channel first > > as i also wanted to talk about the strange computed routes i was getting > > Pretty sure I had the same issue when I tried to report their routing > engine was screwy too. > > Alternatively there is plan B, there is a number of people on this > list that can code, and even more on the dev list, I'm sure between us > we can cook up a suitable patch. Is anyone familiar with the osm2navit > code at all? > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
On Tue, 13 Oct 2009, John Smith wrote: > This is really a bug in their software, specifically the osm2navit > binary, it would be better to file a bug and get them to fix the issue > rather than trying to work around their bugs. > > Have you filed a bug report for this at all? I've sat on irc for nearly a week waiting for some action to discuss it their irc channel first as i also wanted to talk about the strange computed routes i was getting alas #navit has been very quiet -- Writing is easy; all you do is sit staring at the blank sheet of paper until drops of blood form on your forehead. -- Gene Fowler ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
2009/10/13 Elizabeth Dodd : > On Tue, 13 Oct 2009, John Smith wrote: >> This is really a bug in their software, specifically the osm2navit >> binary, it would be better to file a bug and get them to fix the issue >> rather than trying to work around their bugs. >> >> Have you filed a bug report for this at all? > I've sat on irc for nearly a week waiting for some action to discuss it their > irc channel first > as i also wanted to talk about the strange computed routes i was getting Pretty sure I had the same issue when I tried to report their routing engine was screwy too. Alternatively there is plan B, there is a number of people on this list that can code, and even more on the dev list, I'm sure between us we can cook up a suitable patch. Is anyone familiar with the osm2navit code at all? ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit files
2009/10/13 Liz : > navit files built form osm data don't show a road if there is a reuse of the > admin boundary for the road > > and so navigation around adelaide is going to be really interesting > as a lot of main roads have just disappeared > > please guys > don't reuse admin boundaries ( like ABS ) to replace perfectly good data that > was already in the database. This is really a bug in their software, specifically the osm2navit binary, it would be better to file a bug and get them to fix the issue rather than trying to work around their bugs. Have you filed a bug report for this at all? > Just like i used to reuse the post office node for the place name and then the > rendering rules changed for mapnik and none of those named places showed up on > the map. Did you file a bug report against mapnik for this? There is no reason a single node can't be used for both. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
[talk-au] navit files
navit files built form osm data don't show a road if there is a reuse of the admin boundary for the road and so navigation around adelaide is going to be really interesting as a lot of main roads have just disappeared please guys don't reuse admin boundaries ( like ABS ) to replace perfectly good data that was already in the database. Just like i used to reuse the post office node for the place name and then the rendering rules changed for mapnik and none of those named places showed up on the map. :( ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] Navit
>> I tried a few random points on the Hume and on the way down to > wollongong and >> 'it works for me' ;-( >> >> I'm using a recentish svn version and one of John's recentish >> Australia.bin >> files >> >> cheers >> > > i've got a very recent (yesterday) bin file from John, but I'll try the > upgrade from svn > I'm not sure when i installed navit on the machine > Once I convinced apt-get that an svn version from pini was newer than the official package I was able to install a new navit and I can drive along the motorways now. Except for a bit of road at Robertson which is an admin boundary the routing is now pretty good. I'm going to split the road off the boundary and re-enter the road separately. Hope no one objects. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] Navit
> I tried a few random points on the Hume and on the way down to wollongong and > 'it works for me' ;-( > > I'm using a recentish svn version and one of John's recentish > Australia.bin > files > > cheers > i've got a very recent (yesterday) bin file from John, but I'll try the upgrade from svn I'm not sure when i installed navit on the machine ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] Navit
> Interesting, It puts me on motorways all the time (sometimes it would be > nice to > have more choice). Have you checked the oneways etc. Have you got a > permalink > to a problematic road > > cheers > > 150km of Hume Highway and a few km in wollongong same thing motorway > sends me to go down first exit any direction nothing to do with where i want to go ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] Navit
