[OSM-talk] Tagging the source
Philip Barnes wrote: I would prefer to keep the source tag with the object. Within a changeset I will often have some roads where source is GPS, have traced some buildings from bing, and added a few pub/shop names where source is survey I did put my hand up for a tag which is automatically applied for those of us who forget it ;) If I have a background layer up it automatically adds that tag to each object. If I'm selecting stuff from another source then that gets added. ALL of this could well be in one change set, and in fact I may switch between bing and OS layers simply to add other details, so a 'changeset' tag would not be suitable? But AUTOMATICALLY adding that data per object would mean that it does get added :) Of cause I still object to a lot of this stuff being 'free format' ... a 'layer_id' on these tags would be fine, and then the rest of the details are pulled up from that! -- Lester Caine - G8HFL - Contact - http://lsces.co.uk/wiki/?page=contact L.S.Caine Electronic Services - http://lsces.co.uk EnquirySolve - http://enquirysolve.com/ Model Engineers Digital Workshop - http://medw.co.uk Rainbow Digital Media - http://rainbowdigitalmedia.co.uk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
We don't have to choose between the 2 solutions (source on objects vs. source on change set). It's possible to use a fall back system, something like : For display : If no source is available on an object, show the source of the last change set (if available) For editing : If no source is set on a change set, set it from the layers automatically. It requires to modify the tools, but it should make every mapper happy. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
On Sep 28, 2012, at 1:56 PM, Olivier Croquette wrote: For display : If no source is available on an object, show the source of the last change set (if available) Actually that should read: show the sources of all change sets that created or modified the object. The whole thing would need a better specification, but you get the idea. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
The tag 'source' on the changeset has some pros and cons: Pros: - size in database - works only if the changeset comes from a single source Cons: - impossible to modify after changeset is closed - attribution lost in data extracts, planet (separate file for changesets) - doesn't allow more than one source per changeset - rebuild one element history is not trivial (one API call per element + one per changeset multiplied by amount of versions) The tag 'source' on elements: Pros: - support multiple sources during edit session - attribution present in planet, extracts - can be modified at any time - retrieve one element history is easy (one API call) Cons: - size in database - works fine only for small edit sessions (or imports) Pieren ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
Pieren wrote: - attribution lost in data extracts, planet (separate file for changesets) ACTUALLY that is probably the killer? Local working with truncated extracts SHOULD still report this type of data? -- Lester Caine - G8HFL - Contact - http://lsces.co.uk/wiki/?page=contact L.S.Caine Electronic Services - http://lsces.co.uk EnquirySolve - http://enquirysolve.com/ Model Engineers Digital Workshop - http://medw.co.uk Rainbow Digital Media - http://rainbowdigitalmedia.co.uk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
2012/9/28 Lester Caine les...@lsces.co.uk: I did put my hand up for a tag which is automatically applied for those of us who forget it ;) If I have a background layer up it automatically adds that tag to each object. If I'm selecting stuff from another source then that gets added. ALL of this could well be in one change set, and in fact I may switch between bing and OS layers simply to add other details, so a 'changeset' tag would not be suitable? But AUTOMATICALLY adding that data per object would mean that it does get added :) would you also opt for automatically removing these tags, e.g. when a imagery layer is not activated and I move a node? Are you advocating multivalue-source-lists if during the edit of an object the mapper looked at several different images? Honestly I am against a automatic mechanism to add tags, and I'd also doubt that these tags would improve our data quality. I see a huge overhead for really little benefit. Please think about it: what is the benefit for other mappers to know that you traced over imagery provider A's imagery and not that of B? Either you think it is correct or you don't and you will improve it. In rare cases it might help you to know whether this is based on outdated sources, but more often you will already see this by the date of the edit. If the original source is outdated and you know this from knowledge of the real situation you will anyway improve the data. You might be able to find areas which are based on outdated imagery to resurvey. but you can find inactive areas just as well by looking at the changes and dates of them. Even more, a comment like tracing from bing aerial imagery doesn't even tell you which version of their imagery you used (supposedly that around the time the edit was done, but you won't know in 2 years time which version or which date this was, especially if it isn't your home region). It mostly boils down to the simple question: does the mapper have local knowledge or doesn't he. Did he recently survey the area? That's not enough reason to stuff the db with mostly pointless metadata like active imagery layers in the editor during an edit. It might be an idea to add this information automatically to changesets. Btw.: does anybody on this list know if there is metadata for Bing (for a given area) available (all based on the zoomlevel of course, e.g. when did they survey, what was the resolution, where are the seems of their imagery/survey, if the data is not from them, who originally produced it, etc.) cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
