XKOS and guidelines for rendering of statistical classifications in RDF

[catecara]

It was noted that some properties introduced in the current draft could be reused by other vocabularies (cf https://github.com/linked-statistics/xkos/issues/26, https://github.com/linked-statistics/xkos/issues/29 https://github.com/linked-statistics/xkos/issues/32, https://github.com/linked-statistics/xkos/issues/69 ... )

I think an SKOS extension to express statistical classifications should only include properties that are unique to XKOS and cannot be rendered by any other well-known vocabulary. Next to the XKOS spec, a guideline document could be produced, to provide indication of how to use XKOS in conjunction with other vocabularies to express statistical classifications.

[nichtich]

The current draft tries to be both, an extension to SKOS for statistical classifications, and a general collection of extensions to SKOS. It would be better to split these two parts. At least it should be made explicit which extension are for statistical classifications in particular and which are more general.

[tfrancart]

• XKOS is clearly aiming at statistical classifications, as stated in the "Background and Motivation" section :

  Specifically, we address the use of the Simple Knowledge Organization System (SKOS), a LOD specification, to satisfy the requirements of classification systems and concept management in general for the statistical community.

• If some aspects of it are of general interests for others, that is good; but the primary use-cases remain on statistical classification modeling.

• Equivalences have been introduced with iso-thes;

• A best practice document will give example on using XKOS in conjunction with other vocabularies