[OSM-talk] correct (scholarly) attribution?

[Tom Pfeifer]

On 17.05.2019 18:33, Greg Troxel wrote:
> Frederik Ramm <frederik at remote.org> writes:
> 
>> Hi,
>>
>> if someone writes a scientific paper and wants to reference an OSM data
>> set they used, what would be the correct way to do that? Typically such
>> mentions contain author and name of the work, and publication place and
>> year. Or maybe the web-like "retrieved on ..."?
> 
> I vote for not retrieved on but the date the data was extracted from the
> main database, looking like "openstreetmap.org vector map data,
> OpenStreetMap Contributors, extracted YYYY-MM-DD HHMM UTC".
> 
> My impression is that the publication date is for things where a journal
> publishes them, and perhaps self-published blog posts, but for something
> that is continuously modified, it doesn't really make sense.
> 
> It's also good to give a second citation to a journal article that gives
> an overview the project, and surely that exists by now.  But I think the
> most important thing is an unambiguous pointer to both OSM and the epoch
> of data.

I'd like to agree here, and to add a 'it depends': on what the paper is about.
A core issue in a scientific paper is about the repeatability of the described experiment.

Thus when the core message is that the author has measured 123456 km of motorways in country X,
she should describe the precise source and version of the data set used, and the method of counting.
Probably not in the literature reference, but in an appropriate section of the main text.

When, on the other hand, the paper is about the sociology of repetitive vandalism patterns,
the precise date of the database extract might be less relevant, and appropriate case studies would 
need to be cited.

tom