[topicmapmail] Why do we need <subjectIdentity> for Published Subjects?

[Marc de Graauw]

[ Lars Marius Garshol ]

>
> * Marc de Graauw
> |
> | Why exactly do we need the construct
> | <subjectIdentity><subjectIndicatorRef .../></subjectIdentity> for
> | Published Subjects? Can't we achieve exactly the same thing using
> | scope, a name and the Topic Naming Constraint?
>
> Actually, no.
>
> The <subjectIndicatorRef/> element above provides a subject identifier
> (a URI). This plays much the same role as a base name + scope, in that
> it identifies as subject, but it is different in that it refers to a
> subject indicator, a resource that explains what the subject is.
>
Right. That certainly is a difference. I never read it that way and thought 
subject identifiers were just meaningless strings, and URI's were used to make 
sure they are unique, much like URI's are used in XML Namespaces. In 13250 this 
was apparently not the case, as it says:

NOTE 7: Subject descriptors may be offline resources.

and I found nothing to the contrary about public subject descriptors.

XTM does say:

A published subject is any subject for which a subject indicator has been made 
available for public use and is accessible online via a URI.

So it is a requirement that there actually is something at the other end of the 
URI?

Example: In my InterOperability Topic Map (http://www.marcdegraauw.com/itm/) I 
needed unique identifiers for elements in B2B vocabularies. I established 
subject identity by taking the URI of the B2B vocabulary and the name of the 
element, for instance: http://www.xcbl.org/xcbl35/xcbl35.html#OrderNumber Note 
that these do not actually point to anything, they only make the elements 
unique. If I understand you correctly, this approach is wrong, at least if I 
wanted to use these as Published Subjects? (Just as an example, I am not really 
going to provide Published Subjects.) And is this approach allright if I just 
want to provide subject indicators (as opposed to Published Subject 
Indicators)? If not, then it seems to me that the scope + name approach is the 
better one in this case. I use http://www.xcbl.org/xcbl35/xcbl35.html as scope 
and OrderNumber as name.( The scope quite conveniently points to a resource 
that humans can read to see what this OrderNumber actually is, but it would 
probably be better to give the theme an occurrence that points to this 
resource.) So what is the best way to do this: use subject identities or name + 
scope?

What it boils down to is the question whether there is a fundamental difference 
between identities and unique names. (I know one: normally names would be 
something one understands, and identities could be either comprehensible or 
not, but this difference does not seem very fundamental because nothing would 
require an identity to be incomprehensible.)

I see another very practical advantage of scope + name: there hardly are any 
Topic Map Published Subjects available right now. If I use the scope + name 
approach I can use any vocabulary that provides unique names right away, wit  
hout having to wait until someone provides the Published Subjects I need.

> Beyond that the differences are practical. The URI approach has a
> number of practical and polictical advantages, while the name + scope
> approach causes great difficulties.
>

Strangely enough I find the first part of your argument more convincing than 
the last part ;-) What advantages? What difficulties?

Thanks Lars,

Marc