Revisit meaning of "semantics"

[thclark]

Currently, the top-level meaning of semantics is:

Agreed upon, often hierarchical, vocabulary.

@kilojoules and I discussed here about turning it into the following:

Agreed upon, often hierarchical, set of terms and structures to describe data. This includes vocabulary, schema, [taxonomies][taxonomy.md], and [ontologies][ontology.md]

My comment was as follows:

I think the best thing to do would be stick to what we agreed in the metadata challenge for this.
Schema, taxonomies and ontologies use semantics, but this change suggests that they are semantics, breaking self-consistency.
I see what you're doing though; you're responding to our own definition "often hierarchical", because how could a vocabulary be hierarchical without it being a taxonomy?!
We should raise this in the next metadata challenge meeting.

We should revisit our top-level definition of semantics. For my money, we should just directly reference wikipedia's entry:

In computer science, the term semantics refers to the meaning of language constructs, as opposed to their form (syntax). According to Euzenat, semantics "provides the rules for interpreting the syntax which do not provide the meaning directly but constrains the possible interpretations of what is declared".[23]
[23] Euzenat, Jerome. Ontology Matching. Springer-Verlag Berlin Heidelberg, 2007, p. 36.

[kilojoules]

@time-trader Can you weigh in here?

[kilojoules]

The definition of syntax contrasts itself to semantics, and says that semantics is the meaning of data. I think that is correct. The meaning of data is understood through vocabulary, schema, and ontology. Or am I missing something?

It seems somewhat narrow to view the meaning of data as simply the definitions of the words used.

In the example I gave, I say you can't say whether or not a tomato is a vegetable before agreeing to semantics. In particular, I am thinking of a taxonomy (i.e., taxonomy of plants versus taxonomy of foods).