Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

I'm reviewing the fixes and adding changeset comments. The fixes are
mostly good. It seems Hannah is occasionally adding area=yes on
accident.

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

Hi, and thank you for the cleanup efforts!

Please note that in some cases, people added buildings on top of
buildings. Depending on your QA software, this can be hard to spot
because it does not show up in diffs.

> Have you already corrected or
> reverted some of the errors that were noted?

I reverted two serious cases:

mdavenport1985(uid 6744070) was adding nonsense:
https://www.openstreetmap.org/changeset/53501083
https://www.openstreetmap.org/changeset/53429550

The Data Working group stepped in to stop him:
https://www.openstreetmap.org/user/mdavenport1985/blocks

A_N_D_R_E_W(uid 6876174) was intentionally vandalizing:
https://www.openstreetmap.org/changeset/53347778

I also reverted one non serious case of buildings on top of buildings:

https://www.openstreetmap.org/changeset/53347357

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

[OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Donna Baranski-Walker]

Hi All,
Ray, Hannah, Ryan and I have been correcting and validating for the last
few hours to bring things back to good.  Have you already corrected or
reverted some of the errors that were noted?

We noticed and corrected the following errors:
- lines used as area markers
- regions were not tagged with anything (just areas, instead of buildings)
- buildings were missing
- random 'other' points

Please give us your feedback on what we did tonight.

We were impressed by some of the advanced mapping that was done before us.

What we learned from this:
-  Hold closed mapathons for the time being
-  Be sure to have a sign-up sheet that includes everyone OSM id name
-  Provide better training with a hands-on example
-  Give newcomers a standard first assignment and check it with them before
giving them a real assignment
- Just after the mapathon, have all the organizers head over to an
organizer's house or a coffeehouse to spend an hour or two working together
to verify
- We thought we asked for OSM expert mappers to join us but no one
responded.  Ryan sent a request message to talk...@openstreetmap.org, and
h...@openstreetmap.org  -  Is more notice needed?

Looking forward to your replies.

Sincerely,
Donna and our Mapathon Team

P.S.  a question:
  #3757 Jabal al Baba - there's a marker stating "demolition orders"... is
there a way to add notes?


-- Forwarded message --
From: Donna Baranski-Walker 
Date: Thu, Nov 2, 2017 at 4:26 PM
Subject: Re: [OpenStreetMap] Problem mapping in your HOT Palestine projects
To: bgirardot 
Cc: Ryan Gee , Mikel Maron ,
Ray Rischpater , Hannah Pan 


Hi Blake,

I am Ryan's supervisor at Rebuilding Alliance and I asked him to let me
reply since he was unable to join us for our mapathon on Oct. 28th. That
said, I've been out with a nasty cough for the past few days.  Please
pardon my delay in getting back to you.  I've read through many of the
messages - I signed up for OSM-Talk, but I'm not quite sure how to join the
conversation.

I can answer some questions, and I have other questions too.  I am copying
our recent intern, Hannah, who organized the Mapathon.  Hannah and I are
beginners and we've participated in (and I organized) Mapping Palestine
mapathons before, including one at RightsCon.  A friend and experienced
mapper, Ray Rischpater, joined us on Saturday and I'm copying Ray on this
message too.

Here's what we did:

1.  We held a mapathon on Oct. 28th and there were 40 or more attending.
Most were beginners - many were high school students, one was a 6th grader
(he was really good at this), one was in her 80's.  Some watched the 'how
to map on OSM' video - and many did not.   For future mapathons we will
need to do a much better job of instructing the participants on how to map.

2.  On OSM Talk, someone noted,

There also seems to be a pattern of people removing and readding
buildings or adding buildings on top of buildings. (Why are they doing
that? Perhaps they have an assignment of a minimum additions and it's a
way to cheaply meet some quota?)

I think this has to do with our custom data set.  Our partners in
Palestine are the Geospatial mapping team at the Ministry of Local
Governments.
They worked with Mikel at Mapbox to import a high res. custom data set
but the custom set if offset from the Bing map.
We instructed our mapathon mappers to correct the mapping done to date
by moving the boxes into place -- but if you are looking at the
outcome on
the standard Bing map, it will look offset once again, and you will
wonder why people are re-adding buildings.

Ryan has an important observation about this and a suggestion for
Mikel and the GeoMoLG team that could remedy this.
I sent that forward in a separate message.

3.  I want to bring our marathon organizing team together this weekend to
verify and make corrections.  Hannah, Ryan, and Ray might we meet on
Saturday or on Sunday late afternoon?

4.  The beginners on Rebuilding Alliance's team are going to be learning a
lot more about OSM now - and I want to invite more people to join
Rebuilding Alliance's mapathon team so that future US Mapping Palestine
mapathons are well-supported and seamless.

5.  We’re working with the GeoMoLG team to get ready for OSM Mapathons in
Palestine with the villagers participating.  If all goes well, our next US
Mapathons will be more accurate and the villagers can do verification and
naming, as well as mapping themselves.  Before we begin, we are seeking OSM
experts in Palestine who can meet with the GeoMoLG team and attend the
mapathons too.  A prof at Birzeit University has expressed interest -
hoping to be in touch with him soon.

6.  Now for my questions:  are you seeing any evidence of intentional
spamming?  Some of the high school students may have attended just for
community service points and were not 

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

On Sat, 2017-11-04 at 15:23 -0400, john whelan wrote:
> You need to be given permission but having editted on OSM is not a
> requirement.
> 
> Cheerio John
> 
> On 4 Nov 2017 3:20 pm, "Safwat Halaby"  wrote:
> 
> > Can anyone create HOT OSM tasks? Is there an entry barrier?
> > 
> > The creator of task http://tasks.hotosm.org/project/3441 only has 5
> > edits.
> > 
> > ___
> > talk mailing list
> > talk@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk
> > 

Thanks for the info!

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[john whelan]

You need to be given permission but having editted on OSM is not a
requirement.

Cheerio John

On 4 Nov 2017 3:20 pm, "Safwat Halaby"  wrote:

> Can anyone create HOT OSM tasks? Is there an entry barrier?
>
> The creator of task http://tasks.hotosm.org/project/3441 only has 5
> edits.
>
> ___
> 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] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

Can anyone create HOT OSM tasks? Is there an entry barrier?

The creator of task http://tasks.hotosm.org/project/3441 only has 5
edits.

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Andy Townsend]

On 31/10/2017 16:00, john whelan wrote:


Often they have no knowledge of OpenStreetMap or their local area.  To 
them it's just helping the Red Cross or whatever.


Everybody (unless they're "serving life without parole") has knowledge 
of _some_ local area.  If people really don't know what they're doing in 
OSM, maybe they could better help the Red Cross in some other way?


More generally, it seems to be quite cruel to take someone with no 
knowledge of the climate, landscape and architecture a foreign country 
and with little knowledge of geography and (in the most recent case) map 
scales, and expect them to make sensible additions to OSM.  It's not 
their fault that their contributions are net negative; they've been set 
a near-impossible task.  It's even worse if they're students and they're 
actually going to be assessed on the results.


Someone's creating these tasks and someone (not necessarily the same 
person) is assigning them to brand new mappers.  It seems a real shame 
that they're not bothering to look at what has worked in the past and 
learning from that, not asking for feedback or assistance as they go on 
and not helping HOT volunteers do the thing that they're trying to help 
with - the technical solutions (task manager etc.) are in place but the 
human bit's missing.


Best Regards,
Andy


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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[john whelan]

I don't think that Andy's suggestions will work very well in practice.

The reason is that a maperthon is often organised by someone such as a
teacher who is looked on as an authority figure so their instructions will
carry more weight with the new mappers and the mappers themselves often
take the view that you don't need to read instruction books before use.

I worked with half a dozen maperthon mappers from a US university once.
The person leading the maperthon had optimistic ideas on how much mapping a
new mapper could do or how much would need explaining.  I think I could
have done the mapping they accomplished in a third of the time it took me
to handhold them through the process.  They were what I term disposable
mappers.  They did their project and as far as I know never mapped again.

Often they have no knowledge of OpenStreetMap or their local area.  To them
it's just helping the Red Cross or whatever.  They have no concept of the
economic value of the map.  Nor of the idea that anyone else than the Red
Cross or whomever they are thinking of at the time will want to use it.
They probably have no idea how to respond to a changeset comment or even
what a changeset is.

I'm not sure if anyone has done a study on the economic impact of
OpenStreetMap but I might be interesting to see.

We know that some businesses use OSM to illustrate their location.  I
understand in Bangladesh one of Uber's competitors is paying people to map
in OSM so their drivers know where to go.

Blake mentioned the validation process in HOT.  I spend a fair chunk of
time validating.  It works best with one validator on a project.  You get
to know the mappers and giving them feedback helps motivate them.

Task Manager 3 is supposed to have some improvements over TM2 for
validators.  It looks promising but there are a number of issues that need
sorting out.

The process itself currently means someone with no experience can validate
others work. Fine except you end up with lots of green tiles but you don't
know if you can trust the mapping or the validation.

My personal view is HOT works best where there is little prior mapping and
the impact of adding fifty untagged ways is low.  In an area such as the
West Bank which is policitally sensitive precautions should have been taken
to ensure an experienced validation team was on hand that could verify the
mapping shortly after it was done. The better organised project managers do
have a validation teams for their projects.

HOT project managers are often enthusiastic but sometimes have little idea
of what works best or even how to map.  HOT is working hard to improve
this.  Typically they set a project up then move on.  One or two keep an
eye on their projects but my feeling is that is not the general rule.  They
often have no idea who is organising maperthons using their project.

Instead of map everything in the tile it more common these days to ask for
just buildings or just roads.  This recognises that its simpler to train
new mappers on one or two things than bring them up to speed quickly on
everything.

OSM does retain some of the HOT mappers so not all the work is lost.

As you can tell I'm not one of the world's enthusiastic optimists but it is
getting better slowly and one day my phone will stop correcting my words.

Cheerio John

On 31 Oct 2017 11:00 am, "Andy Townsend"  wrote:

> For info, here's an example changeset discussion comment I've just
> written.  Where new mappers are not quite getting the hang of things it'd
> be great if before someone leaps to the mailing list to either complain
> abot or defend HOT and HOT mappers they could offer them advice and help
> about how to get better:
>
>
> "Hello and welcome to OpenStreetMap!
> Looking at some of the changes here such as $example I think you may be
> "overinterpreting" the imagery somewhat.
> If you look at the different imagery choices available you can get a
> feeling for the scale quite well, and its clear that none or almost none of
> the buildings that you've added here actually are buildings (that's clear
> with both Bing and your custom imagery - in fact your custom imagery, if
> later, suggests that buildings shown on Bing have been removed).
> Assuming you're not in the West Bank at the moment, what I'd suggest is
> that you try mapping things more locally to you first - often it's useful
> to look at the imagery of a place you know well to see how certain features
> look from the air.  It can be difficult to tell roads from e.g. riverbeds
> if you're not careful.
> If you're in $country (guessing, based on your editor locale of $locale
> above) then there are lots of local groups both HOT-based and non-HOT based
> where you can discuss mapping; there are also mailing lists and other
> forums - https://wiki.openstreetmap.org/wiki/Contact has lots of details.
> Best Regards,
> Andy"
>
> It'd also be great to refer mappers to other learning resources like the
> wiki beginners' guide or 

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Andy Townsend]

For info, here's an example changeset discussion comment I've just 
written.  Where new mappers are not quite getting the hang of things 
it'd be great if before someone leaps to the mailing list to either 
complain abot or defend HOT and HOT mappers they could offer them advice 
and help about how to get better:



"Hello and welcome to OpenStreetMap!
Looking at some of the changes here such as $example I think you may be 
"overinterpreting" the imagery somewhat.
If you look at the different imagery choices available you can get a 
feeling for the scale quite well, and its clear that none or almost none 
of the buildings that you've added here actually are buildings (that's 
clear with both Bing and your custom imagery - in fact your custom 
imagery, if later, suggests that buildings shown on Bing have been removed).
Assuming you're not in the West Bank at the moment, what I'd suggest is 
that you try mapping things more locally to you first - often it's 
useful to look at the imagery of a place you know well to see how 
certain features look from the air.  It can be difficult to tell roads 
from e.g. riverbeds if you're not careful.
If you're in $country (guessing, based on your editor locale of $locale 
above) then there are lots of local groups both HOT-based and non-HOT 
based where you can discuss mapping; there are also mailing lists and 
other forums - https://wiki.openstreetmap.org/wiki/Contact has lots of 
details.

Best Regards,
Andy"

It'd also be great to refer mappers to other learning resources like the 
wiki beginners' guide or learnosm, but that can be a bit difficult since 
it's difficult to know what they've seen so far


Regards,
Andy


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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Blake Girardot]

Hi,

I spoke with the person responsible for the projects and they are out
sick but aware of this thread and as soon as they are back on their
feet this week, will get to working on the issue, it is a high
priority for them.

