How verbose are you with your data?

Hi All,

I'm curious to know how verbose you are when it comes to publishing 
linked data, specifically in regards to setting the rdf:type of things 
we describe, and with consideration for what I can only call secondary 
resources, things that we aren't primarily describing.

An example may be (although please consider all data you publish):

  :me a foaf:Person; #typed
    foaf:knows x:joe .
  x:joe foaf:name "Joe" . #untyped, needs inferred or dereferenced


I'm asking because for those who are familiar with ontologies they use - 
knowing the domain and range of the common properties - then it may be a 
common pattern (even instinctively) to publish data with the 
consideration that consuming clients will have the same, or at least 
some, awareness of the ontologies too - i.e. can do some inferring.

I guess one could also include awareness for owl:sameAs relations, 
inverse functional properties, and perhaps the most common of all, 
publishing data without strict ^^ datatypes.

Thus, if data is being published in this way, and it's more than a minor 
edge case, then consideration for these factors may need to be added to 
clients sooner rather than later.

Many thanks in advance for any responses,

Nathan