- From: Gannon Dick <gannon_dick@yahoo.com>
- Date: Tue, 30 Sep 2014 12:11:18 -0700
- To: Frans Knibbe | Geodan <frans.knibbe@geodan.nl>, Andrea Perego <andrea.perego@jrc.ec.europa.eu>
- Cc: Ghislain Atemezing <auguste.atemezing@eurecom.fr>, "public-locadd@w3.org Mailing list" <public-locadd@w3.org>
Hi All, <frans.knibbe@geodan.nl> wrote: > > [snip] ... to try to model this in RDFS only? I would favour this option. [I do also, see below] --------------------------------------------------- BTW, another common way of specifying spatial resolution is by using an equivalent scale. If I'm not mistaken, this can be expressed in QUDT by using dimensionless quantities - in particular, quantity:DimensionlessRatio [1]. http://www.qudt.org/qudt/owl/1.0.0/quantity/Instances.html#DimensionlessRatio Another option is to define something like locn:equivalentScaleSpatialResolution, with xsd:integer as range. ---------------------------------------------------- It is fairly easy to demonstrate that the "DimensionlessRatio" is arbitrarily close to a constant (sqrt(2)/2) or the resonant (2/sqrt(2)). The other option, locn:equivalentScaleSpatialResolution, is better, I think. For Labels (RDFS), the set is square (RxC), the indexes on a spreadsheet are not however ((C-26)xC). This might lead you to believe that numbers derived from close packing of spheres (circles) means "something interesting" when in fact it means you have either 1) an incomplete set of labels or 2) some of the labels have alternate spelling. The "Standard" is "close packing of boxes". Horribly boring, BTW. You can reach the same conclusion from a statistical point-of-view by looking at the Probability Density Function. For an (RxC) set of labels, either the Mean moves or the Variance moves but you can't have both without 1) or 2) above. Labels do not inherit the Open World Assumption, although Values do (sleep well, Semantic Web). --Gannon Andrea
Received on Tuesday, 30 September 2014 19:11:46 UTC