Regards,
Blake

On Tue, Oct 31, 2017 at 3:21 PM, Blake Girardot  wrote:
> Greetings,
>
> I have also reached out to the person responsible for the projects
> both originally when brought up and just again today, but not heard
> back from them yet. Should I not hear back from the project creator
> today or tomorrow, I will probably temporarily remove the projects and
> the person who created the projects permissions to manage projects on
> the HOT Tasking Manager until they check in and we can be sure there
> is a plan in place to make sure the mappers on the project are doing a
> good job and the data is good quality.
>
> We should also remember that people who create projects do not always
> have any control or input on who contributes to the projects they
> create, but they should obviously be concerned data quality either
> way.
>
> It should also be noted, that HOT has a two step process that reviews
> all mapping, we call it validation, and that is typically where these
> kinds of issues get identified and feedback to the mappers is made.
> Anyone can participate in the validation process and experienced OSM
> mappers and folks who like to help new mappers in OSM are encouraged
> to be a part of that process. It is not a perfect process, but we are
> always working to make that process go better and more quickly to try
> and catch stuff like this as soon as possible.
>
> And, finally, I will note that when this was first raised, I dove
> right in and started to look at the data and make corrections and give
> feedback. But as soon as I spent time making fixes just to hit
> conflicts on upload because people were reverting, I stopped.
>
> Respectfully,
> Blake
>
> On Tue, Oct 31, 2017 at 2:08 PM, Andy Townsend  wrote:
>> On 31/10/2017 12:43, Frederik Ramm wrote:
>>>
>>>
>>> Have we found out anything about this yet?
>>
>>
>> None of the users who have entries in changeset discussions have commented
>> in those discussions, including the person who created the task at HOT.  I
>> also don't see any attempt to tidy up the data after the event (but to be
>> fair it is only 2 days - I'd usually wait a few more days before assuming
>> that I won't get a reply).
>>
>> It'd be great if some of the people in this thread could volunteer to help
>> these mappers achieve what they are trying to achieve and learn how to edit
>> OSM productively.  Unfortunately the DWG can't be everywhere - we rely very
>> much on everyone else in the community to help new mappers.  As has been
>> said before most of the problems here (as elsewhere) are just new mapper
>> issues and are relatively easily resolved.
>>
>> Also, don't forget there are really good examples of how to do this
>> elsewhere - look at the work the Irish community has done with both remote
>> and local mapping in Lesotho, for example.
>>
>> Best Regards,
>>
>> Andy
>>
>>
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>
>
>
> --
> 
> Blake Girardot
> OSM Wiki - https://wiki.openstreetmap.org/wiki/User:Bgirardot
> HOTOSM Member - https://hotosm.org/users/blake_girardot
> skype: jblakegirardot
> Live OSM Mapper-Support channel - https://hotosm-slack.herokuapp.com/



-- 

Blake Girardot
OSM Wiki - https://wiki.openstreetmap.org/wiki/User:Bgirardot
HOTOSM Member - https://hotosm.org/users/blake_girardot
skype: jblakegirardot
Live OSM Mapper-Support channel - https://hotosm-slack.herokuapp.com/

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Blake Girardot]

Greetings,

I have also reached out to the person responsible for the projects
both originally when brought up and just again today, but not heard
back from them yet. Should I not hear back from the project creator
today or tomorrow, I will probably temporarily remove the projects and
the person who created the projects permissions to manage projects on
the HOT Tasking Manager until they check in and we can be sure there
is a plan in place to make sure the mappers on the project are doing a
good job and the data is good quality.

We should also remember that people who create projects do not always
have any control or input on who contributes to the projects they
create, but they should obviously be concerned data quality either
way.

It should also be noted, that HOT has a two step process that reviews
all mapping, we call it validation, and that is typically where these
kinds of issues get identified and feedback to the mappers is made.
Anyone can participate in the validation process and experienced OSM
mappers and folks who like to help new mappers in OSM are encouraged
to be a part of that process. It is not a perfect process, but we are
always working to make that process go better and more quickly to try
and catch stuff like this as soon as possible.

And, finally, I will note that when this was first raised, I dove
right in and started to look at the data and make corrections and give
feedback. But as soon as I spent time making fixes just to hit
conflicts on upload because people were reverting, I stopped.

Respectfully,
Blake

On Tue, Oct 31, 2017 at 2:08 PM, Andy Townsend  wrote:
> On 31/10/2017 12:43, Frederik Ramm wrote:
>>
>>
>> Have we found out anything about this yet?
>
>
> None of the users who have entries in changeset discussions have commented
> in those discussions, including the person who created the task at HOT.  I
> also don't see any attempt to tidy up the data after the event (but to be
> fair it is only 2 days - I'd usually wait a few more days before assuming
> that I won't get a reply).
>
> It'd be great if some of the people in this thread could volunteer to help
> these mappers achieve what they are trying to achieve and learn how to edit
> OSM productively.  Unfortunately the DWG can't be everywhere - we rely very
> much on everyone else in the community to help new mappers.  As has been
> said before most of the problems here (as elsewhere) are just new mapper
> issues and are relatively easily resolved.
>
> Also, don't forget there are really good examples of how to do this
> elsewhere - look at the work the Irish community has done with both remote
> and local mapping in Lesotho, for example.
>
> Best Regards,
>
> Andy
>
>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk



-- 

Blake Girardot
OSM Wiki - https://wiki.openstreetmap.org/wiki/User:Bgirardot
HOTOSM Member - https://hotosm.org/users/blake_girardot
skype: jblakegirardot
Live OSM Mapper-Support channel - https://hotosm-slack.herokuapp.com/

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Frederik Ramm]

Hi,

On 31.10.2017 14:50, Pierre Béland wrote:
> I suggest to transform this to an «Awesome OSM Geoweek» with focus on
> the quality of edits, quality of leadership, of coaching of new
> contributors. 

Yes, but we must be careful to explain what "quality" means to us.

For example, if someone traces a house, and they trace the roof and not
the footprint, and they just add building=yes instead of something more
specific, and they don't make it rectangular even if they are tracing in
an area where you can assume all buildings to be rectangular ... that is
in my opinion still a valuable contribution (on which they can improve
later if they stay with us and learn).

If, on the other hand, the object isn't even tagged building=*, or if an
orchard is traced and tagged as building=yes name=Orchard, or if a
building is traced across an existing other feature, or if existing
objects are accidentally deleted, re-tagged, or nodes moved around the
country, then the edit starts to become useless or damaging.

If you create a building or a POI in JOSM today, you will be greeted
with a giant input form on which you can add more details than anyone
can possibly know.

When we say "we want quality edits" then that must not be misunderstood
as requesting perfection, all form fields filled out (can horses use
this road... hm... we don't have horses in this country but I guess in
theory...) and one-centimetre tracing precision.

So we need to find a definition for a quality edit that is achievable
for newbie mappers, yet doesn't require cleaning up after. And then we
can say to people running mapathons: "You must ensure that your
participants are in a position to make quality edits". (Individual
participants might still make non-quality edits but if the person
running the mapathon has put them in a position to make quality edits
then it's the individual's fault for choosing not to...)

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Pierre Béland]

For the OSM Geoweek in november, Invitations are made for groups, universities, 
etc. to create their own Mapathon.
I suggest to transform this to an «Awesome OSM Geoweek» with focus on the 
quality of edits, quality of leadership, of coaching of new contributors.  
Otherwise, this will simply be a «Pulse of new contributors», editing for a few 
hours and leaving problems to others afterward and publication of «Big numbers 
of Contributors» that represent minimal impact on number of Contributions but 
significative Quality problems. 

The same with this task. The organizers should tell us what they do to correct 
the situation.
 
Pierre 
 

   Frederik Ramm wrote:
Maybe it is possible to do better QA, improve editing software, or
easily fix the broken data, as suggested in various posts in this
thread, but I'd like to find the root cause of this and work on that.

Why and how did one person or a group of people who apparently lacked
the capabilities to make this activity a success, start it in the first
place? What warnings, what training material, what message of caution
could have led them to seek advice from people with the relevant
experience - or what over-optimistic "everyone can do it, no training
required" message enticed them to carry on recklessly?

Bringing dozens of new people to OSM only to delete (or significantly
overhaul) 99% of their contribution later helps nobody; it causes
unnecessary work for experienced mappers (and the DWG), tarnishes the
reputation of HOT and OSM, and discourages these people (who would have
made a valuable contribution given proper training) from contributing
further.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

___
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] Sudden influx of bad HOTosm edits in West Bank

[Andy Townsend]

On 31/10/2017 12:43, Frederik Ramm wrote:


Have we found out anything about this yet?


None of the users who have entries in changeset discussions have 
commented in those discussions, including the person who created the 
task at HOT.  I also don't see any attempt to tidy up the data after the 
event (but to be fair it is only 2 days - I'd usually wait a few more 
days before assuming that I won't get a reply).


It'd be great if some of the people in this thread could volunteer to 
help these mappers achieve what they are trying to achieve and learn how 
to edit OSM productively.  Unfortunately the DWG can't be everywhere - 
we rely very much on everyone else in the community to help new 
mappers.  As has been said before most of the problems here (as 
elsewhere) are just new mapper issues and are relatively easily resolved.


Also, don't forget there are really good examples of how to do this 
elsewhere - look at the work the Irish community has done with both 
remote and local mapping in Lesotho, for example.


Best Regards,

Andy



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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Frederik Ramm]

Hi,

On 29.10.2017 19:42, Andy Townsend wrote:
> I'm sure that these new mappers mean well, but they
> clearly haven't understood the task that they've been asked to perform
> (or perhaps that task was somewhat inappropriate for brand-new users in
> a place as densely populated as the West Bank). Try and find out if
> they're volunteers (e.g. they're volunteered to give up their own time
> to help improve the mapping in this area) or if they've perhaps been
> told to "map 100 buildings as part of this HOT task" in order to gain a
> credit for part of a course they're doing.

Have we found out anything about this yet?

Maybe it is possible to do better QA, improve editing software, or
easily fix the broken data, as suggested in various posts in this
thread, but I'd like to find the root cause of this and work on that.

Why and how did one person or a group of people who apparently lacked
the capabilities to make this activity a success, start it in the first
place? What warnings, what training material, what message of caution
could have led them to seek advice from people with the relevant
experience - or what over-optimistic "everyone can do it, no training
required" message enticed them to carry on recklessly?

Bringing dozens of new people to OSM only to delete (or significantly
overhaul) 99% of their contribution later helps nobody; it causes
unnecessary work for experienced mappers (and the DWG), tarnishes the
reputation of HOT and OSM, and discourages these people (who would have
made a valuable contribution given proper training) from contributing
further.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

On Sun, 2017-10-29 at 15:47 -0700, Mark Wagner wrote:
> On Sun, 29 Oct 2017 22:15:18 +0200
> Safwat Halaby  wrote:
> 
> > - Adding closed ways with area=yes instead of building=yes, or with
> > no
> > tags at all
> 
> A closed way with "area=yes" is a *very* common newbie mistake with
> iD:
> the user traced an area, then forgot to tag it, or didn't realize
> they
> needed to apply additional tags.
> 

Yes but that's not the issue with those edits per se. For instance
adding area=yes (or building=yes) on pre existing buildings en masse is
not a very common newbie mistake. See my other messages for more
details

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[john whelan]

Actually area=yes in Africa isn't that bad.  In JOSM I quite often come
across a settlement with thirty buildings drawn but not tagged and its a
lot faster to select the area, validate it then add tags to thirty
buildings at once rather than map thirty buildings.  You do need to check
them visually but it is still fairly quick and easy to do.

Cheerio John

On 30 October 2017 at 04:52, Michał Brzozowski  wrote:

> Some time ago I suggested that the user should choose the preset and
> *then* draw a feature. Such reversal is logical, allows to disable
> incompatible geometry types (such as building=* on lines and points) and
> even opens the door for displaying a concise tip on how to draw the object
> based on preset chosen.
> From my experience - we have imported most of address points in Poland.
> People add POIs as names to existing address nodes without any
> shop/amenity/etc. tags. This is an obvious UX issue.
>
> Michał
>
> 30.10.2017 1:53 AM "Bryan Housel"  napisał(a):
>
>> Haha I promise I won’t be offended.  I welcome the criticism - this is
>> part of working on something that matters to a lot of people.
>>
>> Anyway, iD already does train the user in the walkthrough how to assign
>> tags, and iD does warn the user on the save screen if they are uploading
>> untagged features.
>>
>> So, what would you prefer iD do.. Just not upload untagged things?  We
>> could do this, and I don’t have any strong opinion one way or another.
>>
>> In the past, the thinking has been that it’s better to accept an
>> imperfect contribution than to turn away an imperfect contributor.   The
>> map is used for a lot of things these days - so maybe we should rethink
>> this as a community how we handle obviously bad edits.
>>
>> Bryan
>>
>>
>>
>>
>> On Oct 29, 2017, at 7:23 PM, Jo  wrote:
>>
>> Yes, it's amazing that after all these years of reporting it, that BUG in
>> iD still hasn't been resolved. But don't dare to complain, we might offend
>> Bryan.
>>
>> Polyglot
>>
>> 2017-10-29 23:47 GMT+01:00 Mark Wagner :
>>
>>> On Sun, 29 Oct 2017 22:15:18 +0200
>>> Safwat Halaby  wrote:
>>>
>>> > - Adding closed ways with area=yes instead of building=yes, or with no
>>> > tags at all
>>>
>>> A closed way with "area=yes" is a *very* common newbie mistake with iD:
>>> the user traced an area, then forgot to tag it, or didn't realize they
>>> needed to apply additional tags.
>>>
>>> --
>>> Mark
>>>
>>> ___
>>> talk mailing list
>>> talk@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk
>>>
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>>
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>>
> ___
> 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] Sudden influx of bad HOTosm edits in West Bank

[Michał Brzozowski]

Some time ago I suggested that the user should choose the preset and *then*
draw a feature. Such reversal is logical, allows to disable incompatible
geometry types (such as building=* on lines and points) and even opens the
door for displaying a concise tip on how to draw the object based on preset
chosen.
>From my experience - we have imported most of address points in Poland.
People add POIs as names to existing address nodes without any
shop/amenity/etc. tags. This is an obvious UX issue.

