Re: [OSM-talk-ie] Is data being corrupted?

2018-01-22 Thread moltonel 3x Combo 
That changeset was a newbie's first, under my supervision. I was
making an introduction to OSM at my coworking space, which is located
right there by the train station. When reviewing the session's edits I
noticed the strange "delete one POI and edit another identical one"
behaviour but not that the car park had been broken as a result, and
mistook that for a POI de-duplication edit.

What I think happened:
* There was no duplicate POI as I initially thought.
* The user didn't intend to edit the POI; maybe he played with the
undo button a bit, but his changeset was about editing a completely
unrelated feature.
* For whatever reason iD deleted node 3754515929 (the original POI)
and assigned its tags and location to node 1165425462 (a corner of the
car park)
* I could blame an iD bug, or I could blame a total newbie selecting
two nodes and merging them by mistake and without noticing anything.
* User uploads his intended edit along with the iD data mixing.

I don't think there's anything to be gained in not cleaning up the
current data, so I'll do that as cleanly as possible. The database
isn't getting corrupted randomly. It could be an iD bug, but it's also
easily explained as a newbie boo-hoo.

Thanks for the heads-up.

On 22/01/2018, Marc Gemis  wrote:
> the last edit date of ways or relations are only updated when their
> tags or the list of their members change. Not when the members in the
> relation of way change location or tags.
> So the error you describe is only 6 days old, and Vincent de Philly
> wrote a changeset comment 5 days ago
> (https://www.openstreetmap.org/changeset/55504147)
>
> m.
>
> On Mon, Jan 22, 2018 at 10:53 AM, Colm Moore 
> wrote:
>> Hi,
>>
>> Please don't correct this error until we know what is happening. :)
>>
>> Of late, I've noticed some mapping errors, e.g. a node on a road that has
>> been dragged off-road by a few hundred metres. When I check the last edit,
>> it might not have been edited in 5 years. I thought it implausible that
>> such obvious errors would remain in place for so long. Indeed, I had
>> sometimes edited in the area and would have thought that I would have
>> spotted an obvious error (I readily admit I make errors also). Many of the
>> issues are thrown up by OSMI http://tools.geofabrik.de/osmi/
>>
>> Today, using OSMI, I see that the car park in Kilkenny train station is
>> wonky, but it hasn't been changed in two years. The boundary wall hasn't
>> been edited in 4 years. The thing is, I viewed that area recently (I don't
>> recollect why) and had to do a forced refresh of the browser (Ctrl F5) of
>> the area to make the background map match the object outline. Several
>> users have edited in the area in recent years and surely one of us would
>> have spotted the error. Interestingly, the car park is now attached to a
>> node for a shop.
>>
>> https://www.openstreetmap.org/way/216942338 Edited about 4 years ago by
>> Vincent de Phily Version #2 · Changeset #20131140
>>
>> https://www.openstreetmap.org/way/100847087 Edited about 2 years ago by
>> wheelmap_visitor Version #7 · Changeset #35421480
>>
>> https://www.openstreetmap.org/node/1165425462 Edited 6 days ago by Reilsky
>> Version #4 · Changeset #55504147
>>
>> --
>>
>> It may be justifiable for the last edit date not to be updated where an
>> object within a relation is edited, but the relation itself isn't edited.
>>
>> Any thoughts?
>>
>> Colm
>>
>> PS Thanks to OSMI people for the service they provide. :)
>>
>>
>> ---
>> Never doubt that a small group of thoughtful, committed citizens can
>> change the world. Indeed, it is the only thing that ever has. Margaret
>> Mead
>> ___
>> Talk-ie mailing list
>> Talk-ie@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk-ie
>
> ___
> Talk-ie mailing list
> Talk-ie@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-ie
>

___
Talk-ie mailing list
Talk-ie@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ie

Re: [OSM-talk-ie] Is data being corrupted?

2018-01-22 Thread Rory McCann 

On 22/01/18 10:53, Colm Moore wrote:

It may be justifiable for the last edit date not to be updated where

an object within a relation is edited, but the relation itself isn't edited.


Any thoughts?

That's exactly what happens.

When a way (or relation's) tags, or the list of nodes that make it up is
changed (or reorded) then the way (or rel.) gets a new version number
(and hence datetime). But if a node is moved, then that node gets a new
version number (and datetime), but it doesn't change the way.

Which can be a bit annoying if you want do real deep dives into the OSM
history, but it's not that big a problem.