Interesting, It puts me on motorways all the time (sometimes it would be nice to have more choice). Have you checked the oneways etc. Have you got a permalink to a problematic road cheers On Sun, Sep 27, 2009 at 7:26 PM, wrote: > this might just be a silly / basic question > > how does one tell Navit to allow one to drive on the motorway? > > I double checked, its set to car, not to horse cycle or pedestrian, and it > won't send me along a motorway > > or is this a problem with au data?? > > Liz > who luckily knew the way because it would have been a very long journey > along the backroads > > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
[talk-au] Navit
this might just be a silly / basic question how does one tell Navit to allow one to drive on the motorway? I double checked, its set to car, not to horse cycle or pedestrian, and it won't send me along a motorway or is this a problem with au data?? Liz who luckily knew the way because it would have been a very long journey along the backroads ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
[talk-au] Navit Update
I seem to have overlooked the update script since rebuilding the server, I've manually run this and a new navit file is now available: http://map-data.bigtincan.com/data/Australia-20090831.navit.bin I've now setup a cronjob to do this automatically after lunch on Sundays. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
I think the latter. - Ben. -Original Message- From: b.schulz...@scu.edu.au Sent: 07 August 2009 13:33 To: talk-au@openstreetmap.org Subject: Re: [talk-au] navit I wonder, how hard would it be to write a script which looked at the ABS boundaries and placed an is_in= tag on every way/node/whatever which is completely within the multipolygon relation? Or would it be better to modify the routing software to look at the ABS boundaries instead? Just throwing ideas around Brent - Original Message - From: Sam Couter Date: Friday, August 7, 2009 1:07 pm Subject: Re: [talk-au] navit To: talk-au@openstreetmap.org > John Smith wrote: > > Nothing special, just followed the directions on this wiki page: > > > > > http://www.rigacci.org/wiki/doku.php/doc/appunti/hardware/eeepc_navit > As a result, most of the towns and suburbs in Australia can't be > searched for due to a lack of is_in tags. My patch for osm2navit > (attached) is a bit heavy-handed but trivial. If you've got a build > environment set up for navit, give it a go. > -- > Sam Couter > | mailto:s...@couter.id.au > OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 > 8530 03AE DE89 C75C >___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
--- On Fri, 7/8/09, b.schulz...@scu.edu.au wrote: > I'm just trying to brainstorm a way of achieving what > you're describing while still conforming to strict > boundary sources as opposed to drawing a boundary around a > town by hand. As far as I'm aware this is the information we have available to us, even if Franc hasn't uploaded it all yet: http://wiki.openstreetmap.org/wiki/Import/Catalogue/ABS_Data For anything else, like Greater Sydney Regional etc we'd need to either re-use existing boundaries and making polygons from it, similar to how the state boundary exists. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
Does the ABS have other boundaries which we can use for this? For example "Sydney" currently would just refer to the "Sydney" LGA, rather than that area which people would call "Greater Sydney". Basically, do they publish boundaries such as "greater sydney" "the newcastle region" etc? If they don't then can we get LGA boundaries? That would work for basically smaller than Sydney. For instance the Lismore LGA encompasses Lismore, Goonellabah, Chilcotts Grass...Well, pretty much everything from half-way to Casino up to the Byron Shire. Otherwise could there be a way for us to define regions such as "Greater Newcastle" or "Gympie and surrounds" by forming larger multipolygon relations from a set of suburb boundaries manually? I'm just trying to brainstorm a way of achieving what you're describing while still conforming to strict boundary sources as opposed to drawing a boundary around a town by hand. Brent > This is also why I was suggesting using admin_level=9 for town > boundaries, for towns like gympie that don't match postcode or > admin_level=10 boundaries. Then that information can be used to > find any street within the Gympie town area, not just the ABS > Gympie boundary. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
--- On Fri, 7/8/09, Sam Couter wrote: > That one seems to work much better. It's a bit of a hack but it would be much better if navit pulled the is_in information from admin boundaries, there is a boundary for most countries and things within Australia. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
