Re: DQV ISSUE-187 - cardinality of the link between Metric and Dimension


On Sun, Sep 13, 2015 at 7:12 PM, Antoine Isaac <> wrote:
> 1. I am not certain we can rule out the possibility of frameworks that are
> made of non-disjoint dimensions but are nonetheless very useful. For
> example if a SPARQL endpoint has terrible uptime (say, only 50%), I'd be
> tempted to say that this is both a performance and an availability issue.
> Nandana said he had thought of examples, I hope he can provide some.

Regarding the metrics with multiple dimensions in the 2015-09-11 call, I
was referring to the LODQM model [1]. It says "The quality metrics for
assessing these quality dimensions are proposed at the second level of the
model, some of which are used for measuring two quality dimensions." Figure
1 illustrates this (page 5). 8 out of 32 metrics defined in the paper have
2 dimensions associated to them. However, I agree with Jeremy it might be
simpler to have one dimension for metric if one is designing a metrics
framework and so as we all agree it could good to recommend to do so.


We have to think about the flexibility and usefulness in both the quality
measure producers' point of view (e.g., data providers, third party
evaluation tools, etc.) and quality measure consumers' (e.g., data
consumers, dataset selection, ranking, or recommendation tools) point of
view as Antoine and Jeremy have already mentioned.

Best Regards,

Received on Tuesday, 15 September 2015 14:28:01 UTC