Re: DAML Level of Effort for FY03-FY05

Drew,

Since I was the initial proponent of daml:imports on the Joint
Committee, let me address this issue. You are absolutely correct that
the imports statement must be used. Simply refering to a namespace does
not include any ontology definitions. You must make the imports
statement explicit. Period.
The correct way to do this for instance data is:

<rdf:Description about="">
  <daml:imports
rdf:resource="http://www.daml.org/experiment/ontology/beta/elements-ont#"/>
</rdf:Description>

The problem with using RDF namespaces to decide which schemas are
relevant is that multiple files may contain different definitions about
the same URI. See the attached GIF for example. The URI for Dolphin has
additional definitions in two schemas, good-schema and bad-schema. These
definitions are simply rdfs:subClassOf statements that happen to have
orig-schema#Dolphin as their subject. The problem with simply using
namespaces is I can't say that my-doc agrees with the definitions of
Dolphin found in good-schema but not those found in bad-schema. This is
why daml:imports was an essential component of the language. Note that
this issue is also addressed in the Web Ontology Requirements document
[1].

Now this does mean that there will be some apparent redundancy between
the namespace declarations and the imports declarations in any DAML
documents. This is unfortunate, but is the result of XML namespaces
being defined solely to solve the namespace problem for languages before
the W3C was concerned with ontologies. The ideal solution would be to
have the imports introduce namespaces for each ontology (as is done with
most pre-Web ontology languages), but alas this solution would go
against the grain of pre-existing web standards.

Theoretically we could say that the use of a namespace implies that you
import that ontology, but this does have the issue (raised later in this
thread) of distinguishing ontologies from other XML namespaces. Also, it
would not allow someone to provide completely alternate definitions for
a common set of terms (whether this is a good thing or a bad thing, I'm
not sure). Regardless, there is nothing in the DAML+OIL specs that says
that such default import statements exist. Although some tools may
decide to implement this default, it is nonstandard, and is akin to to
having HTML tools that are very flexible about guessing what you mean
when you provide malformed HTML.

Jeff Heflin

[1]
http://www.w3.org/TR/2002/WD-webont-req-20020307/#goal-shared-ontologies

Drew McDermott wrote:
> 
> Hello, Dave --
> 
> In your capacity as keeper of the DARPA/SONAT ontologies, I have a
> question about the "instance data."  I am cc-ing www-rdf-logic because
> there seems to be a general issue here about how DAML datasets work.
> 
> These ontologies are found at
> 
> http://www.daml.org/experiment/ontology/
> 
> and I don't think they're sensitive (or they would be https: instead
> of http:).
> 
> The file af-a.daml is RDF, rather than DAML, and appears to consist of
> "instance data," not ontology.
> 
> Here is the beginning of that file:
> 
> <rdf:RDF
>   xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'
>   xmlns:rdfs='http://www.w3.org/2000/01/rdf-schema#'
>   xmlns:NS0='http://www.daml.org/experiment/ontology/beta/elements-ont#'
>  >
> 
>   <rdf:Description rdf:about='http://www.daml.org/2001/12/factbook/af.daml#A1202'>
>     <NS0:country rdf:resource='http://www.daml.org/2001/09/countries/fips#AF'/>
>     <rdf:type rdf:resource='http://www.daml.org/2001/12/factbook/factbook-ont#Port'/>
>     <rdfs:label>Kheyrabad</rdfs:label>
>   </rdf:Description>
> 
> My problem is this:
> 
> Wouldn't it be much more informative if this line
> 
>   xmlns:NS0='http://www.daml.org/experiment/ontology/beta/elements-ont#'
> 
> were supplemented by something like this:
> 
>   <daml:imports rdf:resource="http://www.daml.org/experiment/ontology/beta/elements-ont#"/>
> 
> But it's not clear where this would go.
> 
> Is it supposed to be the case that just mentioning an ontology's
> namespace means that the file uses that ontology?  I find that rather
> disconcerting.  For the time being, we'll make that assumption, but
> this seems like a bug (or gap) in the DAML spec.  There ought to be an
> official way to say "This RDF file is based on ontologies X, Y, and
> Z."
> 
>                                              -- Drew McDermott