Re: [Talk-us] Changing Data Attribution
Very nice. I did not know about that convention. Will use in the future for sure. Zeke On Sun, Jun 13, 2010 at 12:57 PM, Toby Murray wrote: > According to the Key:source page on the wiki, an object can have > multiple source tags. So if you go out and survey a TIGER road and > discover that the name is incorrect you could change the name and add > a "source:name=survey" tag. I guess this allows you to distinguish the > source of specific elements of an object. > > According to this I guess I should tag all the I-70 exits in western > Kansas that I recently added ref= tags to with > "source:ref=KDOT" since I used a KDOT map (these are public domain) to > get the exit numbers. > > > Toby > > > On Sun, Jun 13, 2010 at 10:40 AM, Zeke Farwell wrote: > > The way I see it, any given feature may have many components to it's > data. > > With a road you've got the geometry, name, classification, surface, etc…. > . > > These components may all come from one source, or they make come from > > several. If the road is unmodified since the TIGER import then the > source > > of all these components is TIGER. If I've taken a GPS trace of the road, > > noted the name, classification, and surface, then I can make a new way > (or > > modify the old one) and tag it "source=GPS" (or source=GPS; survey if you > > want to get detailed). The source of the geometry is a GPS and the > source > > of the other components is my survey. Since I've collected all new data > > from the ground, TIGER is no longer the source of any component. If > however > > I've re-aligned the geometry of a road off Yahoo imagery, I have not > > verified the other components. The name, surface, and classification > are > > still from TIGER alone so it should be preserved in the source tag > > (source=Yahoo; TIGER). Of course if it's a road in my neighborhood that > I > > know is called Spruce Ave, and should be tertiary then I'd tag it > > "source=Yahoo; local_knowledge". Geometry source from Yahoo, other > > components from my own knowledge of the area. Then the next mapper will > see > > that someone with knowledge of the area edited that way last and the data > > can be trusted more than TIGER. Basically I just think about what the > > actual source of the data is, and since there are multiple pieces of data > > about each feature it's perfectly reasonable to have more than one > source. > > The attribution tag is a bit different. It's not about the source of the > > data, but about giving credit to the person, organization, government, > etc > > who made the data available to OSM. As long as attribution is > required and > > the source of any component of a feature is still said organization, then > > the attribution tag should be preserved. TIGER data is public domain, so > no > > attribution is required. This is why data from the TIGER import only has > a > > source tag. Source is really for the benefit of the next mapper, so they > > can gauge if the data they have might be better than what already exists. > > Zeke > > Burlington, VT > > > > > > > > On Sat, Jun 12, 2010 at 10:00 PM, Josh Kraayenbrink > > wrote: > >> > >> This is in relation to Lars' question with attribution on nodes and > ways. > >> I have been thinking, possibly incorrectly, about attribution on data. > The > >> Tiger import was great, but as you all know, not perfectly accurate. I > have > >> been "reviewing" and almost all roads, ways, etc that have been imported > in > >> my area are now corrected. The problem comes in the ways we review this > >> imported data. If I go around and get .gpx traces and use that to move > the > >> imported roads and mark the road as reviewed, is it really still to be > >> attributed to the Tiger import, or does the source/attribution actually > >> change to my trace? And what about a trace of aerial photography, make a > >> difference? I do believe attributing the data to Tiger is no longer > >> accurate, but not one hundred percent > >> On top of that, do I delete the current data and create a new, more > >> accurate piece of data in its place, or simply move the Tiger data and > >> change the attribution. While this makes absolutely no difference to the > >> current state of the map, it does make a difference for the history of > the > >> data. Is my data actually a newer improved version of the Tiger data, or > is > >> it a new piece of data? > >> This does not just apply to Tiger data either. If you map something, but > I > >> edit or move it, where does the attribution lie in this? Just something > I > >> have been scratching my head on as I have picked up my mapping > >> and actually getting in the field as of late. Curious on the trains of > >> thought or consensus on this. > >> ___ > >> Talk-us mailing list > >> Talk-us@openstreetmap.org > >> http://lists.openstreetmap.org/listinfo/talk-us > >> > > > > > > ___ > > Talk-us maili
Re: [Talk-us] MassGIS source/attribution tags
On 06/13/2010 07:39 AM, Greg Troxel wrote: > So if you are considering writing a bot that does > > for each node in mass > > if the node has only massgis attribution tags, and no others, delete > the tags (because surely it's either pointless and should be garbage > collected, or is a member of a way) > > then I'd say that sounds good. Yep, that was precisely what I had in mind. Alright, it's a simple algorithm, so hopefully I'll have time to write a bot that does this. -- Lars Ahlzen l...@ahlzen.com ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Changing Data Attribution
