Re: [OSM-talk] temporary=yes tag
On Monday 13 October 2008 00:17:39 you wrote: > If you're looking for a one word English tag, I think seasonal may be > an option. It is usually used for things that are only in summer, or > winter, etc, but can be used for holidays as well. > Excellent, thank you, that's surely the right word! I'm replying to the list which I guess was what you intended? seasonal=yes sounds perfect :-) ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] temporary=yes tag
On Sunday 12 October 2008 22:58:40 you wrote: > Hi, > > > It's not temporary "until it's done", it's sometimes there, sometimes > > not. Like christmas decorations, only with those you know they are around > > in December. > > Then, logically speaking, the item they want to tag is not a "bullring > that is sometimes not there" - it is a "place where they sometimes have > a bullring". Much like we'd tag a fairground a fairground and not a > temporary fair... isn't that so? I wanted to check how fairs are mapped currently but the wiki appears to be down. Are you suggesting we tag plain ground? I'm just not sure what you mean. Fairs here are placed over ground that doesn't have anything particularly fair-related (parking lots, plazas, parks, etc). I also think keeping the bullring tag makes sense, for searching and such. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] temporary=yes tag
On Sunday 12 October 2008 22:55:24 you wrote: > 2008/10/13 Matias D'Ambrosio <[EMAIL PROTECTED]>: > > On Sunday 12 October 2008 22:32:17 Ryszard Mikke wrote: > >> 2008/10/3 Matias D'Ambrosio <[EMAIL PROTECTED]>: > >> > While discussing leisure=bullring in the talk-es (spanish speaking > >> > list), the issue came up of bullrings that are only set up on certain > >> > occasions (holidays and other special events). This is also common for > >> > fairs, circuses, etc. So we thought something like a tag > >> > temporary=yes/no would be nice for them, I searched in the wiki and > >> > found nothing of the sort, is something of the sort available? > >> > opening_hours is not fit for this, obviously. I'm not sure "temporary" > >> > is the right word (specially since it might give the idea that the OSM > >> > tag is temporary, which is not intended). > >> > >> Wouldn't it be better to have "expires=date" instead? > >> So trhat it would disappear automatically after the date. > > > > It's not temporary "until it's done", it's sometimes there, sometimes > > not. Like christmas decorations, only with those you know they are around > > in December. > > So maybe "periodical" is your word? > I guess that's closer, but periodical indicates regularity. It's still better than temporary (which is just too confusing). ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] temporary=yes tag
Hi, > It's not temporary "until it's done", it's sometimes there, sometimes not. > Like christmas decorations, only with those you know they are around in > December. Then, logically speaking, the item they want to tag is not a "bullring that is sometimes not there" - it is a "place where they sometimes have a bullring". Much like we'd tag a fairground a fairground and not a temporary fair... isn't that so? Bye Frederik -- Frederik Ramm ## eMail [EMAIL PROTECTED] ## N49°00'09" E008°23'33" ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] temporary=yes tag
On Sunday 12 October 2008 22:32:17 Ryszard Mikke wrote: > 2008/10/3 Matias D'Ambrosio <[EMAIL PROTECTED]>: > > While discussing leisure=bullring in the talk-es (spanish speaking > > list), the issue came up of bullrings that are only set up on certain > > occasions (holidays and other special events). This is also common for > > fairs, circuses, etc. So we thought something like a tag temporary=yes/no > > would be nice for them, I searched in the wiki and found nothing of the > > sort, is something of the sort available? opening_hours is not fit for > > this, obviously. I'm not sure "temporary" is the right word (specially > > since it might give the idea that the OSM tag is temporary, which is not > > intended). > > Wouldn't it be better to have "expires=date" instead? > So trhat it would disappear automatically after the date. > It's not temporary "until it's done", it's sometimes there, sometimes not. Like christmas decorations, only with those you know they are around in December. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] temporary=yes tag
2008/10/3 Matias D'Ambrosio <[EMAIL PROTECTED]>: > While discussing leisure=bullring in the talk-es (spanish speaking list), the > issue came up of bullrings that are only set up on certain occasions > (holidays and other special events). This is also common for fairs, circuses, > etc. So we thought something like a tag temporary=yes/no would be nice for > them, I searched in the wiki and found nothing of the sort, is something of > the sort available? opening_hours is not fit for this, obviously. I'm not > sure "temporary" is the right word (specially since it might give the idea > that the OSM tag is temporary, which is not intended). Wouldn't it be better to have "expires=date" instead? So trhat it would disappear automatically after the date. Ryszard Mikke -- -- ==> Gardzi miodem - szalony! <== -- -- ==> -- Kubus Puchatek <== -- [EMAIL PROTECTED][EMAIL PROTECTED] ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Pedestrians on cycleways