--- On Thu, 6/8/09, b.schulz...@scu.edu.au wrote: > I wonder, how hard would it be to > write a script which looked at the ABS boundaries and placed > an is_in= tag on every way/node/whatever which is completely > within the multipolygon relation? You don't need the is_in tags, you use administrative boundaries to find that information out. > Or would it be better to modify the routing software to > look at the ABS boundaries instead? Exactly, although to make it easier on apps they can pre-process the boundaries and the stuff in the boundaries to know what a town or way is in, for is_in to actually be useful it would have to be applied consistently to everything, not just boundaries, and that isn't likely to happen nor should it when that information can be extracted from boundaries. This way if you wanted to find a post code, you can just look for the post code boundary. This is also why I was suggesting using admin_level=9 for town boundaries, for towns like gympie that don't match postcode or admin_level=10 boundaries. Then that information can be used to find any street within the Gympie town area, not just the ABS Gympie boundary. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
John Smith wrote: > I just ran the new osm2navit over a data file a couple of days old and it > went from 40M to 50M as a result of the extra is_in data. > > http://maps.bigtincan.com/data/Australia-20090804.navit.bin That one seems to work much better. -- Sam Couter | mailto:s...@couter.id.au OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 8530 03AE DE89 C75C signature.asc Description: Digital signature ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
I wonder, how hard would it be to write a script which looked at the ABS boundaries and placed an is_in= tag on every way/node/whatever which is completely within the multipolygon relation? Or would it be better to modify the routing software to look at the ABS boundaries instead? Just throwing ideas around Brent - Original Message - From: Sam Couter Date: Friday, August 7, 2009 1:07 pm Subject: Re: [talk-au] navit To: talk-au@openstreetmap.org > John Smith wrote: > > Nothing special, just followed the directions on this wiki page: > > > > > http://www.rigacci.org/wiki/doku.php/doc/appunti/hardware/eeepc_navit > As a result, most of the towns and suburbs in Australia can't be > searched for due to a lack of is_in tags. My patch for osm2navit > (attached) is a bit heavy-handed but trivial. If you've got a build > environment set up for navit, give it a go. > -- > Sam Couter > | mailto:s...@couter.id.au > OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 > 8530 03AE DE89 C75C > ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
I just ran the new osm2navit over a data file a couple of days old and it went from 40M to 50M as a result of the extra is_in data. http://maps.bigtincan.com/data/Australia-20090804.navit.bin ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
--- On Thu, 6/8/09, Sam Couter wrote: > As a result, most of the towns and suburbs in Australia > can't be > searched for due to a lack of is_in tags. My patch for > osm2navit > (attached) is a bit heavy-handed but trivial. If you've got > a build > environment set up for navit, give it a go. I didn't have a need for one, but I'll check their code out of SVN. I've been trying to figure out a SQL search, based on the AU data in the pgsql DB, the problem is it's been too long since I needed to do something this complex, do you know of an easy way to turn something like this: George St, Sydney, NSW into a SQL query? ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
John Smith wrote: > Nothing special, just followed the directions on this wiki page: > > http://www.rigacci.org/wiki/doku.php/doc/appunti/hardware/eeepc_navit As a result, most of the towns and suburbs in Australia can't be searched for due to a lack of is_in tags. My patch for osm2navit (attached) is a bit heavy-handed but trivial. If you've got a build environment set up for navit, give it a go. -- Sam Couter | mailto:s...@couter.id.au OpenPGP fingerprint: A46B 9BB5 3148 7BEA 1F05 5BD5 8530 03AE DE89 C75C Index: navit/osm2navit.c === --- navit/osm2navit.c (revision 2448) +++ navit/osm2navit.c (working copy) @@ -1424,7 +1424,9 @@ if (item_is_town(*item_bin) && attr_strings[attr_string_label]) { char *tok,*buf=is_in_buffer; if (!buf[0]) -strcpy(is_in_buffer, "Unknown"); +strncpy(buf, "Australia", BUFFER_SIZE); + else +strncat(buf, ", Australia", BUFFER_SIZE - strlen(buf)); while ((tok=strtok(buf, ","))) { while (*tok==' ') tok++; signature.asc Description: Digital signature ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
Thanks, weekly is very useful anyway. I've just returned from a mapping trip, that used your navit map on my car computer - it was very useful for finding missed streets cheers On Sun, Aug 2, 2009 at 2:51 PM, John Smith wrote: > > > > --- On Tue, 28/7/09, Franc Carter wrote: > > > Excellent thanks - I'll be interested in the Aussie > > file > > It's too much of a hassle to do daily files, but I'm setting up scripts to > run after lunchtime on Sundays at present, the Aussie file is up, and the NZ > one is still being produced. > > http://maps.bigtincan.com/data/Australia-20090802.navit.bin > > and when it's completed > > http://maps.bigtincan.com/data/NZ-20090802.navit.bin > > > > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