Rory

___
Talk-ie mailing list
Talk-ie@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ie

Re: [OSM-talk-ie] Is data being corrupted?

2018-01-22 Thread Marc Gemis 
the last edit date of ways or relations are only updated when their
tags or the list of their members change. Not when the members in the
relation of way change location or tags.
So the error you describe is only 6 days old, and Vincent de Philly
wrote a changeset comment 5 days ago
(https://www.openstreetmap.org/changeset/55504147)

m.

On Mon, Jan 22, 2018 at 10:53 AM, Colm Moore  wrote:
> Hi,
>
> Please don't correct this error until we know what is happening. :)
>
> Of late, I've noticed some mapping errors, e.g. a node on a road that has 
> been dragged off-road by a few hundred metres. When I check the last edit, it 
> might not have been edited in 5 years. I thought it implausible that such 
> obvious errors would remain in place for so long. Indeed, I had sometimes 
> edited in the area and would have thought that I would have spotted an 
> obvious error (I readily admit I make errors also). Many of the issues are 
> thrown up by OSMI http://tools.geofabrik.de/osmi/
>
> Today, using OSMI, I see that the car park in Kilkenny train station is 
> wonky, but it hasn't been changed in two years. The boundary wall hasn't been 
> edited in 4 years. The thing is, I viewed that area recently (I don't 
> recollect why) and had to do a forced refresh of the browser (Ctrl F5) of the 
> area to make the background map match the object outline. Several users have 
> edited in the area in recent years and surely one of us would have spotted 
> the error. Interestingly, the car park is now attached to a node for a shop.
>
> https://www.openstreetmap.org/way/216942338 Edited about 4 years ago by 
> Vincent de Phily Version #2 · Changeset #20131140
>
> https://www.openstreetmap.org/way/100847087 Edited about 2 years ago by 
> wheelmap_visitor Version #7 · Changeset #35421480
>
> https://www.openstreetmap.org/node/1165425462 Edited 6 days ago by Reilsky 
> Version #4 · Changeset #55504147
>
> --
>
> It may be justifiable for the last edit date not to be updated where an 
> object within a relation is edited, but the relation itself isn't edited.
>
> Any thoughts?
>
> Colm
>
> PS Thanks to OSMI people for the service they provide. :)
>
>
> ---
> Never doubt that a small group of thoughtful, committed citizens can change 
> the world. Indeed, it is the only thing that ever has. Margaret Mead
> ___
> Talk-ie mailing list
> Talk-ie@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-ie

___
Talk-ie mailing list
Talk-ie@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ie

[OSM-talk-ie] Is data being corrupted?

2018-01-22 Thread Colm Moore 
Hi,

Please don't correct this error until we know what is happening. :)

Of late, I've noticed some mapping errors, e.g. a node on a road that has been 
dragged off-road by a few hundred metres. When I check the last edit, it might 
not have been edited in 5 years. I thought it implausible that such obvious 
errors would remain in place for so long. Indeed, I had sometimes edited in the 
area and would have thought that I would have spotted an obvious error (I 
readily admit I make errors also). Many of the issues are thrown up by OSMI 
http://tools.geofabrik.de/osmi/

Today, using OSMI, I see that the car park in Kilkenny train station is wonky, 
but it hasn't been changed in two years. The boundary wall hasn't been edited 
in 4 years. The thing is, I viewed that area recently (I don't recollect why) 
and had to do a forced refresh of the browser (Ctrl F5) of the area to make the 
background map match the object outline. Several users have edited in the area 
in recent years and surely one of us would have spotted the error. 
Interestingly, the car park is now attached to a node for a shop.

https://www.openstreetmap.org/way/216942338 Edited about 4 years ago by Vincent 
de Phily Version #2 · Changeset #20131140

https://www.openstreetmap.org/way/100847087 Edited about 2 years ago by 
wheelmap_visitor Version #7 · Changeset #35421480

https://www.openstreetmap.org/node/1165425462 Edited 6 days ago by Reilsky 
Version #4 · Changeset #55504147

--

It may be justifiable for the last edit date not to be updated where an object 
within a relation is edited, but the relation itself isn't edited.

Any thoughts?

Colm

PS Thanks to OSMI people for the service they provide. :)


---
Never doubt that a small group of thoughtful, committed citizens can change the 
world. Indeed, it is the only thing that ever has. Margaret Mead   
___
Talk-ie mailing list
Talk-ie@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ie