The 'OSM tags for routing/Access-Restrictions' page http://wiki.openstreetmap.org/index.php/OSM_tags_for_routing/Access-Restrictions has it that pedestrians are not allowed on cycleways unless an explicit 'foot=yes' is added. This seems to reflect the situation in Germany. However, in Belgium and The Netherlands, the default is that pedestrians are allowed on cycle tracks. I'd like to invite people to provide input on how the rules are in other countries, preferably by adding default-access tables to the wiki page. Philip Homburg ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-legal-talk] Paid services from OSM
On 10/12/08, Richard Fairhurst <[EMAIL PROTECTED]> wrote: > Goodness me, that's an enormously confrontational-sounding posting, Whoops, it was meant to be more controversial than confrontational. > * I would like OSMF to publish the current licence > * IMHO OSMF should publish the licence e.t.c Publish the licence +1 If things have been sorted and addressed, then please tell us! If things haven't been sorted and addressed, then tell us and we can continue debating use cases about avant-garde performance artists. (I've a couple ready in draft) ___ legal-talk mailing list [EMAIL PROTECTED] http://lists.openstreetmap.org/listinfo/legal-talk
[OSM-talk] Georeferencing images WAS ping OpenAerialMap
> - Original Message - > From: "andrew" <[EMAIL PROTECTED]> > To: > Sent: Thursday, October 09, 2008 9:02 PM > Subject: Re: [OSM-talk] ping OpenAerialMap > > > > On Thursday 09 October 2008 16:20:56 Richard Fairhurst wrote: >> Concerns were >> expressed that some OAM content had been georeferenced by use of >> Google Maps. > > So should all georeferencing be done by visiting control points with > a gps? > > I've some aerial photos that I need to find how to georeference, 4MB > jpegs and 17 overlapping tiles, is there a program you can > recommend? > > Andrew Heggie > How about Quantum GIS with the georeferencing plugin http://www.qgis.org/ David ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] ping OpenAerialMap
sylva.icuklive.co.uk> writes: > > On Sun, 12 Oct 2008 08:48:36 + (UTC), Jukka Rahkonen wrote: > > >> I've some aerial photos that I need to find how to georeference, 4MB > >> jpegs and 17 overlapping tiles, is there a program you can > >> recommend? > > > > > >Could you tolerate a non-integrated system and measure pixel coordinates and > >corresponding true world coordinates with a separate program and feed in > >these > >ground control points from command line? > > If you mean can I attribute a single pixel in the image with a > coordinate on the ground then yes. I'm not a programmer or a > cartographer just a pc user but can use the command line. > > >In this case you could use > >gdal-programs (gdal_translate and gdalwarp) and get a single mosaiced > >georeferenced image. Download FWTools package, it includes both gdal > tools and > OpenEV Viewer that can be used for defining ground control points. > > OK I'll have a look, thanks. I've previously used a free windows > mapping program but it uses a metre grid and even when projected with > british OS datum and co ordinates its 70m out (even the control points > get displaced) so I have some sort of problem with the datum and > projection. Hi, First step is to measure ground control points and insert them into image file. That is done with gdal_translate and -gpc option. Read: http://gdal.org/gdal_translate.html Command line will be something like: gdal_translate -of GTiff -gcp [...] -gcp [...] -gcp [...] input.jpg with_gpc.tif Now you'll have a new image "with_gpc.tif". Next step is to use gdalwarp. If you have given coordinates as WGS84 lat/lon degrees and you'd like to keep the same projection (epsg:4326) it should go this way: gdalwarp -of GTiff -s_srs epsg:4326 -t_srs epsg:4326 with_gpc.tif final_georeferenced_tif_image.tif This is right from my head and I cannot guarantee that commands are totally correct but they should be close to working. You can output into whatever projection you like, output to different file format. You can also mosaic all your images together once they have ground control points measured. However, try to keep things simple in the begining and start playing with individual images. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] lanes