--- On Tue, 28/7/09, Franc Carter wrote: > Excellent thanks - I'll be interested in the Aussie > file It's too much of a hassle to do daily files, but I'm setting up scripts to run after lunchtime on Sundays at present, the Aussie file is up, and the NZ one is still being produced. http://maps.bigtincan.com/data/Australia-20090802.navit.bin and when it's completed http://maps.bigtincan.com/data/NZ-20090802.navit.bin ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
--- On Wed, 29/7/09, Mark Hetherington wrote: > I'm very interested in the Australia image. I'm curious how > it's being processed though, when I did the processing > myself I found searching in navit did not work. My > investigations to this point indicated I needed a patched > version of osm2navit to put all nodes "in" Australia. Any > special tricks up your sleeve? Nothing special, just followed the directions on this wiki page: http://www.rigacci.org/wiki/doku.php/doc/appunti/hardware/eeepc_navit I downloaded the data for Australia, rather than Italy though :) ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] navit
Hi John, I'm very interested in the Australia image. I'm curious how it's being processed though, when I did the processing myself I found searching in navit did not work. My investigations to this point indicated I needed a patched version of osm2navit to put all nodes "in" Australia. Any special tricks up your sleeve? Cheers, Mark John Smith wrote: > For those that are interested, I'll be building a daily or weekly navit file, > I ran it over aussie data only and the file comes out to about 41M, and if > there is any interest I could keep building a NZ only file and it'll be about > 4M. > > http://maps.bigtincan.com/data/Australia-20090728.navit.bin > http://maps.bigtincan.com/data/NZ-20090728.navit.bin > > > > > ___ > 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
Re: [talk-au] navit
Excellent thanks - I'll be interested in the Aussie file cheers On Wed, Jul 29, 2009 at 1:44 PM, John Smith wrote: > > For those that are interested, I'll be building a daily or weekly navit > file, I ran it over aussie data only and the file comes out to about 41M, > and if there is any interest I could keep building a NZ only file and it'll > be about 4M. > > http://maps.bigtincan.com/data/Australia-20090728.navit.bin > http://maps.bigtincan.com/data/NZ-20090728.navit.bin > > > > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au > -- Franc ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
[talk-au] navit
For those that are interested, I'll be building a daily or weekly navit file, I ran it over aussie data only and the file comes out to about 41M, and if there is any interest I could keep building a NZ only file and it'll be about 4M. http://maps.bigtincan.com/data/Australia-20090728.navit.bin http://maps.bigtincan.com/data/NZ-20090728.navit.bin ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] Navit
Forgot to mention, that routing over largish distances makes the router go all screwy, 600km and it thinks I have to go 50m or I'm already at my destination etc. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
[talk-au] Navit
Ok, I'll preface this by saying that navit seems to be a really really good start, but the UI needs a lot of spit and polish. When you are running the GTK interface you can't make it go to full screen. Which if you are on a small screen to begin with that's a lot of space wasted displaying menus and title bars you only need to access infrequently. When you are running the internal interface you can just point and click (or right click) to mark a destination, you have to go through a bunch of different screens similar to a satnav device. Even then I couldn't get it to route, not sure if it's something I did or didn't do or what. Clicking options to change something, like 2D to 3D display doesn't update the display to show it has done something, or drop you to the previous menu screen or anything, so you think nothing has happened, or you are left assuming it might have happened. Also for some reason the ABS boundaries seem to be given an unusually high preference over roads, when you zoom out (f8 key, f7 is zoom in) you don't see many roads at all, it's mostly ABS boundaries which is very confusing to navigate. On a side note, it took 130M of bzip compressed OSM data and turned it into a 50M data file. I can't figure out how to dump the data from pgsql to OSM as I'd be happy to script OSM to navit data files on the map server I built. Does anyone know how to get the info back out of pgsql? ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au