Re: Datatypes

Reto Bachmann-Gmür wrote:
> Garret Wilson wrote:
>   
>> It all depends on the use case. Javadoc was created to document
>> packages, methods, and members, so it prefers to reference
>> finer-grained things.
>>
>> In a semantic description context (such as in RDF), though, it makes
>> more sense to treat packages like namespaces, classes like resources,
>> and members like resource properties. 
>>     
> I think that in a semantic description context all relevant resources
> should be addressable, this is at least Classes and Methods, they are
> both appropriate subjects for adding properties describing them.
> Probably it would also be useful to have a Class (owl:Class rdfs:Class)
> for packages.
>   
>> That would give us something like this:
>>
>> <rdf:RDF
>>  xmlns:package="info:java/com/example/package#",
>>  xmlns:property="http://ploop.org/namespaces/property#"
>>     
>>  <package:Class>
>>    <property:member>My value assigned to
>> com.example.package.Class.member.</property:member>
>>  </package:Class>
>> </rdf:RDF>
>>     
> I'm not seeing what the literal "My value assigned to
> com.example.package.Class.member." is supposed to be,

You must have missed this line: "See http://www.ploop.org/ to see where 
I'm going with this."

> what if just want
> to express that a certain class has a certain method?

You'd have to do that in some other way. Like I said earlier, it all 
comes down to use cases. There will always be a certain level of 
granularity you'll have to choose for the URI fragment. What if you want 
to identify method arguments, for example?


> What happens if
> two methods have the same name (but a different signature)?
>   

You must have missed this line: "See http://www.ploop.org/ to see where 
I'm going with this."

> I don't think that it makes sense to model a method as a particular kind
> of property

You're sort of late, here, as the JavaBean specification has been out 
for quite a while. :) C# (following Delphi) and ECMAScript have perhaps 
more elegant ways of assigning functionality to properties assignment 
and retrieval. But I think it all comes down to your missing the line, 
"See http://www.ploop.org/ to see where I'm going with this."

> and a JavaClass as a class of resources:

Actually, that part seems completely natural to me. They're even both 
named "class," oddly enough.

> Why should
> different JavaClass-Resources have the same type of property if they
> happen to have a method with the same name (what's the rdfs:domain of
> property:member)?

Hey, why are we talking about methods, by the way? "Member" was the word 
being used, and that doesn't necessarily refer to methods. In fact, I 
wasn't even referring to member variables---I was referring to properties.

Read the JavaBean specification, and then read http://www.ploop.org/ . 
I'll bet that answers all your questions, but if you have more I'll be 
happy to address them.

Best,

Garret

Received on Wednesday, 12 September 2007 14:25:42 UTC