Re: Domain of schema:numberOfEmployees?

I agree with the simple fix.

However, as you mention this solution comes short when you need to qualify
the value, such as with temporal validity. This is my case. I have numbers
of employees that are reported to specific dates. Currently, I'm thinking
of abusing both Schema.org and Data Cube Vocabulary to something like:

@prefix qb: <http://purl.org/linked-data/cube#> .
@prefix schema: <http://schema.org/> .
@prefix sdms-dimension: <http://purl.org/linked-data/sdmx/2009/dimension#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

[] schema:numberOfEmployees [
    a qb:Observation ;
    schema:value 42 ;
    sdmx-dimension:refTime "2000-01-01"^^xsd:date
  ] .

I believe this could be done with schema:QuantitativeValue (instead of
qb:Observation) if it had a way of expressing reference time (e.g.,
schema:referenceTime instead of sdmx-dimension:refTime).

Qualification of assertions (such as time-indexing) is a wider issue, so
I'd not delve into it now and instead go for the simple fix.

When it comes to validating the number of employees as you propose, I think
there are cases in which it's not that straightforward. For example, in the
case of the dataset I'm working on, I have both number of employees in
terms of persons and full-time equivalents (FTE).

- Jindřich

-- 
Jindřich Mynarz
http://mynarz.net/#jindrich

On Tue, Apr 28, 2015 at 1:14 PM, Dan Brickley <danbri@google.com> wrote:

> On 28 April 2015 at 10:41, Sarven Capadisli <info@csarven.ca> wrote:
> > On 2015-04-28 11:02, Jindřich Mynarz wrote:
> >>
> >> Hi,
> >>
> >> why is the domain of <http://schema.org/numberOfEmployees> set to
> >> schema:BusinessAudience? Shouldn't it be schema:Organization instead?
>
> Thanks both. This is worth addressing. I've filed an issue:
> https://github.com/schemaorg/schemaorg/issues/456
>
> There is a natural temptation to try to generalize this further, but
> (as I've argued in the issue tracker) things get complex quickly.
> Adding Organization as an appropriate type for numberOfEmployees is a
> quick and simple improvement.
>
> Incidentally it does also improve (very slightly) the possibility for
> verifying certain kinds of information, or at least probing for
> consistency. For example consider an Organization that has Ali, Bob
> and Carla listed as 'employee'. If we also had numberOfEmployees: 3,
> we would have some additional confidence that we have a complete
> description of the organization's employees (at that time, at least),
> whereas if we had been told numberOfEmployees: 2, we might instead
> care to re-evaluate what exactly the data was telling us. Since
> schema.org (and microdata/rdfa/json-ld) descriptions are always
> partial accounts of a larger reality, having these simple count
> properties can provide helpful checks.
>
> Dan
>
> >> Best,
> >>
> >> Jindřich
> >>
> >> --
> >> Jindřich Mynarz
> >> http://mynarz.net/#jindrich
> >
> >
> > Beside the point that the rdfs:comment only emphasizing on "business",
> > looking at all of the rdfs:subClassOf schema:Organizations, it appears
> to be
> > that "number of employees" is suitable for all of them.
> >
> > I agree that domain schema:Organization seems more appropriate. Thinking
> > ahead, are there organisations (currently undeclared) which do not have
> > "employees"?
> >
> > -Sarven
> > http://csarven.ca/#i
> >
>
>