On 2020-05-12 15:28, Jean-Marc Liotier wrote:
> On 5/12/20 2:52 PM, Colin Smale wrote:
>
>> As you and many others frequently remind us: OSM is first and foremost about
>> the data and not any specific use-case or rendering thereof.
> Yes - but a data model is not a neutral representation of reality: it is a
> projection through a use-case, mapping reality to a construct fit for
> specific purposes. In the relational database world, we cannot produce a
> satisfactory schema without knowledge of what sort of queries are intended.
> Flattening the highly dimensional reality into any data model involves such
> choices. Closer to the daily preoccupations of Openstreetmap, even lists of
> attribute values are reductionist and finding the appropriate tradeoff cannot
> be achieved isolatedly: it requires input from those who will deal with the
> consequences of the choices - the data-consuming users.
Absolutely, well put. Any kind of modelling requires reductions of
complexity, trade-offs and compromises. Which nuances are we going to
include, and which are we going to consciously choose to leave out? If
we don't leave anything out, we have not modelled reality, we have
duplicated it.
The role I expect of the data consumers is to articulate how they would
like to view the data (including what attributes they are expecting),
and not to dictate how that data is stored/represented internally.
Cartography, geography, statistics etc are very different skills to data
modelling and database design.
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk
