[geotk] Is WKT always sufficient to define a CRS?

[Martin Desruisseaux]

Hello Jon

Le 02/08/11 15:17, Jon Blower a écrit :
> I’m wondering how coordinate reference systems may be serialized and 
> exchanged. Is WKT a good serialization format, i.e. can all CRSs be expressed 
> completely (and decoded therefrom)?
It is a good format for the most common CRS if you don't need metadata (area of 
validity, scope, accuracy, aliases, remarks). For less common CRS, it may fall 
short in the following cases:

  * Temporal CRS (however I think we could push for an official extension at OGC
    if there is interest for that)
  * Stereographic projections over poles having axes like "South along 90°
    meridian" (but extension coud also be considered)

> Are there any pros and cons in this respect of WKT vs Proj4 syntax?
>
I think that are pretty similar in fonctionalities. My guess:

Proj4 advantages:

  * More compact string
  * More reliable when the information come from the PostGIS "spatial_ref_sys"
    table. The WKT declared in PostGIS have errors, but those errors seem to be
    mostly unoticed or have low priority (the PostGIS team didn't replied to my
    bug report) because PostGIS doesn't use them - it uses only the column
    containing Proj4 parameter strings.

WKT advantages:

  * Standardized (however we have to nuance: there is different "flavors" of
    WKT. For example Oracle and ESRI WKT are slightly different than the
    standard one).
  * Maybe more flexibility in the way to declare projection parameters and axes
    order.


Regards,

Martin