According to the Key:source page on the wiki, an object can have multiple source tags. So if you go out and survey a TIGER road and discover that the name is incorrect you could change the name and add a "source:name=survey" tag. I guess this allows you to distinguish the source of specific elements of an object. According to this I guess I should tag all the I-70 exits in western Kansas that I recently added ref= tags to with "source:ref=KDOT" since I used a KDOT map (these are public domain) to get the exit numbers. Toby On Sun, Jun 13, 2010 at 10:40 AM, Zeke Farwell wrote: > The way I see it, any given feature may have many components to it's data. > With a road you've got the geometry, name, classification, surface, etc…. . > These components may all come from one source, or they make come from > several. If the road is unmodified since the TIGER import then the source > of all these components is TIGER. If I've taken a GPS trace of the road, > noted the name, classification, and surface, then I can make a new way (or > modify the old one) and tag it "source=GPS" (or source=GPS; survey if you > want to get detailed). The source of the geometry is a GPS and the source > of the other components is my survey. Since I've collected all new data > from the ground, TIGER is no longer the source of any component. If however > I've re-aligned the geometry of a road off Yahoo imagery, I have not > verified the other components. The name, surface, and classification are > still from TIGER alone so it should be preserved in the source tag > (source=Yahoo; TIGER). Of course if it's a road in my neighborhood that I > know is called Spruce Ave, and should be tertiary then I'd tag it > "source=Yahoo; local_knowledge". Geometry source from Yahoo, other > components from my own knowledge of the area. Then the next mapper will see > that someone with knowledge of the area edited that way last and the data > can be trusted more than TIGER. Basically I just think about what the > actual source of the data is, and since there are multiple pieces of data > about each feature it's perfectly reasonable to have more than one source. > The attribution tag is a bit different. It's not about the source of the > data, but about giving credit to the person, organization, government, etc > who made the data available to OSM. As long as attribution is required and > the source of any component of a feature is still said organization, then > the attribution tag should be preserved. TIGER data is public domain, so no > attribution is required. This is why data from the TIGER import only has a > source tag. Source is really for the benefit of the next mapper, so they > can gauge if the data they have might be better than what already exists. > Zeke > Burlington, VT > > > > On Sat, Jun 12, 2010 at 10:00 PM, Josh Kraayenbrink > wrote: >> >> This is in relation to Lars' question with attribution on nodes and ways. >> I have been thinking, possibly incorrectly, about attribution on data. The >> Tiger import was great, but as you all know, not perfectly accurate. I have >> been "reviewing" and almost all roads, ways, etc that have been imported in >> my area are now corrected. The problem comes in the ways we review this >> imported data. If I go around and get .gpx traces and use that to move the >> imported roads and mark the road as reviewed, is it really still to be >> attributed to the Tiger import, or does the source/attribution actually >> change to my trace? And what about a trace of aerial photography, make a >> difference? I do believe attributing the data to Tiger is no longer >> accurate, but not one hundred percent >> On top of that, do I delete the current data and create a new, more >> accurate piece of data in its place, or simply move the Tiger data and >> change the attribution. While this makes absolutely no difference to the >> current state of the map, it does make a difference for the history of the >> data. Is my data actually a newer improved version of the Tiger data, or is >> it a new piece of data? >> This does not just apply to Tiger data either. If you map something, but I >> edit or move it, where does the attribution lie in this? Just something I >> have been scratching my head on as I have picked up my mapping >> and actually getting in the field as of late. Curious on the trains of >> thought or consensus on this. >> ___ >> Talk-us mailing list >> Talk-us@openstreetmap.org >> http://lists.openstreetmap.org/listinfo/talk-us >> > > > ___ > Talk-us mailing list > Talk-us@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-us > > ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] MassGIS source/attribution tags
On Sun, Jun 13, 2010 at 7:39 AM, Greg Troxel wrote: > then I'd say that sounds good. > +1 -- Bill n1...@arrl.net bill.n1...@gmail.com ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Changing Data Attribution
The way I see it, any given feature may have many components to it's data. With a road you've got the geometry, name, classification, surface, etc…. . These components may all come from one source, or they make come from several. If the road is unmodified since the TIGER import then the source of all these components is TIGER. If I've taken a GPS trace of the road, noted the name, classification, and surface, then I can make a new way (or modify the old one) and tag it "source=GPS" (or source=GPS; survey if you want to get detailed). The source of the geometry is a GPS and the source of the other components is my survey. Since I've collected all new data from the ground, TIGER is no longer the source of any component. If however I've re-aligned the geometry of a road off Yahoo imagery, I have not verified the other components. The name, surface, and classification are still from TIGER alone so it should be preserved in the source tag (source=Yahoo; TIGER). Of course if it's a road in my neighborhood that I know is called Spruce Ave, and should be tertiary then I'd tag it "source=Yahoo; local_knowledge". Geometry source from Yahoo, other components from my own knowledge of the area. Then the next mapper will see that someone with knowledge of the area edited that way last and the data can be trusted more than TIGER. Basically I just think about what the actual source of the data is, and since there are multiple pieces of data about each feature it's perfectly reasonable to have more than one source. The attribution tag is a bit different. It's not about the source of the data, but about giving credit to the person, organization, government, etc who made the data available to OSM. As long as attribution is required and the source of any component of a feature is still said organization, then the attribution tag should be preserved. TIGER data is public domain, so no attribution is required. This is why data from the TIGER import only has a source tag. Source is really for the benefit of the next mapper, so they can gauge if the data they have might be better than what already exists. Zeke Burlington, VT On Sat, Jun 12, 2010 at 10:00 PM, Josh Kraayenbrink wrote: > This is in relation to Lars' question with attribution on nodes and ways. > > I have been thinking, possibly incorrectly, about attribution on data. The > Tiger import was great, but as you all know, not perfectly accurate. I have > been "reviewing" and almost all roads, ways, etc that have been imported in > my area are now corrected. The problem comes in the ways we review this > imported data. If I go around and get .gpx traces and use that to move the > imported roads and mark the road as reviewed, is it really still to be > attributed to the Tiger import, or does the source/attribution actually > change to my trace? And what about a trace of aerial photography, make a > difference? I do believe attributing the data to Tiger is no longer > accurate, but not one hundred percent > > On top of that, do I delete the current data and create a new, more > accurate piece of data in its place, or simply move the Tiger data and > change the attribution. While this makes absolutely no difference to the > current state of the map, it does make a difference for the history of the > data. Is my data actually a newer improved version of the Tiger data, or is > it a new piece of data? > > This does not just apply to Tiger data either. If you map something, but I > edit or move it, where does the attribution lie in this? Just something I > have been scratching my head on as I have picked up my mapping > and actually getting in the field as of late. Curious on the trains of > thought or consensus on this. > > ___ > Talk-us mailing list > Talk-us@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-us > > ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Changing Data Attribution
Josh Kraayenbrink writes: > I have been thinking, possibly incorrectly, about attribution on data. The > Tiger import was great, but as you all know, not perfectly accurate. I have > been "reviewing" and almost all roads, ways, etc that have been imported in > my area are now corrected. The problem comes in the ways we review this > imported data. If I go around and get .gpx traces and use that to move the > imported roads and mark the road as reviewed, is it really still to be > attributed to the Tiger import, or does the source/attribution actually > change to my trace? And what about a trace of aerial photography, make a > difference? I do believe attributing the data to Tiger is no longer > accurate, but not one hundred percent There is a difference between attribution and source traceability. Attribution is about OSM acknowledging that we used Tiger day, and is similar to the BY requirement in CC-BY-SA. If you adjust a way that was imported from TIGER, then really you've produced something that is a derived work from both of you. The history shows that you edited it, and that seems adequate. As far as attribution, that's really just "Thanks to TIGER for giving us some bits to start with", and that remains. pgpkqR5m1GrHI.pgp Description: PGP signature ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] MassGIS source/attribution tags
Lars Ahlzen writes: > Linear features imported from MassGIS data have the appropriate source= > and attribution= tags. That makes sense. However, each individual node > has these identical tags as well, including nodes that have no other tags. > > For such nodes with no other tags, this seems redundant. It's a bit > similar to TLIDs on nodes from the TIGER import (which were removed), i.e: > http://wiki.openstreetmap.org/wiki/TIGER_fixup/node_tags > > Not only does this take extra space, but it complicates editing. JOSM > (and probably other editors) highlights nodes that have attributes, > nodes don't get deleted with the parent way, etc. > > Is there a good reason to keep such tags around, other than "it takes > effort to change it"? I don't think we need attribution tags on nodes which solely exist to be points in ways. The attribution request is social and perhaps a bit legal, not technical, and is more along the lines of "OSM should acknowledge that it has massgis-based data" than "it should be possible to determine from the database the provenance of each massgis-derived object". I recently suffered from the nodes-remaining-when-way-deleted problem in JOSM, but didn't realize why at the time. There aren't too many massgis-sourced roads that don't exist, but I found one at Cape Cod Light in North Truro (but then again the whole lighthouse was moved). So if you are considering writing a bot that does for each node in mass if the node has only massgis attribution tags, and no others, delete the tags (because surely it's either pointless and should be garbage collected, or is a member of a way) then I'd say that sounds good. pgpd81DKtsGNg.pgp Description: PGP signature ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
