Re: [Talk-GB] Ficticious embankments? Vandalism.
On 21/3/20 11:02 pm, ael wrote: On Sat, Mar 21, 2020 at 10:45:53AM +1100, Warin wrote: On 18/3/20 1:42 am, ael wrote: On Tue, Mar 17, 2020 at 11:25:24AM +, Devonshire wrote: On Tue, Mar 17, 2020, at 2:08 AM, Warin wrote: On 17/3/20 8:02 am, ael wrote: The inability to mark an object's location as "authorititive" has always seemed like a massive shortcoming of the project to me. Stopping people re-aligning things based on a bad phone GPS or badly aligned aerial imagery is impossible and even realising that things have been incorrectly moved is random at best. I agree entirely and have often wished for exactly that. I sometimes use source=gps_surveys (plural) to try to convey that this is not just one random gps trace. "source=average of multiple gps surveys, high accuracy" Be really descriptive... the 's' on the end of gps surveys is really easy to miss. Well, yes, and I do quite often expand the source tag to try to convey more. But in your example "high" accuracy is a problem. If I was using differential gps with cm accuracy, I would call that "high" accuracy. In the present case, the accuracy is not really known, but probably approaching a meter. But I guess that sort of thing could be included in a source tag, although free form text might be better in a note tag. Is not the source tag free form? Indeed any OSM tag is 'free form' - i.e. "Any tags you like". I don't think any one uses the source tag other than mappers looking at where the data came from. As such it can be anything you think suitable. The 'high accuracy' is a relative term simply there to help those that don't understand the previous 'average of multiple gps surveys' But my impression is that many armchair mappers just don't look. In this case any tags will be ignored. Pointless coming up with another tag. ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On Sat, Mar 21, 2020 at 10:45:53AM +1100, Warin wrote: > On 18/3/20 1:42 am, ael wrote: > > On Tue, Mar 17, 2020 at 11:25:24AM +, Devonshire wrote: > > > On Tue, Mar 17, 2020, at 2:08 AM, Warin wrote: > > > > On 17/3/20 8:02 am, ael wrote: > > > The inability to mark an object's location as "authorititive" has always > > > seemed like a massive shortcoming of the project to me. Stopping people > > > re-aligning things based on a bad phone GPS or badly aligned aerial > > > imagery is impossible and even realising that things have been > > > incorrectly moved is random at best. > > I agree entirely and have often wished for exactly that. I sometimes use > > source=gps_surveys (plural) to try to convey that this is not just one > > random gps trace. > > "source=average of multiple gps surveys, high accuracy" > > Be really descriptive... the 's' on the end of gps surveys is really easy to > miss. Well, yes, and I do quite often expand the source tag to try to convey more. But in your example "high" accuracy is a problem. If I was using differential gps with cm accuracy, I would call that "high" accuracy. In the present case, the accuracy is not really known, but probably approaching a meter. But I guess that sort of thing could be included in a source tag, although free form text might be better in a note tag. But my impression is that many armchair mappers just don't look. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On 18/3/20 1:42 am, ael wrote: On Tue, Mar 17, 2020 at 11:25:24AM +, Devonshire wrote: On Tue, Mar 17, 2020, at 2:08 AM, Warin wrote: On 17/3/20 8:02 am, ael wrote: The inability to mark an object's location as "authorititive" has always seemed like a massive shortcoming of the project to me. Stopping people re-aligning things based on a bad phone GPS or badly aligned aerial imagery is impossible and even realising that things have been incorrectly moved is random at best. I agree entirely and have often wished for exactly that. I sometimes use source=gps_surveys (plural) to try to convey that this is not just one random gps trace. In this case, I just had source=gps_survey. "source=average of multiple gps surveys, high accuracy" Be really descriptive... the 's' on the end of gps surveys is really easy to miss. I too regret the awful smartphone (and satnav) gps traces which suggest all gps is rubbish. I try not to upload any gps which is not reasonably accurate. And add a note if the gps quality is poor when it still has value, perhaps because there are no other traces in the area. I suppose that we ought to start a discussion on the tagging list to suggest source:accuracy = low|medium|high|differential Why? This can be placed in to source=phone for location, poor accuracy. This then is where you look for information on the source ... rather than yet another key. ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism
On Thu, Mar 19, 2020 at 03:08:39PM +, ael wrote: > On Sat, Mar 14, 2020 at 06:09:26PM +, ael wrote: > > I have just noticed some new "Embankments" added around a fortnight ago. > > These were added to some stone circles in Cornwall which I know well and > > > > I have added a changeset comment. > > https://www.openstreetmap.org/changeset/81640861 > > The user has reverted the changes to the Hurlers in Cornwall, and > apologised. He has just replied again to the changeset and says that he has reverted all the other embankments. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism
On Sat, Mar 14, 2020 at 06:09:26PM +, ael wrote: > I have just noticed some new "Embankments" added around a fortnight ago. > These were added to some stone circles in Cornwall which I know well and > > I have added a changeset comment. > https://www.openstreetmap.org/changeset/81640861 The user has reverted the changes to the Hurlers in Cornwall, and apologised. However, I had a quick look at some of his other changesets, and he seems to have done similar things at other sites that have not (?yet) been reverted. One I noticed was a "Coffin Stone" which now has an embankment. Is that real? I doubt it. I thought that he had added another embankment at the Rollright Stones, but I couldn't see that just now. But he did remove two tags without explanation. I wonder whether other points with their history were deleted. I haven't looked any further, but it looks as if this is a widespread problem. I haven't formally alerted the DWG as yet: the user might yet revert all the changes, but I hope others will keep an eye on all of this. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On Tue, Mar 17, 2020 at 11:25:24AM +, Devonshire wrote: > On Tue, Mar 17, 2020, at 2:08 AM, Warin wrote: > > On 17/3/20 8:02 am, ael wrote: > > The inability to mark an object's location as "authorititive" has always > seemed like a massive shortcoming of the project to me. Stopping people > re-aligning things based on a bad phone GPS or badly aligned aerial imagery > is impossible and even realising that things have been incorrectly moved is > random at best. I agree entirely and have often wished for exactly that. I sometimes use source=gps_surveys (plural) to try to convey that this is not just one random gps trace. In this case, I just had source=gps_survey. I too regret the awful smartphone (and satnav) gps traces which suggest all gps is rubbish. I try not to upload any gps which is not reasonably accurate. And add a note if the gps quality is poor when it still has value, perhaps because there are no other traces in the area. I suppose that we ought to start a discussion on the tagging list to suggest source:accuracy = low|medium|high|differential or some such. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On Tue, Mar 17, 2020 at 11:25:24AM +, Devonshire wrote: > On Tue, Mar 17, 2020, at 2:08 AM, Warin wrote: > > On 17/3/20 8:02 am, ael wrote: > > The inability to mark an object's location as "authorititive" has always > seemed like a massive shortcoming of the project to me. Stopping people > re-aligning things based on a bad phone GPS or badly aligned aerial imagery > is impossible and even realising that things have been incorrectly moved is > random at best. I agree entirely and have often wished for exactly that. I sometimes use source=gps_surveys (plural) to try to convey that this is not just one random gps trace. In this case, I just had source=gps_survey. I too regret the awful smartphone (and satnav) gps traces which suggest all gps is rubbish. I try not to upload any gps which is not reasonably accurate. And add a note if the gps quality is poor when it still has value, perhaps because there are no other traces in the area. I suppose that we ought to start a discussion on the tagging list to suggest source:accuracy = low|medium|high|differential or some such. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On Tue, Mar 17, 2020 at 01:08:52PM +1100, Warin wrote: > On 17/3/20 8:02 am, ael wrote: > > > > I have only just got around to looking in more detail, and discovered > > that it is much worse than I had realised: vandalism. > > > > I have taken waypoints on nearly all of the individual stones, and then > > refined those positions with waypoint averaging on multiple visits. > > > In cases like this I would use the source tag on the way so that others have > a very good chance of seeing it and respecting the previous work rather than > simply changing it to what they think it should be. Indeed. I nearly always include a source tag for those sorts of reasons. In this case the user did not just move the points, but deleted them, source tag and all. So destroying the history. > It is too easy to over look hard work that may have gone into establishing > data. > A single GPS trace is fine if that is all there is, better to average many > GPS traces, in some locations I have 50+. In this case, I used a "waypoint averaging" function on my Garmin. That collects fixes over long periods and applies statistical algorithms to refine the coordinates. And this can be done on many occasions to refine the position even further. I had done this on many visits, and each averaging took considerable time. Often in pretty nasty weather. So my coordinates should have been very accurate. Hence my anger that they should have been deleted. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On Tue, Mar 17, 2020 at 09:53:07AM +, John Aldridge wrote: > On 17-Mar-20 02:08, Warin wrote: > > A single GPS trace is fine if that is all there is, better to average many > > GPS traces, in some locations I have 50+. > > Though, AIUI, once you've reached this level of precision, remaining errors > are likely to be systematic (e.g. satellites in a particular direction being > generally received via a -- delaying -- reflection rather than directly). No > amount of averaging will help with that. Well, the Garmin averaging goes some way to improve on that. That is one of the reasons that waypoint averaging can be done over several visits: it gives some sort of averaging over reflections, atmospheric changes even satellite calibration. Not up to professional differential GPS, of course, but potentially accurate to maybe a meter. Certainly way better then using imagery with the poor resolution, parallax errors and all the rest, which is probably what the user used. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On Tue, Mar 17, 2020, at 2:08 AM, Warin wrote: > On 17/3/20 8:02 am, ael wrote: >> In cases like this I would use the source tag on the way so that others have >> a very good chance of seeing it and respecting the previous work rather than >> simply changing it to what they think it should be. It is too easy to over >> look hard work that may have gone into establishing data. A single GPS trace >> is fine if that is all there is, better to average many GPS traces, in some >> locations I have 50+. What would you put in the source tag in this case and does it make a difference? The inability to mark an object's location as "authorititive" has always seemed like a massive shortcoming of the project to me. Stopping people re-aligning things based on a bad phone GPS or badly aligned aerial imagery is impossible and even realising that things have been incorrectly moved is random at best. Kevin ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On 17/3/20 8:53 pm, John Aldridge wrote: On 17-Mar-20 02:08, Warin wrote: A single GPS trace is fine if that is all there is, better to average many GPS traces, in some locations I have 50+. Though, AIUI, once you've reached this level of precision, remaining errors are likely to be systematic (e.g. satellites in a particular direction being generally received via a -- delaying -- reflection rather than directly). No amount of averaging will help with that. Yes, a source of error. When placed over some reasonable imagery any discrepancies can be seen, evaluated and if necessary compensated for. However most users will not be aware of the discrepancy as they would be using another GPS to 'see' it. If the GPS tracks are separated by days, weeks and months then the satellites will have changed, and the satellite positions too. Usually there will be 'outliers' that are more than 2 standard deviations from the average. It is valid to delete these from computations. I don't bother with the calculations and do it by eye, more satisfying than loading it all into a program, taking the result and entering it. And I like to see the raw data anyway. ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On 17-Mar-20 02:08, Warin wrote: A single GPS trace is fine if that is all there is, better to average many GPS traces, in some locations I have 50+. Though, AIUI, once you've reached this level of precision, remaining errors are likely to be systematic (e.g. satellites in a particular direction being generally received via a -- delaying -- reflection rather than directly). No amount of averaging will help with that. John ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On 17/3/20 8:02 am, ael wrote: I have only just got around to looking in more detail, and discovered that it is much worse than I had realised: vandalism. I have taken waypoints on nearly all of the individual stones, and then refined those positions with waypoint averaging on multiple visits. In cases like this I would use the source tag on the way so that others have a very good chance of seeing it and respecting the previous work rather than simply changing it to what they think it should be. It is too easy to over look hard work that may have gone into establishing data. A single GPS trace is fine if that is all there is, better to average many GPS traces, in some locations I have 50+. ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments? Vandalism.
On Mon, Mar 16, 2020 at 04:18:02:PM +, Andy Townsend wrote: > On 16/03/2020 15:36, ael wrote: > > > > There has now been had one short reply essentially admitting tagging for > > the renderer. I haven't replied as yet, but Andy has. > > In this case it looks like the offending data's been removed, though the > tiles haven't rerendered yet (due to the site being busy). If there are I have only just got around to looking in more detail, and discovered that it is much worse than I had realised: vandalism. I have taken waypoints on nearly all of the individual stones, and then refined those positions with waypoint averaging on multiple visits. I now discover that this user has deleted that information and put new points from who knows where. I started replying to the changset comment before I had discovered that horror, and possibly failed to remain polite after it came to light. So it is vandalism. I will see what reply I get. Meanwhile I think everyone should check his other edits to historic sites. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments?
On 16/03/2020 15:36, ael wrote: There has now been had one short reply essentially admitting tagging for the renderer. I haven't replied as yet, but Andy has. In this case it looks like the offending data's been removed, though the tiles haven't rerendered yet (due to the site being busy). If there are possible problems elsewhere I'd suggest asking explicitly in the relevant changeset, or perhaps asking in this changeset if the same thing has happened and needs fixing somewhere else. If it needs the DWG to pick it up then drop a mail to d...@osmfoundation.org with as much detail as possible*, although this user does seem to be responsive to changeset comments. Best Regards, Andy (from the DWG) * if that hasn't happened already of course. I can't see it any a queue anywhere, but someone else may have picked it up and closed it already. ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments?
On Mon, Mar 16, 2020 at 04:08:45PM +0100, Mateusz Konieczny wrote: > > > > Mar 15, 2020, 22:36 by witwa...@disroot.org: > > > On Sun, Mar 15, 2020 at 09:18:59PM +, David Woolley wrote: > > > >> On 14/03/2020 18:09, ael wrote: > >> > I have just noticed some new "Embankments" added around a fortnight ago. > >> > These were added to some stone circles in Cornwall which I know well and > >> > have extensively surveyed. There is no trace of any embankments. No > >> > source was given and the user does not appear to be local. > >> > >> I suspect this is a case of tagging for the renderer, which is, of course, > >> wrong. > >> > > > > That is what I thought, and mentioned in the changeset comment. I have > > had no response as yet from the user. > > > > A *very* brief look at those recent edits suggests this might be a world > > wide problem, and may need a mass revert. There are other changes with > > no source given, and I wonder if copyright material has been used. > > > > Given no reply and confirmed tagging for renderer reverting all similar edits > should be acceptable (and likely - desirable). There has now been had one short reply essentially admitting tagging for the renderer. I haven't replied as yet, but Andy has. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments?
Mar 15, 2020, 22:36 by witwa...@disroot.org: > On Sun, Mar 15, 2020 at 09:18:59PM +, David Woolley wrote: > >> On 14/03/2020 18:09, ael wrote: >> > I have just noticed some new "Embankments" added around a fortnight ago. >> > These were added to some stone circles in Cornwall which I know well and >> > have extensively surveyed. There is no trace of any embankments. No >> > source was given and the user does not appear to be local. >> >> I suspect this is a case of tagging for the renderer, which is, of course, >> wrong. >> > > That is what I thought, and mentioned in the changeset comment. I have > had no response as yet from the user. > > A *very* brief look at those recent edits suggests this might be a world > wide problem, and may need a mass revert. There are other changes with > no source given, and I wonder if copyright material has been used. > Given no reply and confirmed tagging for renderer reverting all similar edits should be acceptable (and likely - desirable). ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments?
On Sun, Mar 15, 2020 at 09:18:59PM +, David Woolley wrote: > On 14/03/2020 18:09, ael wrote: > > I have just noticed some new "Embankments" added around a fortnight ago. > > These were added to some stone circles in Cornwall which I know well and > > have extensively surveyed. There is no trace of any embankments. No > > source was given and the user does not appear to be local. > > I suspect this is a case of tagging for the renderer, which is, of course, > wrong. That is what I thought, and mentioned in the changeset comment. I have had no response as yet from the user. A *very* brief look at those recent edits suggests this might be a world wide problem, and may need a mass revert. There are other changes with no source given, and I wonder if copyright material has been used. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
Re: [Talk-GB] Ficticious embankments?
On 14/03/2020 18:09, ael wrote: I have just noticed some new "Embankments" added around a fortnight ago. These were added to some stone circles in Cornwall which I know well and have extensively surveyed. There is no trace of any embankments. No source was given and the user does not appear to be local. I suspect this is a case of tagging for the renderer, which is, of course, wrong. ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
[Talk-GB] Ficticious embankments?
I have just noticed some new "Embankments" added around a fortnight ago. These were added to some stone circles in Cornwall which I know well and have extensively surveyed. There is no trace of any embankments. No source was given and the user does not appear to be local. I have added a changeset comment. https://www.openstreetmap.org/changeset/81640861 I looked at the user's recent changesets, and saw similar additions to other ancient monuments: the top of the list was the Rollright Stones. I don't know the Rollright stones well enough to know whether there is a real embankment there. But I suspect this is tagging for the renderer to make ancient monumnets more obvious on the standard map. I suggest that every one check their local historic sites for any unjustified modifications. ael ___ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb