[OSM-talk] Unrepentant Vandal

[Andy Townsend]

On 07/11/2015 10:36, Andrew Errington wrote:
> Hi all,
>
> Here is a link to a random point on a light rail system:
> http://osm.org/go/546Jvddtd--?m=
>
> Soon after it opened I travelled on it from end to end, collecting gps
> data and photos of all the station signs.  There are two railway
> lines, one in each direction, and I mapped them both carefully.
>
> Recently I discovered that someone had helpfully deleted one of the
> lines and tagged the other with tracks=2.  I really don't think this
> is acceptable.
>
> I found the changeset and asked the user who did it why they destroyed
> my work.  They replied:
> "The OSM wiki implies that a single way with tracks=2 is the preferred
> way of showing rail lines with two tracks. This was the method used
> most in S. Korea, I was attempting to create consistency."
>
> This is not actually true (and I double-checked the wiki, just in
> case).  I pointed this out but the user did not acknowledge this was a
> mistake, or offer an apology.

I'd definitely suggest that changeset discussions are the best place to 
have this sort of conversation.  That way, it's visible, so that other 
people can be aware of the problem (and also discussions in public 
tended to be conducted with more politeness).  I'm sure that they 
generally believed that they were doing the right thing, but didn't 
think through the implications of what they were doing on data 
consumers* and other mappers.

 From looking at their edit history, which appears to be wide-ranging, I 
suspect that they're a non-surveying mapper who may not actually have 
been to all of the places that they've edited.

>
> So, my question is, am I being unreasonable, or am I right to think
> this is unacceptable?  How can I guard against this?

There are a bunch of "who's been editing where" tools - one that's 
especially worth mentioning is 
http://wiki.openstreetmap.org/wiki/Quality_assurance#WhoDidIt which 
looks for changes in an area and can provide an RSS feed.  I also use 
ITO's OSM Mapper http://wiki.openstreetmap.org/wiki/ITO_World#OSM_Mapper 
.  That works a few days in arrears, but is very useful at helping you 
to visualise what has changed.

> I have no
> problem with people improving the map by improving the data, but I am
> starting to see a lot of deletions, incorrect tagging, and generally
> shoddy work appearing, especially in Korea where I have done a lot of
> original work.

There might be a couple of issues here.  One possibility is people using 
"QA tools" to identify "problems" and "fixing" them.  We've had a lot of 
issues in GB with this - people changing the tagging on oneway roads in 
carparks incorrectly when the _real_ problem was that not all of the 
roads had been mapped is one example that springs to mind.  Where a QA 
tool identifies a problem, it should really be a prompt to carry out a 
local survey rather than apply a "remote guess" of what might be wrong.  
If I spot a problem like that I'd usually add to the changeset 
discussion of the original mapper, or add an OSM note or a fixme, to try 
and get it looked at properly.  Sometimes not of this works and there 
really are no local mappers, and the problem is bad enough that a remote 
fix really is needed (perhaps a newbie has broken a major road by 
mistake), but it's surely best that edits are done by people who either 
actually are there, or at least have been there.  Previous OSM tags + 
imagery don't always give a full sense of what something actually is.

Related to this is people "correcting" tags that are "wrong".  Often a 
"wrong" tag is a really useful indicator that an inexperienced mapper 
has been active, there are other things besides the "wrong" tag that 
might been checking, and the QA report is a useful indicator of this.  
Fixing the "wrong" tag removes the QA report but leaves the other data 
that doesn't match the real world in OSM.

The other possibility to people causing problems using QA tools is 
actually good news - lots of new mappers!  People new to OSM will get 
things wrong, whichever editor they're using, and any attempt to get 
them to trudge through the mire that is the current "beginners' guide" 
in the wiki will cause many to stop mapping before they've even 
started.**.  New mappers need help and understanding rather than "you've 
done it wrong!".  Often asking "what are you trying to map - how can I 
help?" is a good way of getting to the root of the problem.  I'd also 
give new mappers a week or more to "find their feet" - I suspect that 
the feeling of being watched would put some off too.

> Do I have to set up some kind of watch on all of my
> contributions and check them if someone edits them?

I did actually used to do this - I postprocess OSM data before using 
mkgmap to create a Garmin map, and one of the things that that did was 
to enable a list of "things edited since I last edited them" to be 
produced.  It's not something I've looked at of late though, since most 
edits are perfectly valid and there are plenty of other ways of 
identifying problems introduced by new (and old) mappers.

One other thing worth mentioning is whenever I've gone back to resurvey 
something that has been "questionably edited" I've almost always missed 
something that's either changed or that I missed the first time.  I can 
think of more than one example of someone changing "my" X to "their" Y, 
and after I had gone and had a look I found that Y was actually correct now.

Cheers,

Andy (SomeoneElse)


* as a data consumer I don't really care what you're changing to what as 
long as I know it's changing, no information is lost, and it doesn't 
make it impossible to extract the information in the future.

** Compare http://wiki.openstreetmap.org/wiki/Beginners'_guide with 
http://wiki.openstreetmap.org/w/index.php?title=Beginners%27_guide&oldid=1140764 
to see the problem.  Thankfully the person responsible for those changes 
is no longer editing the wiki.