[Tagging] tag proposal "image=http:/... .jpg"
osm at tobias-knerr.de
Sat Feb 6 16:38:32 GMT 2010
Ulf Lamping schrieb:
> Am 05.02.2010 12:26, schrieb Tobias Knerr:
>> Sam Vekemans wrote:
>> The problem with this proposal is that there isn't a definition which of
>> the several images that likely exist for most objects should be
>> referenced. And I expect that it would hard to create one, as there
>> aren't any remotely objective criteria defining /the/ image for an
>> object - it can always only be /an/ image of the object.
> This misses the point. In Wikipedia there's also no "definition which of
> the several images that likely exist for most objects should be
> referenced", but Wikipedia has a lot of geo related articles with an image.
Imo, this isn't applicable for OSM because of two central differences
between OSM and Wikipedia:
1. OSM collects raw data. Wikipedia does not.
2. OSM has many different uses. Wikipedia has wikipedia.org.
Wikipedia primarily creates encyclopaedic texts (I'll ignore the lists
and tables here). This inherently means that whoever writes down
information will also have to decide - in a creative and somewhat
subjective process - how to present the information to the reader:
Choose wording and phrase sentences, create sections, determine the
sections' order, decide how long each section will be, create tables or
graphs to show certain trends, etc. Of course, adding images to
illustrate the text is also part of that process.
OSM delegates presentation to applications such as renderers: Only they
will filter information and determine visual appearance.
Wikipedia is clearly different from OSM in that there is exactly one
relevant way of viewing a Wikipedia entry: Visiting the page on
wikipedia.org. Everything else is an afterthought, used by a small
minority of Wikipedia visitors.
Therefore, every article on Wikipedia is optimized for presentation on
that web site: If the "OpenStreetMap" Wikipedia page should display a
logo at 64px width, then you'll find the file name and the "64px"
directly in the source code.
OSM, however, is not only about openstreetmap.org. We produce a database
that is hopefully useful for many different applications and web sites.
That a project like Wikipedia, which is mostly limited to a single
product, doesn't store information relevant for that product's
presentation separately isn't an argument for doing the same in OSM.
>> My opinion is that personal preferences like that shouldn't be part of
>> the OSM database. No "my favourite Sunday walk" route relations, no
>> subjective "food=extremely_tasty" for restaurants, and no "my favourite
>> image of the object" either.
> There's a big difference between the subjective "my favourite walk" and
> the objective "this is a picture of the object".
It's objective information that "this is a picture of the object". The
choice of the image is, however, highly subjective, as many other
pictures of the object exist.
Similarly, it's objective information that my favourite walk is a route
from X via Y to Z. It's subjective why, of all the possible routes, I
chose exactly this one.
More information about the Tagging