Lester Caine wrote: I did put my hand up for a tag which is automatically applied for those of us who forget it ;) If I have a background layer up it automatically adds that tag to each object. In Potlatch you can simply press 'B' (for 'Background') to add the source= tag for the current background imagery. It isn't added automatically, and won't be, because it doesn't follow that you're necessarily tracing from a background source just because it's displayed. You might be using a GPS track, or your own local knowledge, or a vector background layer, or whatever. cheers Richard -- View this message in context: http://gis.19327.n5.nabble.com/Ref-Re-All-you-ve-ever-wanted-to-know-about-the-french-cadastre-tp5727997p5728076.html Sent from the General Discussion mailing list archive at Nabble.com. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
2012/9/28 Lester Caine les...@lsces.co.uk: Pieren wrote: - attribution lost in data extracts, planet (separate file for changesets) ACTUALLY that is probably the killer? Local working with truncated extracts SHOULD still report this type of data? come on, I think you are overexxagerating this. Have a look how apple handles attribution. They distribute a big blob and a separate (hard to find) text file that says: contains data of a, b, c, , z, aa, bb, cc, ..., zz, aaa, bbb, They don't even tell you where they use which data source. ;-) I am not advocating to act accordingly, but saying that in every data extract there must be full history (that is also attribution, to see who edited the stuff), or changeset information, or source-tags , or a source-tag on every single object might not be necessary. IMHO you also retain attribution if you say: data from openstreetmap.org and on openstreetmap.org you get all the necessary attribution info (e.g. in the wiki, from the API with changeset-comments, etc.). It was already in the past like this: someone who imports stuff with osm2pgsql almost never retains this information (as long as he works with the standard style file). Cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
2012/9/28 Pieren pier...@gmail.com: The tag 'source' on elements: Pros: - support multiple sources during edit session - attribution present in planet, extracts - can be modified at any time - retrieve one element history is easy (one API call) Cons: - size in database - works fine only for small edit sessions (or imports) there are other cons to source tags on objects: when does the original contribution faint? This is a very difficult question and every mapper who wants to improve an object which has a source-tag gets the burden to care about. He might act wrong if he doesn't retain the previous attribution (because he would be obfuscating the source) and he might also act wrong if he keeps it (because he would attribute an object to a source from which there is probably few if anything left). This makes pretty clear IMHO that source tags (in a project like osm, where no data is stable) clearly don't belong to object, but to the changes which create and modify those objects. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
Martin Koppenhoefer wrote: It mostly boils down to the simple question: does the mapper have local knowledge or doesn't he. Did he recently survey the area? Reviewing my own 'processing', even where I have local knowledge I know I should be adding tags when I'm tracing, but when I'm in the flow THAT disrupts things. *I* would like something that flags that *I* pulled something from BING and something else from OS or where ever. Having to type some bloody long string each time means simply it does not happen :) And yes part of the reason would be to flag a version of a data source! That's not enough reason to stuff the db with mostly pointless metadata like active imagery layers in the editor during an edit. It might be an idea to add this information automatically to changesets. Which would not work for all the reasons already given. Only if you save on every change of configuration would that be accurate. Perhaps all I am asking for is a 'default' set of tags that get added which I have to change manually, but the DATA certainly is important, and tags like 'start_date' should be populated by default! -- Lester Caine - G8HFL - Contact - http://lsces.co.uk/wiki/?page=contact L.S.Caine Electronic Services - http://lsces.co.uk EnquirySolve - http://enquirysolve.com/ Model Engineers Digital Workshop - http://medw.co.uk Rainbow Digital Media - http://rainbowdigitalmedia.co.uk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
2012/9/28 Lester Caine les...@lsces.co.uk: I have to change manually, but the DATA certainly is important, and tags like 'start_date' should be populated by default! How would the editor (program) know about the start_date? What are you using this tag for? cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
Martin Koppenhoefer wrote: I am not advocating to act accordingly, but saying that in every data extract there must be full history (that is also attribution, to see who edited the stuff), or changeset information, or source-tags , or a source-tag on every single object might not be necessary. IMHO you also retain attribution if you say: data from openstreetmap.org and on openstreetmap.org you get all the necessary attribution info (e.g. in the wiki, from the API with changeset-comments, etc.). Full history comes from the main database ... I was just advocating 'source' which adds to the information when one is looking at something and saying That is wrong! you may know straight away that it is a candidate to update. -- Lester Caine - G8HFL - Contact - http://lsces.co.uk/wiki/?page=contact L.S.Caine Electronic Services - http://lsces.co.uk EnquirySolve - http://enquirysolve.com/ Model Engineers Digital Workshop - http://medw.co.uk Rainbow Digital Media - http://rainbowdigitalmedia.co.uk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