Michał

30.10.2017 1:53 AM "Bryan Housel"  napisał(a):

> Haha I promise I won’t be offended.  I welcome the criticism - this is
> part of working on something that matters to a lot of people.
>
> Anyway, iD already does train the user in the walkthrough how to assign
> tags, and iD does warn the user on the save screen if they are uploading
> untagged features.
>
> So, what would you prefer iD do.. Just not upload untagged things?  We
> could do this, and I don’t have any strong opinion one way or another.
>
> In the past, the thinking has been that it’s better to accept an imperfect
> contribution than to turn away an imperfect contributor.   The map is used
> for a lot of things these days - so maybe we should rethink this as a
> community how we handle obviously bad edits.
>
> Bryan
>
>
>
>
> On Oct 29, 2017, at 7:23 PM, Jo  wrote:
>
> Yes, it's amazing that after all these years of reporting it, that BUG in
> iD still hasn't been resolved. But don't dare to complain, we might offend
> Bryan.
>
> Polyglot
>
> 2017-10-29 23:47 GMT+01:00 Mark Wagner :
>
>> On Sun, 29 Oct 2017 22:15:18 +0200
>> Safwat Halaby  wrote:
>>
>> > - Adding closed ways with area=yes instead of building=yes, or with no
>> > tags at all
>>
>> A closed way with "area=yes" is a *very* common newbie mistake with iD:
>> the user traced an area, then forgot to tag it, or didn't realize they
>> needed to apply additional tags.
>>
>> --
>> Mark
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
>
>
> ___
> 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] Sudden influx of bad HOTosm edits in West Bank

[Jo]

You can also use OSM inspector to find untagged ways:

http://tools.geofabrik.de/osmi/?view=tagging=-1.25144=5.12425=13=0.63=ways_without_tags

and it's of course the better option instead of adding load to Overpass API.

Polyglot

2017-10-30 2:47 GMT+01:00 Jo :

> Untagged ways in a bbox:
>
> https://overpass-turbo.eu/s/sFU
>
> way({{bbox}})(if:count_tags()==0)->.w1;
> rel(bw.w1);way(r)->.w2;
> (.w1; - .w2;);
> (._; >;);
> out meta;
>
> Michael beat me to the answer for the other query. Way too many comes near
> as an answer.
>
> Jo
>
>
> 2017-10-30 2:01 GMT+01:00 john whelan :
>
>> Many of the untagged ways are from JOSM and that warns before uploading
>> as well.
>>
>> Out of curiosity does any one have a count of the number of untagged ways
>> and area=yes in the database?
>>
>> Thanks John
>>
>> On 29 Oct 2017 8:53 pm, "Bryan Housel"  wrote:
>>
>>> Haha I promise I won’t be offended.  I welcome the criticism - this is
>>> part of working on something that matters to a lot of people.
>>>
>>> Anyway, iD already does train the user in the walkthrough how to assign
>>> tags, and iD does warn the user on the save screen if they are uploading
>>> untagged features.
>>>
>>> So, what would you prefer iD do.. Just not upload untagged things?  We
>>> could do this, and I don’t have any strong opinion one way or another.
>>>
>>> In the past, the thinking has been that it’s better to accept an
>>> imperfect contribution than to turn away an imperfect contributor.   The
>>> map is used for a lot of things these days - so maybe we should rethink
>>> this as a community how we handle obviously bad edits.
>>>
>>> Bryan
>>>
>>>
>>>
>>>
>>> On Oct 29, 2017, at 7:23 PM, Jo  wrote:
>>>
>>> Yes, it's amazing that after all these years of reporting it, that BUG
>>> in iD still hasn't been resolved. But don't dare to complain, we might
>>> offend Bryan.
>>>
>>> Polyglot
>>>
>>> 2017-10-29 23:47 GMT+01:00 Mark Wagner :
>>>
 On Sun, 29 Oct 2017 22:15:18 +0200
 Safwat Halaby  wrote:

 > - Adding closed ways with area=yes instead of building=yes, or with no
 > tags at all

 A closed way with "area=yes" is a *very* common newbie mistake with iD:
 the user traced an area, then forgot to tag it, or didn't realize they
 needed to apply additional tags.

 --
 Mark

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

>>>
>>> ___
>>> talk mailing list
>>> talk@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk
>>>
>>>
>>>
>>> ___
>>> 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] Sudden influx of bad HOTosm edits in West Bank

[Jo]

Untagged ways in a bbox:

https://overpass-turbo.eu/s/sFU

way({{bbox}})(if:count_tags()==0)->.w1;
rel(bw.w1);way(r)->.w2;
(.w1; - .w2;);
(._; >;);
out meta;

Michael beat me to the answer for the other query. Way too many comes near
as an answer.

Jo


2017-10-30 2:01 GMT+01:00 john whelan :

> Many of the untagged ways are from JOSM and that warns before uploading as
> well.
>
> Out of curiosity does any one have a count of the number of untagged ways
> and area=yes in the database?
>
> Thanks John
>
> On 29 Oct 2017 8:53 pm, "Bryan Housel"  wrote:
>
>> Haha I promise I won’t be offended.  I welcome the criticism - this is
>> part of working on something that matters to a lot of people.
>>
>> Anyway, iD already does train the user in the walkthrough how to assign
>> tags, and iD does warn the user on the save screen if they are uploading
>> untagged features.
>>
>> So, what would you prefer iD do.. Just not upload untagged things?  We
>> could do this, and I don’t have any strong opinion one way or another.
>>
>> In the past, the thinking has been that it’s better to accept an
>> imperfect contribution than to turn away an imperfect contributor.   The
>> map is used for a lot of things these days - so maybe we should rethink
>> this as a community how we handle obviously bad edits.
>>
>> Bryan
>>
>>
>>
>>
>> On Oct 29, 2017, at 7:23 PM, Jo  wrote:
>>
>> Yes, it's amazing that after all these years of reporting it, that BUG in
>> iD still hasn't been resolved. But don't dare to complain, we might offend
>> Bryan.
>>
>> Polyglot
>>
>> 2017-10-29 23:47 GMT+01:00 Mark Wagner :
>>
>>> On Sun, 29 Oct 2017 22:15:18 +0200
>>> Safwat Halaby  wrote:
>>>
>>> > - Adding closed ways with area=yes instead of building=yes, or with no
>>> > tags at all
>>>
>>> A closed way with "area=yes" is a *very* common newbie mistake with iD:
>>> the user traced an area, then forgot to tag it, or didn't realize they
>>> needed to apply additional tags.
>>>
>>> --
>>> Mark
>>>
>>> ___
>>> talk mailing list
>>> talk@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk
>>>
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>>
>>
>> ___
>> 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] Sudden influx of bad HOTosm edits in West Bank

[Michael Bemmerl]

Pierre Béland schrieb:
> Bryan Housel wrote:
> > Many of the untagged ways are from JOSM and that warns before
> uploading as well.
> and
> John Whelan wrote:
> > Many of the untagged ways are from JOSM and that warns before
> uploading as well.
>
>
> It is always possible to query Overpass and find such problems.

This query returned nearly 21.000 ways in Western Europe:

way["area"="yes"]({{bbox}})(if:count_tags() == 1);

Regards,
Michael

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Pierre Béland]

Bryan Housel wrote:> Many of the untagged ways are from JOSM and that warns 
before uploading as well.and
John Whelan wrote:> Many of the untagged ways are from JOSM and that warns 
before uploading as well.

It is always possible to query Overpass and find such problems.
But if applications such as iD and JOSM did add to the Changesets metadata a 
Warning tag, this could be used by applications such as OSMCHA to filter 
changesets.  
Something like warning=Orphan nodes; no tags; etc.
 
Pierre 

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Jo]

Next to the buttons for point, line and area, add a button to add
rectangular buildings. Let user click 3 times instead of 5 and tag it as
building=yes.

Polyglot

2017-10-30 1:50 GMT+01:00 Bryan Housel :

> Haha I promise I won’t be offended.  I welcome the criticism - this is
> part of working on something that matters to a lot of people.
>
> Anyway, iD already does train the user in the walkthrough how to assign
> tags, and iD does warn the user on the save screen if they are uploading
> untagged features.
>
> So, what would you prefer iD do.. Just not upload untagged things?  We
> could do this, and I don’t have any strong opinion one way or another.
>
> In the past, the thinking has been that it’s better to accept an imperfect
> contribution than to turn away an imperfect contributor.   The map is used
> for a lot of things these days - so maybe we should rethink this as a
> community how we handle obviously bad edits.
>
> Bryan
>
>
>
>
> On Oct 29, 2017, at 7:23 PM, Jo  wrote:
>
> Yes, it's amazing that after all these years of reporting it, that BUG in
> iD still hasn't been resolved. But don't dare to complain, we might offend
> Bryan.
>
> Polyglot
>
> 2017-10-29 23:47 GMT+01:00 Mark Wagner :
>
>> On Sun, 29 Oct 2017 22:15:18 +0200
>> Safwat Halaby  wrote:
>>
>> > - Adding closed ways with area=yes instead of building=yes, or with no
>> > tags at all
>>
>> A closed way with "area=yes" is a *very* common newbie mistake with iD:
>> the user traced an area, then forgot to tag it, or didn't realize they
>> needed to apply additional tags.
>>
>> --
>> Mark
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>
> ___
> 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] Sudden influx of bad HOTosm edits in West Bank

[john whelan]

Many of the untagged ways are from JOSM and that warns before uploading as
well.

Out of curiosity does any one have a count of the number of untagged ways
and area=yes in the database?

Thanks John

On 29 Oct 2017 8:53 pm, "Bryan Housel"  wrote:

> Haha I promise I won’t be offended.  I welcome the criticism - this is
> part of working on something that matters to a lot of people.
>
> Anyway, iD already does train the user in the walkthrough how to assign
> tags, and iD does warn the user on the save screen if they are uploading
> untagged features.
>
> So, what would you prefer iD do.. Just not upload untagged things?  We
> could do this, and I don’t have any strong opinion one way or another.
>
> In the past, the thinking has been that it’s better to accept an imperfect
> contribution than to turn away an imperfect contributor.   The map is used
> for a lot of things these days - so maybe we should rethink this as a
> community how we handle obviously bad edits.
>
> Bryan
>
>
>
>
> On Oct 29, 2017, at 7:23 PM, Jo  wrote:
>
> Yes, it's amazing that after all these years of reporting it, that BUG in
> iD still hasn't been resolved. But don't dare to complain, we might offend
> Bryan.
>
> Polyglot
>
> 2017-10-29 23:47 GMT+01:00 Mark Wagner :
>
>> On Sun, 29 Oct 2017 22:15:18 +0200
>> Safwat Halaby  wrote:
>>
>> > - Adding closed ways with area=yes instead of building=yes, or with no
>> > tags at all
>>
>> A closed way with "area=yes" is a *very* common newbie mistake with iD:
>> the user traced an area, then forgot to tag it, or didn't realize they
>> needed to apply additional tags.
>>
>> --
>> Mark
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
>
>
> ___
> 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] Sudden influx of bad HOTosm edits in West Bank

[Bryan Housel]

Haha I promise I won’t be offended.  I welcome the criticism - this is part of 
working on something that matters to a lot of people.

Anyway, iD already does train the user in the walkthrough how to assign tags, 
and iD does warn the user on the save screen if they are uploading untagged 
features.  

So, what would you prefer iD do.. Just not upload untagged things?  We could do 
this, and I don’t have any strong opinion one way or another.  

In the past, the thinking has been that it’s better to accept an imperfect 
contribution than to turn away an imperfect contributor.   The map is used for 
a lot of things these days - so maybe we should rethink this as a community how 
we handle obviously bad edits.

Bryan




> On Oct 29, 2017, at 7:23 PM, Jo  wrote:
> 
> Yes, it's amazing that after all these years of reporting it, that BUG in iD 
> still hasn't been resolved. But don't dare to complain, we might offend Bryan.
> 
> Polyglot
> 
> 2017-10-29 23:47 GMT+01:00 Mark Wagner  >:
> On Sun, 29 Oct 2017 22:15:18 +0200
> Safwat Halaby > wrote:
> 
> > - Adding closed ways with area=yes instead of building=yes, or with no
> > tags at all
> 
> A closed way with "area=yes" is a *very* common newbie mistake with iD:
> the user traced an area, then forgot to tag it, or didn't realize they
> needed to apply additional tags.
> 
> --
> Mark
> 
> ___
> talk mailing list
> talk@openstreetmap.org 
> https://lists.openstreetmap.org/listinfo/talk 
> 
> 
> ___
> 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] Sudden influx of bad HOTosm edits in West Bank

[Jo]

Yes, it's amazing that after all these years of reporting it, that BUG in
iD still hasn't been resolved. But don't dare to complain, we might offend
Bryan.

Polyglot

2017-10-29 23:47 GMT+01:00 Mark Wagner :

> On Sun, 29 Oct 2017 22:15:18 +0200
> Safwat Halaby  wrote:
>
> > - Adding closed ways with area=yes instead of building=yes, or with no
> > tags at all
>
> A closed way with "area=yes" is a *very* common newbie mistake with iD:
> the user traced an area, then forgot to tag it, or didn't realize they
> needed to apply additional tags.
>
> --
> Mark
>
> ___
> 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] Sudden influx of bad HOTosm edits in West Bank

[Mark Wagner]

On Sun, 29 Oct 2017 22:15:18 +0200
Safwat Halaby  wrote:

> - Adding closed ways with area=yes instead of building=yes, or with no
> tags at all

A closed way with "area=yes" is a *very* common newbie mistake with iD:
the user traced an area, then forgot to tag it, or didn't realize they
needed to apply additional tags.

-- 
Mark

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

On Sun, 2017-10-29 at 20:13 +0100, Blake Girardot HOT/OSM wrote:
> Greetings,
> 
> I SomeoneElse mentioned this in our HOT IRC channel.
> 
> I have already asked the project creator to take a look.
> 
> But, are you sure they are bad edits? did you use the 2016 imagery
> specified in the mapping projet?
> 
> For example, change set 53330616, the first one i randomly looked at
> from you list looks like bad editing until you use the 2016 imagery,
> please see the linked to two images below, one with Bing/DG Premium
> (they
> look the same) and one with the imagery supplied with the project.
> 
> So, please let us not rush to revert if you are not using the
> proper imagery.
> 
> As I said, I will or someone else from HOT will look into it further,
> thank you very much for bringing it to our attention, as well as the
> other issues you pointed out, which I have not had a chance to look
> at
> yet.
> 
> https://screenpresso.com/=l5lhb (old bing)
> 
> https://screenpresso.com/=TPYpg (new aerial)

I apologize for spamming the list with multiple split messages. Had
some technical issues. Last message in the series:

I'm afraid I suspect you're not using QA tools properly. Specifically
in JOSM, click the MIDDLE button to to reveal old buildings buried
beneath the newer buildings.

Also, JOSM will not show you when the users deleted buildings and then
re-added them. You'll only see the new buildings.

Also, these are not your typical newbie errors.

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Andy Townsend]

On 29/10/2017 21:33, Safwat Halaby wrote:

User "Andrew" added things like "world's biggest refrigerator", "WHEEL
OF FORTUNE", "ANDREWS BIGGEST CAR", "Concrete Thing", etc.


For completeness, http://overpass-turbo.eu/s/sFM shows "Andrew's Dream Car".

Reverting edits such as this makes sense to me.

Best Regards,
Andy


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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

On Sun, 2017-10-29 at 20:10 +0100, Blake Girardot HOT/OSM wrote:
> But, are you sure they are bad edits? did you use the 2016 imagery
> specified in the mapping projet?
> 
> For example, change set 53330616, the first one i randomly looked at
> from you list looks like bad editing until you use the 2016 imagery,
> please see the attached two images, one with Bing/DG Premium (they
> look the same) and one with the imagery supplied with the project.

This point has been repeated so I'll say one last time that this has
nothing to do with imagery alignment. It has to do with destroying
previous buildings and re-adding them, or adding new buildings on top
of them without destroying them, or other various weird edits.

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

On Sun, 2017-10-29 at 22:24 +0100, Blake Girardot wrote:
> Safwat,
> 
> You are not helping things by unilaterally reverting things and
> accusing people of vandalism after you asked HOT to look into it.
> 
> At this point all I have seen from your first list was some newbie
> edits with some newbie mistakes and some stuff that looked totally
> fine when using the correct imagery.
> 
> and before you started reverting all of the andrew users work, I saw
> decent mapping and bad tagging from a new mapper. clearly not
> vandalism.
> 
> Granted I am not the best at reviewing things by changeset so I might
> have missed something, that is totally possible.
> 
> At this point I am going to stop looking into it since you have
> decided to take matters into your own hands with your own solutions.
> 
> Regards,
> 
> Blake
> 
> 
> On Sun, Oct 29, 2017 at 9:15 PM, Safwat Halaby 
> wrote:
> > 
> > Due to a mailing client config error, I'm unsure if I sent my
> > messages privately or to the list (or at all). At the risk of
> > repeating
> > myself:
> > 
> > The problems are not related to Imagery at all. They're blatantly
> > obvious mapping issues like:
> > - removing things and re-adding them for no apparent reason
> > - Adding buildings on top of pre-existing buildings
> > - Adding closed ways with area=yes instead of building=yes, or with
> > no
> > tags at all
> > - Intentional vandalism in some cases.
> > 
> > ___
> > talk mailing list
> > talk@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk
> 
> 
> 


Dear blake, I'm afraid you're misunderstanding this.

I've not touched anything since contacting the DWG. I've done exactly 2
reverts prior to contacting them.

Revert 1:

The user has added duplicate buildings on top of already existing
buildings, except that those duplicate buildings had no building=yes
tag. I have not accused that user of vandalism and simply reverted. I
then realized this is not an isolated case, so left things for the DWG.

https://www.openstreetmap.org/changeset/53347357

Revert 2:
User "Andrew" added things like "world's biggest refrigerator", "WHEEL
OF FORTUNE", "ANDREWS BIGGEST CAR", "Concrete Thing", etc. This is
blatantly obvious, high priority, destructive vandalism. Sure, the user
may have done some good changes in between, but why should I bother
spending time bisecting the edits of a vandalist? In cases like this,
you revert and ask questions later, and perhaps later extract the good
bits if you've got the time to work through the junk.

https://www.openstreetmap.org/changeset/53347778

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Blake Girardot]

Safwat,

You are not helping things by unilaterally reverting things and
accusing people of vandalism after you asked HOT to look into it.

At this point all I have seen from your first list was some newbie
edits with some newbie mistakes and some stuff that looked totally
fine when using the correct imagery.

and before you started reverting all of the andrew users work, I saw
decent mapping and bad tagging from a new mapper. clearly not
vandalism.

Granted I am not the best at reviewing things by changeset so I might
have missed something, that is totally possible.

At this point I am going to stop looking into it since you have
decided to take matters into your own hands with your own solutions.

Regards,

Blake


On Sun, Oct 29, 2017 at 9:15 PM, Safwat Halaby  wrote:
>
> Due to a mailing client config error, I'm unsure if I sent my
> messages privately or to the list (or at all). At the risk of repeating
> myself:
>
> The problems are not related to Imagery at all. They're blatantly
> obvious mapping issues like:
> - removing things and re-adding them for no apparent reason
> - Adding buildings on top of pre-existing buildings
> - Adding closed ways with area=yes instead of building=yes, or with no
> tags at all
> - Intentional vandalism in some cases.
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk



-- 

Blake Girardot
OSM Wiki - https://wiki.openstreetmap.org/wiki/User:Bgirardot
HOTOSM Member - https://hotosm.org/users/blake_girardot
skype: jblakegirardot
Live OSM Mapper-Support channel - https://hotosm-slack.herokuapp.com/

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

Due to a mailing client config error, I'm unsure if I sent my
messages privately or to the list (or at all). At the risk of repeating
myself:

The problems are not related to Imagery at all. They're blatantly
obvious mapping issues like: 
- removing things and re-adding them for no apparent reason
- Adding buildings on top of pre-existing buildings
- Adding closed ways with area=yes instead of building=yes, or with no
tags at all
- Intentional vandalism in some cases. 

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Nicolás Alvarez]

2017-10-29 16:13 GMT-03:00 Blake Girardot HOT/OSM :
> Greetings,
>
> I SomeoneElse mentioned this in our HOT IRC channel.
>
> I have already asked the project creator to take a look.
>
> But, are you sure they are bad edits? did you use the 2016 imagery
> specified in the mapping projet?
>
> For example, change set 53330616, the first one i randomly looked at
> from you list looks like bad editing until you use the 2016 imagery,
> please see the linked to two images below, one with Bing/DG Premium (they
> look the same) and one with the imagery supplied with the project.
>
> So, please let us not rush to revert if you are not using the
> proper imagery.
>
> As I said, I will or someone else from HOT will look into it further,
> thank you very much for bringing it to our attention, as well as the
> other issues you pointed out, which I have not had a chance to look at
> yet.
>
> https://screenpresso.com/=l5lhb (old bing)
>
> https://screenpresso.com/=TPYpg (new aerial)

Hmm those seem to be the same image. Did you upload the wrong screenshot maybe?

-- 
Nicolás

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Blake Girardot HOT/OSM]

Greetings,

I SomeoneElse mentioned this in our HOT IRC channel.

I have already asked the project creator to take a look.

But, are you sure they are bad edits? did you use the 2016 imagery
specified in the mapping projet?

For example, change set 53330616, the first one i randomly looked at
from you list looks like bad editing until you use the 2016 imagery,
please see the linked to two images below, one with Bing/DG Premium (they
look the same) and one with the imagery supplied with the project.

So, please let us not rush to revert if you are not using the
proper imagery.

As I said, I will or someone else from HOT will look into it further,
thank you very much for bringing it to our attention, as well as the
other issues you pointed out, which I have not had a chance to look at
yet.

https://screenpresso.com/=l5lhb (old bing)

https://screenpresso.com/=TPYpg (new aerial)

Regards,
Blake

On Sun, Oct 29, 2017 at 6:18 PM, Safwat Halaby  wrote:
> There's a sudden influx of bad edits by different users related to
> Palestine HotOSM tasks. I don't know why. Perhaps it's a poorly-trained
> group?
>
> Incomplete list of relevant hotOSM tasks: 3441, 3447, 3759, 3768
>
> Incomplete List of bad changesets:
>
> https://www.openstreetmap.org/changeset/53330583 (reverted)
> https://www.openstreetmap.org/changeset/53330562
> https://www.openstreetmap.org/changeset/53330596
> https://osmcha.mapbox.com/changesets/53330616
> https://osmcha.mapbox.com/changesets/53330636
> https://www.openstreetmap.org/changeset/53330694
> https://www.openstreetmap.org/changeset/53330695
> https://www.openstreetmap.org/changeset/53330707 (intersecting ways)
> https://www.openstreetmap.org/changeset/53330589
> https://www.openstreetmap.org/changeset/53330728
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk



-- 

Blake Girardot
Humanitarian OpenStreetMap Team

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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Andy Townsend]

On 29/10/2017 17:18, Safwat Halaby wrote:

There's a sudden influx of bad edits by different users related to
Palestine HotOSM tasks.


There's a "I've seen a problem; what should I do?" section on 
http://wiki.osmfoundation.org/wiki/Data_Working_Group which explains the 
steps to follow.  You've already been doing that (for which thanks), but 
I mention it for the benefit of anyone else reading this who's 
unfamiliar with the process.


Where these are new mappers I'd suggest a "hello and welcome" message 
that asks a bit more about how they're contributing and what the 
problems are with it (for example - 
https://www.openstreetmap.org/way/536519974 has no tags but appears to 
be tracing over https://www.openstreetmap.org/way/519139010 from 2 
months ago).  I'm sure that these new mappers mean well, but they 
clearly haven't understood the task that they've been asked to perform 
(or perhaps that task was somewhat inappropriate for brand-new users in 
a place as densely populated as the West Bank). Try and find out if 
they're volunteers (e.g. they're volunteered to give up their own time 
to help improve the mapping in this area) or if they've perhaps been 
told to "map 100 buildings as part of this HOT task" in order to gain a 
credit for part of a course they're doing.


I'd also (and this would greatly help both the new mapper and the DWG) 
point out specific issues in each changeset, such as "way xyz does not 
have any tags, and without any tags no-one trying to use OSM data will 
know what it is" or "building abc overlaps with another building, which 
clearly can't be the case in reality".


Often there are a mix of edits from well-meaning brand-new HOT mappers - 
some that need reverting immediately because they actually break stuff 
(e.g. inadvertant node drags), some that need tidying and some that are 
perfectly OK to leave.  Having changeset discussion comments explaining 
what is what makes any future post-newbie tidy-up much easier.


Best Regards,

Andy (DWG member)


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

Re: [OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

Here's some intentional vandalism (among the other horrible edits).

https://www.openstreetmap.org/user/A_N_D_R_E_W

There also seems to be a pattern of people removing and readding
buildings or adding buildings on top of buildings. (Why are they doing
that? Perhaps they have an assignment of a minimum additions and it's a
way to cheaply meet some quota?)

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

[OSM-talk] Sudden influx of bad HOTosm edits in West Bank

[Safwat Halaby]

There's a sudden influx of bad edits by different users related to
Palestine HotOSM tasks. I don't know why. Perhaps it's a poorly-trained 
group?

Incomplete list of relevant hotOSM tasks: 3441, 3447, 3759, 3768

Incomplete List of bad changesets:

https://www.openstreetmap.org/changeset/53330583 (reverted)
https://www.openstreetmap.org/changeset/53330562
https://www.openstreetmap.org/changeset/53330596
https://osmcha.mapbox.com/changesets/53330616
https://osmcha.mapbox.com/changesets/53330636
https://www.openstreetmap.org/changeset/53330694
https://www.openstreetmap.org/changeset/53330695
https://www.openstreetmap.org/changeset/53330707 (intersecting ways)
https://www.openstreetmap.org/changeset/53330589
https://www.openstreetmap.org/changeset/53330728

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