On Sun, Oct 12, 2008 at 9:13 AM, Gervase Markham <[EMAIL PROTECTED]>wrote: > graham wrote: > > - Say I have a bus lane (or cycle lane) running along one side of a > > two-way road (the most common situation where I am). Just attaching a > > 'left' tag to it makes it dependent on nobody ever reversing the > > arbitrary direction attached to the way. > > No, it doesn't, because editors would make flipping the left/right part > of the "reverse way" command, just as they now flip "oneway"s and so on. > There are various consistency checks editors need to do when making > changes - this is just another one. > > > - More serious, I think: it just feels quite arbitrary as a solution: I > > would have to tag a lane as being 'left' in relation to the random > > direction the arrows on the way happen to be pointing, rather than in > > relation to anything in the real world. > > How else do you unambiguously intrinsically define the side of a way? > Unless it's a closed way, where you have "inside" and "outside", the > only way is to give it a direction (which it has) and say "right" or > "left". > > You can define it extrinsically using things like "north" or "west" but > that runs into problems with roads which curve, or even run in a circle. > > > 2. have an 'origin' tag to be used on the first node of a way > > independent of direction; if the way direction flipped, the origin would > > stay in place. 'Left' or 'right' would be in relation to the origin > > node. Still completely arbitrary where the origin goes, and how do you > > find it on a long way? > > If my proposal has disadvantages, then this has them all but adds some > new ones too :-) > > > 3. Make more of a separation between internal representation of ways and > > user views in josm/potlatch. All ways have a direction which is > > independent of the 'arrowed' direction which can be displayed to users, > > and is fixed - a totally arbitrary value used only as a fixed reference. > > Why? Why not just use the arrowed direction? After all, it's not exactly > complicated to flip tags when you change the direction of the way. > That's why the scheme is generic :left and :right - so it can be used on > any tags. > > Gerv > Just as a note, someone already added automatic left/right tag changing to JOSM when a way is reversed. It works for a few different schemes, including :left and :right. Karl ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] More vandalism
I have previously reported how someone with the screen name 'thebigfatgeek' has messed up a roundabout in Pretoria. Well, yesterday he changed a lot more stuff, most of it clearly being wrong. See for example http://www.openstreetmap.org/browse/way/4259221/history Joining ways with different names, creating evil semicolons : http://www.openstreetmap.org/browse/way/4616530/history Changing bridleways to residential roads. http://www.openstreetmap.org/browse/way/26165816/history Local members have messaged him before and after yesterday, but the damage is already done. Clearly OSM can do without the help of someone who pays so little attention to detail. I would be grateful if Frederik can run his revert tool again. In fact I wouldn't mind if it's in a cron job. Regards, Nic ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] lanes
graham wrote: > - Say I have a bus lane (or cycle lane) running along one side of a > two-way road (the most common situation where I am). Just attaching a > 'left' tag to it makes it dependent on nobody ever reversing the > arbitrary direction attached to the way. No, it doesn't, because editors would make flipping the left/right part of the "reverse way" command, just as they now flip "oneway"s and so on. There are various consistency checks editors need to do when making changes - this is just another one. > - More serious, I think: it just feels quite arbitrary as a solution: I > would have to tag a lane as being 'left' in relation to the random > direction the arrows on the way happen to be pointing, rather than in > relation to anything in the real world. How else do you unambiguously intrinsically define the side of a way? Unless it's a closed way, where you have "inside" and "outside", the only way is to give it a direction (which it has) and say "right" or "left". You can define it extrinsically using things like "north" or "west" but that runs into problems with roads which curve, or even run in a circle. > 2. have an 'origin' tag to be used on the first node of a way > independent of direction; if the way direction flipped, the origin would > stay in place. 'Left' or 'right' would be in relation to the origin > node. Still completely arbitrary where the origin goes, and how do you > find it on a long way? If my proposal has disadvantages, then this has them all but adds some new ones too :-) > 3. Make more of a separation between internal representation of ways and > user views in josm/potlatch. All ways have a direction which is > independent of the 'arrowed' direction which can be displayed to users, > and is fixed - a totally arbitrary value used only as a fixed reference. Why? Why not just use the arrowed direction? After all, it's not exactly complicated to flip tags when you change the direction of the way. That's why the scheme is generic :left and :right - so it can be used on any tags. Gerv ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] When does Potlatch upload?
Ben Laenen wrote: > Well, it only starts happening when I've done quite a lot of > mapping, so > pinpointing the exact trigger for it is difficult... I can only tell > that it happens regularly after mapping for a while. Got it, I think; will be fixed in 0.11e (trac #1256). cheers Richard ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] When does Potlatch upload?
On Sunday 12 October 2008, Richard Fairhurst wrote: > Ben Laenen wrote: > > just a little question about recent Potlatch versions: I've noticed > > that > > from time to time changes are uploaded while objects are still > > selected. I remember that it used to be that it only uploaded > > changes on deselecting. > > If it's changed, that's a bug, steps to reproduce would be great. Well, it only starts happening when I've done quite a lot of mapping, so pinpointing the exact trigger for it is difficult... I can only tell that it happens regularly after mapping for a while. Ben ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] ping OpenAerialMap
On Sun, 12 Oct 2008 08:48:36 + (UTC), Jukka Rahkonen wrote: >> I've some aerial photos that I need to find how to georeference, 4MB >> jpegs and 17 overlapping tiles, is there a program you can >> recommend? > > >Could you tolerate a non-integrated system and measure pixel coordinates and >corresponding true world coordinates with a separate program and feed in these >ground control points from command line? If you mean can I attribute a single pixel in the image with a coordinate on the ground then yes. I'm not a programmer or a cartographer just a pc user but can use the command line. >In this case you could use >gdal-programs (gdal_translate and gdalwarp) and get a single mosaiced >georeferenced image. Download FWTools package, it includes both gdal tools and >OpenEV Viewer that can be used for defining ground control points. OK I'll have a look, thanks. I've previously used a free windows mapping program but it uses a metre grid and even when projected with british OS datum and co ordinates its 70m out (even the control points get displaced) so I have some sort of problem with the datum and projection. I also am looking at the GRASS gis in Linux but not learning very fast. Andrew Heggie ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] lanes
Gervase Markham wrote: > LeedsTracker wrote: >> As Shaun says, the unresolved issue of lane handed-ness seems to be >> blocking this lane issue. > > This is anothe occasion where a generic :left/:right proposal would be > useful... > I had missed that thread till now. But it doesn't seem to have come to any usable conclusion. I have two problems with the left/right solution (neither original): - Say I have a bus lane (or cycle lane) running along one side of a two-way road (the most common situation where I am). Just attaching a 'left' tag to it makes it dependent on nobody ever reversing the arbitrary direction attached to the way. I quite often find I'm reversing sections of bidirectional ways (eg to join two of them); if anyone ever did that with a way with 'lane:left' style tags the lane would suddenly flip to the wrong side of the road. - More serious, I think: it just feels quite arbitrary as a solution: I would have to tag a lane as being 'left' in relation to the random direction the arrows on the way happen to be pointing, rather than in relation to anything in the real world. Unfortunately the only suggestions I can come up with are pretty lame, but maybe they might trigger better ideas in someone else: 1. draw ways as multiple parallel lanes. Maybe all two-way roads should be represented as two parallel ways in opposite directions? I'm sure everyone would love the work involved in that ;-). 2. have an 'origin' tag to be used on the first node of a way independent of direction; if the way direction flipped, the origin would stay in place. 'Left' or 'right' would be in relation to the origin node. Still completely arbitrary where the origin goes, and how do you find it on a long way? 3. Make more of a separation between internal representation of ways and user views in josm/potlatch. All ways have a direction which is independent of the 'arrowed' direction which can be displayed to users, and is fixed - a totally arbitrary value used only as a fixed reference. Consider all bidirectional ways 'as if' they were two ways orientated in reference to the hidden direction. When a user selects 'add a lane to a way' (or bus-stop or any other 'sided' feature) josm/potlatch switch to a two-way representation of the way; the user clicks on the particular one of the two he/she wants to add the lane to. Internally this could be named 'left' or 'right' lane; but the user would never see the words 'left' or 'right' or need to explicitly tag the sidedness of the lane. Dumps all the work onto developers and makes it hard for end-users to fix things that go wrong since the inner workings have become invisible :-( As far as bus lanes go, it looks like the easiest thing to do is use 'busway' with exactly the same tags as cycleway. So once the sidedness problems are resolved for cycleways busways will be fine too. In the meantime I guess I'll hold off and start mapping historic buildings instead ;-) Graham > > > Gerv > > > ___ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] lanes
Gerv wrote: > LeedsTracker wrote: > > As Shaun says, the unresolved issue of lane handed-ness seems > to be > > blocking this lane issue. > > This is anothe occasion where a generic :left/:right proposal > would be > useful... > > At a risk of re-opening a discussion, what is the unresolved issue of handedness? Surely if you can have oneway=yes in the direction of the arrows and oneway=-1 for oneway in the opposite direction of the arrows, then left and right can also be defined relative to the direction of the way. Completely no use for nodes of course which are points and don't have any form of direction. Ed ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] ping OpenAerialMap
andrew sylva.icuklive.co.uk> writes: > > On Thursday 09 October 2008 16:20:56 Richard Fairhurst wrote: > > Concerns were > > expressed that some OAM content had been georeferenced by use of > > Google Maps. > > So should all georeferencing be done by visiting control points with > a gps? > > I've some aerial photos that I need to find how to georeference, 4MB > jpegs and 17 overlapping tiles, is there a program you can > recommend? Could you tolerate a non-integrated system and measure pixel coordinates and corresponding true world coordinates with a separate program and feed in these ground control points from command line? In this case you could use gdal-programs (gdal_translate and gdalwarp) and get a single mosaiced georeferenced image. Download FWTools package, it includes both gdal tools and OpenEV Viewer that can be used for defining ground control points. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] lanes
LeedsTracker wrote: > As Shaun says, the unresolved issue of lane handed-ness seems to be > blocking this lane issue. This is anothe occasion where a generic :left/:right proposal would be useful... Gerv ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk