[OSM-talk] very long multiple int_name tags in Near East

Maarten Deen mdeen at xs4all.nl
Thu Nov 21 06:14:07 UTC 2013


On 2013-11-20 23:49, Florian Lohoff wrote:
> On Wed, Nov 20, 2013 at 01:13:34PM +0100, Martin Koppenhoefer wrote:
>> 2013/11/20 Peter Wendorff <wendorff at uni-paderborn.de>
>> I agree, but what will we do on a practical level? I checked a dozen
>> arbitrary nodes and apparently nobody has cleaned up any of these in 
>> the
>> past 4 years:
>> http://www.openstreetmap.org/browse/changeset/3118180
>> and we are talking about 24174 nodes.
>> 
>> What is the general opinion on putting all those int_names (ideally 
>> only
>> those with more than one value, but how do you check this?) into a new
>> temporary tag "fixme:int_name" like Peter suggested? Or are there 241
>> people interested in solving this manually, so everybody only has to
>> check/fix 100 items?
> 
> As the order on the int_names might have changed due to edits 
> afterwards
> i'd be very careful to convert them automatically to the
> name:<languagecode> version which is the right way to do it.
> 
> If one could reconstruct the order/tag from the original import and its
> the same all the time i'd vote for changing them.

You'd just need to get the first version of the nodes. That's not very 
hard to do.

> That nobody cared in 4 years is not an argument to delete it or change
> it.

Very true.

> Probably people have been happy with the results of nominatim
> which parses it?

Does Nominatim parse it? Or does it only parse the name:<lang> tags? I 
can't imagine it does something (certainly not something smart) with 
those concatenated int_name tags.

Maarten



More information about the talk mailing list