Re: delete Primer section on Boolean-valued properties?

Dan Connolly wrote:

> With apologies for giving you conflicting advice...
> 
> On Tue, 2002-11-05 at 16:39, Frank Manola wrote:
> 
>>Section 4.3.2 of the Primer describes how to do Boolean-valued 
>>properties in RDF (because at the time RDF lacked any built-in values 
>>for TRUE and FALSE.)  The section was added because I was explicitly 
>>directed to cover it, but this was before we had datatypes.
>>
> 
> Yes... I might have asked you to do that.
> 
> I remember, now, that there are all these "how do I do XYZ?"
> questions among RDF users, and it would be nice for the
> primer to answer them. (hence my offer to clarify
> rdf:Alt, which I gather you ended up dealing with
> without my help.)


Well, you might want to have a look at whether it's clarified 
satisfactorily or not (Brian has made some further wording suggestions).


> 
> So while I'd like to minimize the explanation of generalities
> and background material in the primer, I support explaining
> how users should go about specific tasks.


Yes, the boolean section would make sense as part of a section like that 
  ("RDF Design Patterns", or something);  it's just in a sort of 
"miscellaneous" section now.


> 
> I think 2 or 3 examples of how to deal with boolean characteristics
> is in order... one using a class, say GoodThing, but noting we
> don't have rdf:notType, so this doesn't give you BadThing for
> free... you might mention owl:complementOf in there...
> and another example using xsd:boolean.


This is sort of what the section does now (doesn't mention 
owl:complement of;  it does say you need a language like OWL with more 
power to express disjointness).


> 
> 
>> Now that we 
>>*do* have datatypes, the section looks a bit strange (you can just use 
>>xsd:boolean, a point I recently added to the section).  Brian has 
>>suggested this section be removed (which I thought about doing in the 
>>process of producing the current Primer version), but I was reluctant to 
>>do this on "editor's discretion" alone, because the section is only 
>>there because the WG asked for it.  If anyone has any good reasons for 
>>keeping it, let me know, otherwise I'd like to delete it.
>>
> 
> I don't mind deleting it too much.


Part of the reason it looks odd now is that it's alone in not describing 
a particular RDF feature, just describing how to use RDF to do 
something.  If it had "brothers", there would be more reason to keep it.


> 
> But while you're thinking about this sort of stuff...
> Do you have examples of dealing with n-ary relations?


Section 4.3.1 talks about this some, in the context of describing 
rdf:value.  You might want to see if this is the sort of thing you had 
in mind.


> 
> I saw recent discussion of things versus their names;
> I'm intersted to have that explained in the primer.
> 


What specific points do you have in mind?  There's a little bit in 
Section 2.3 (just above figure 3) that talks about the advantages of 
identifying the creator of the Web page by a URI, rather than using the 
literal "John Smith".  And there's some material about halfway through 
section 2.4 that talks about not (say) using people's email addresses as 
their URIs, but instead using a blank node as the subject of the various 
identifying properties that they may actually have (like an email 
address, an employee number, and so on).  But I think what you have in 
mind is a separate section on each of those topics, right?

--Frank

 


-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-875

Received on Tuesday, 5 November 2002 18:57:02 UTC