2012/9/28 Lester Caine les...@lsces.co.uk: Full history comes from the main database ... I was just advocating 'source' which adds to the information when one is looking at something and saying That is wrong! you may know straight away that it is a candidate to update. yes, but often when something is wrong, the source-tag is as well ;-). I have seen lots of source=PSG (coastline) where the data obviously was far too detailed to be from PSG, it is because people hardly remove those (meanwhile unvalid) source-tags. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
On 28 September 2012 15:13, Martin Koppenhoefer dieterdre...@gmail.com wrote: yes, but often when something is wrong, the source-tag is as well ;-). I have seen lots of source=PSG (coastline) where the data obviously was far too detailed to be from PSG, it is because people hardly remove those (meanwhile unvalid) source-tags. Agree on that. There is lots of UK data with source=NPE or PGS that has obviously been subsequently adjusted from Bing or elsewhere, probably done this myself many times and forgotten to change the source tag. That's why I would prefer some automation. If I have Bing and an OS based layer open and active in JOSM then I don't see a problem with automatically adding those as sources to the changeset. Those tags should then be visible when retrieving object history. Mappers could still override that with source tags on the objects if required. Kevin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
Martin Koppenhoefer wrote: I have to change manually, but the DATA certainly is important, and tags like 'start_date' should be populated by default! How would the editor (program) know about the start_date? What are you using this tag for? Eventually a corrected date can be added ... we have sufficient historic mapping on line now, but NEW stuff going forward would at least be properly tagged from day one ... and in 10 years time you know that an object IS at least 10 years old ... it's just a matter of planning for the future. Again I'm just as lax at adding it, mainly because the editors don't present it! -- Lester Caine - G8HFL - Contact - http://lsces.co.uk/wiki/?page=contact L.S.Caine Electronic Services - http://lsces.co.uk EnquirySolve - http://enquirysolve.com/ Model Engineers Digital Workshop - http://medw.co.uk Rainbow Digital Media - http://rainbowdigitalmedia.co.uk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging the source
The only real problem I see mentioned above is the overall size of the database. It seems to me that we are somehow confusing problems caused by the data itself with problems caused by its storage in the database. Couldn't we simply work on a scheme that would normalize the database, so that we'd have to store each piece of information (e.g. source value as well as all the other tags) only once? I haven't worked much around the OSM pgsql database schemas, so it may not be as easy as I'd hope. Jerome On Fri, Sep 28, 2012 at 7:39 AM, Kevin Peat k...@k3v.eu wrote: On 28 September 2012 15:13, Martin Koppenhoefer dieterdre...@gmail.com wrote: yes, but often when something is wrong, the source-tag is as well ;-). I have seen lots of source=PSG (coastline) where the data obviously was far too detailed to be from PSG, it is because people hardly remove those (meanwhile unvalid) source-tags. Agree on that. There is lots of UK data with source=NPE or PGS that has obviously been subsequently adjusted from Bing or elsewhere, probably done this myself many times and forgotten to change the source tag. That's why I would prefer some automation. If I have Bing and an OS based layer open and active in JOSM then I don't see a problem with automatically adding those as sources to the changeset. Those tags should then be visible when retrieving object history. Mappers could still override that with source tags on the objects if required. Kevin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
[OSM-talk] [tagging] Against source=survey
The tag source=survey hides the fact that those coordinates were derived from the signal provided by a particular infrastructure: the US satellites. Maybe those ways derived from GPS tracks should be tagged source=NASA or source=Pentagon, instead of source=survey, which does not cite the true source. Just to prevent people from believing that cartography consists in switching the on/off button of a toy called Garmin... Lucas ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] [tagging] Against source=survey
Why not go all the way and say source=cesium? In all seriousness, though, what's wrong with source=GPS as an alternative? Survey, to me, implies a crew out with tripods and such. Karl On Tue, May 6, 2008 at 12:32 PM, Juan Lucas Dominguez Rubio [EMAIL PROTECTED] wrote: The tag source=survey hides the fact that those coordinates were derived from the signal provided by a particular infrastructure: the US satellites. Maybe those ways derived from GPS tracks should be tagged source=NASA or source=Pentagon, instead of source=survey, which does not cite the true source. Just to prevent people from believing that cartography consists in switching the on/off button of a toy called Garmin... Lucas ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] [tagging] Against source=survey
Karl Newman wrote: Sent: 06 May 2008 8:58 PM To: Juan Lucas Dominguez Rubio Cc: talk@openstreetmap.org Subject: Re: [OSM-talk] [tagging] Against source=survey Why not go all the way and say source=cesium? In all seriousness, though, what's wrong with source=GPS as an alternative? Survey, to me, implies a crew out with tripods and such. I use source=GPS but see no reason why source=survey wouldn't be valid. The Ordnance Survey use GPS for their work nowadays and in reality don't normally operate as a crew, the one man survey is the way they work. In other words, very little difference to the way that OSMers gather data. The principal difference is that we don't work at the +/- 3cm tolerance and a large portion of our surveying is done kinetically. Cheers Andy ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk