[sc34wg3] Questions on N0396: (16) locators

[Lars Marius Garshol]

* Jan Algermissen
| 
| N0396 says in 3.1:
| 
| "A locator is a string that references one or more information resources"
| 
| The note at the end of 3.4.5 says:
| 
| "Locators which refer directly to subjects which are not information
| resources must be used with caution. They should not be used in the
| [subject addresses] property, as this is intended only for references
| to information resources. Rather, they should be placed in the
| [subject identifiers] property."
| 
| The two statements contradict each other.

They do, though I would perhaps rather say that the note expands on
the definition. It turns out that whether we want it or not there are
URIs that do not reference information resources.
 
| I strongly advice us NOT to try any 'fixups' towards
| interoperability with RDF. I think that topic maps are far better of
| if they stick to the idea that "a URI allways refers to the
| information resource that (might) be returned by an HTTP GET. There
| are no locators that refer to abstract concepts." Anything else is
| introducing evil, IMHO.

The trouble is that URIs were defined by the IETF and we can't change
them. It's a fact of life that there are quite a few URI schemes that
do not reference information resources and there's not a whole lot we
can do about it. It's impossible to outlaw them in any meaningful way,
and to pretend that the problem does not exist is the worst we can do.

Note that there are many other kinds of URIs than just "http://" ones.
We can't disallow ftp, nntp, gopher, news, mailto, ... URIs, nor the
more dubious ones like urn:isbn:..., tdb:..., and so on.

This is SAM issue locator-reference, BTW:

<URL: http://www.ontopia.net/omnigator/models/topic_complete.jsp?tm=tm-standards.xtm&id=locator-reference >

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >