Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
I'm writing with a proposed extension to this mechanical edit: adjusting the type and wood tags to current conventions. It's easy to do and saves future changeset churn. I'm also adding "denoation=historic" on certain trees, and removing "denotation=cluster" from things like water towers. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
On Thu, Mar 5, 2015 at 2:40 PM, Bryce Nesbitt wrote: > Ok, call for additional perhaps final comments on this worldwide > mechanical cleanup for the impacts of the denotation=cluster import. > Fire in the hole: the first changeset is 29373065 ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
Ok, call for additional perhaps final comments on this worldwide mechanical cleanup for the impacts of the denotation=cluster import. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
On Wed, Mar 4, 2015 at 11:31 AM, SomeoneElse wrote: > > Well if the aim was to just undo the original mechanical edits that would > of course still be an option - not by looking at "node last touched by" but > by looking at the node lists from the original mechanical edit changesets. UK mappers smartly reverted before it got messy. But a semi revert of a node later edited by someone else adds risk and trouble to what's a pretty simple situation: a tag that is so compromised whatever value it could have had, it does not now have. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
On 04/03/2015 19:20, Bryce Nesbitt wrote: Andy, It appears from an audit that there's been a modest amount of editing,and mechanical copying of the trees affected by the original import. In the UK that import was reverted, but worldwide it is messier. The denotation=cluster tag itself is problematic at best, and better methods of finding tree clusters have been demonstrated. Thus it really seems that the most pragmatic choice is a wholesale purge of the cluster value, and a selective (no touching manual mapping) purge of the fixme. I can certainly see arguments for that (I was around at the time the tag was introduced and never exactly understood what "denotation=cluster" was supposed to be for), but suspect that a consultation with some of the other mappers using it since would make sense. If you're getting rid of _all_ "denotation=cluster" worldwide then the wiki page that currently says "This was a mechanical edit based on proximity to other trees." needs changing. Limiting the purge to the nodes last touched by user Nop would, unfortunately, be half baked. The cluster value was created by Nop out of a disagreement with the concept of mapping individual trees. The problem spread from there. Well if the aim was to just undo the original mechanical edits that would of course still be an option - not by looking at " node last touched by" but by looking at the node lists from the original mechanical edit changesets. Cheers, Andy ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
On 04/03/2015 18:18, Bryce Nesbitt wrote: Ok, making a possibly final call for input on the proposed "cluster" mechanical edit: http://wiki.openstreetmap.org/wiki/Mechanical_Edits/Bryce_C_Nesbitt Just to be clear - you're only removing these tags where they were added by the original problematical mechanical edits, not where they have been manually added by other mappers? Cheers, Andy ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
Ok, making a possibly final call for input on the proposed "cluster" mechanical edit: http://wiki.openstreetmap.org/wiki/Mechanical_Edits/Bryce_C_Nesbitt ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
Yes, that tag sounds like it should be removed. -- John F. Eldredge -- j...@jfeldredge.com "Darkness cannot drive out darkness; only light can do that. Hate cannot drive out hate; only love can do that." Dr. Martin Luther King, Jr. On March 2, 2015 2:55:29 PM moltonel 3x Combo wrote: On 02/03/2015, Martin Koppenhoefer wrote: > He was interested in > "special" trees and was asuming that trees close to other trees were less > "special" (something I don't agree with per se, but in practice might have > worked back then, because the mappers mapping "special trees" were > typically mapping only those special trees, hence there was less > probability of other trees _mapped_ nearby, even if there were actual trees > in the real world). Ok, that's a reasonable intent. But not a reasonable method, because the heuristic is flawed, because "storing the result of an osm query in osm data" is bad practice, and because a list of "normal" trees is insanely harder to maintain than a list of "special" trees. So there's not much to redeem the tag AFAICS. I'm happy to see it deleted from objects, surely starting with that one import and then double-checking the other changesets. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
On 02/03/2015, Martin Koppenhoefer wrote: > He was interested in > "special" trees and was asuming that trees close to other trees were less > "special" (something I don't agree with per se, but in practice might have > worked back then, because the mappers mapping "special trees" were > typically mapping only those special trees, hence there was less > probability of other trees _mapped_ nearby, even if there were actual trees > in the real world). Ok, that's a reasonable intent. But not a reasonable method, because the heuristic is flawed, because "storing the result of an osm query in osm data" is bad practice, and because a list of "normal" trees is insanely harder to maintain than a list of "special" trees. So there's not much to redeem the tag AFAICS. I'm happy to see it deleted from objects, surely starting with that one import and then double-checking the other changesets. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
2015-03-02 18:41 GMT+01:00 moltonel 3x Combo : > Out of curiosity, how would you render denotation=cluster differently > than other denotations ? Automatically create a forest polygon around > them ? Render them narrower than "normal" trees ? Why ? I can see the > interest in rendering landmark and natural_monument more prominently, > but the usecase for cluster is much harder to define (and if it > exists, a spatial query would probably still be better ?). > the intention of the mapper who introduced it (by performing a database query and storing the results in the osm db, something you shouldn't do, we all agree) was to omit those trees in his renderings. He was interested in "special" trees and was asuming that trees close to other trees were less "special" (something I don't agree with per se, but in practice might have worked back then, because the mappers mapping "special trees" were typically mapping only those special trees, hence there was less probability of other trees _mapped_ nearby, even if there were actual trees in the real world). cheers, Martin ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
On 02/03/2015, Martin Koppenhoefer wrote: > I can imagine people using it to determine importance of trees for rendering. Out of curiosity, how would you render denotation=cluster differently than other denotations ? Automatically create a forest polygon around them ? Render them narrower than "normal" trees ? Why ? I can see the interest in rendering landmark and natural_monument more prominently, but the usecase for cluster is much harder to define (and if it exists, a spatial query would probably still be better ?). ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Talk-de] Formal proposal: mechanically reverting fixme=set␣better␣denotation / denotation=cluster
2015-03-01 22:04 GMT+01:00 Bryce Nesbitt : > and since nobody is maintaining the denotation tag. I do use the denotation tag for trees which I add manually. The "cluster" value is not in the set of values I typically use, still I can imagine people using it to determine importance of trees for rendering. cheers, Martin ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk