Andreas Trawoeger schrieb am 04.04.2012 15:16:
- Spatial Reference System Identifier (SRID [1]) definition is missing
Introducing a SRID field would also imply to implement coordinate transformation services behind WikiData and most likely would complicate data usability.
IMO it is sufficient to constrain geodata to WGS84. For most purposes (at least in Wikipedia) it should be accurate enough (± 5 m). We don't do land surveying here, right?
That means, we'll basically ignore influences from plate tectonics. Local or continent-related CRS like NAD83 (for North America) or ETRS89 (for Europe) have the advantage of being more stable pertaining to a particular plate. That's why we use them in Europe and North America. In a global CRS like WGS84 theoretically we have to update geo-coordinates due to crustal motion more often. But we speak of <20 m within 100 years (maybe more in some areas like Hawaii). Relevant to Wikidata? Probably not.
On the other hand there are some historic CRS (like NAD27, Gauss-Krüger). But - as John said - these geo-coordinates can be transformed into WGS84 (usually with an accuracy of <10 m).
In Wikipedia we also have some coordinates for other globes (Moon, Mars) and also celestial coordinates as Lat/lon. But I have no clue on what definitions those coordinates based on. Presumably it would be better to separate these things from geodata.
Regards Alex