Re: Information resources?

> |> Assertions that http://xmlns.com/wordnet/1.6/Hoary_Marmot is a web
> |> page or has a particular creator or last modified date or
> |> what-have-you are inconsistent.
> |
> | How could web client software not have such assertions interally,
> | whether or not expressed as RDF?  My RDF web client software [2],
> [...]
> | Do you see a way around this problem, other than by having wordnet
> | switch to using some kind of indirection as I've suggested?=20
> 
> Yes. Change your code so that it does not implicitly make assertions
> about the things identified by URIs based on the metadata it gets back
> from doing retrieval operations.
> 
> At the very least, don't make those implicit assertions if what you
> get back is an RDF representation. Assume the RDF tells you everything
> that the publisher wished you to know about the resource. If you want
> to use the other metadata to make your application smarter about
> caching, by all means do so, but don't assume that because you got an
> expires header back that you can assert that the resource expires at
> the specified time.
> 
> Your system does not have to make those assertions and if I want to
> make those assertions, I can make them in a different way.

I'll think about it more, but it sounds like you're suggesting
crippling my application.  It got lots of good data from its HTTP
activities, which the user may want to know about, like when it
fetched this representation and the Last-Modified date according to
the server.  You're asking it not to store that data, I guess.

It could store it in RDF safely if it put the URI it used into a
string literal.   That would probably work.   

    http://xmlns.com/wordnet/1.6/Hoary_Marmot is a species, a subclass
      of Marmot, etc.

    "http://xmlns.com/wordnet/1.6/Hoary_Marmot" is web address, which
      we last dereferenced at time ____, and got a last-modify data
      of ___, etc.

Hrm.

> | Wordnet
> | (as currently published at http://xmlns.com/wordnet/1.6/) is one of
> | the few vocabularies with this problem; I expect to motivate DanBri
> | to fix it soon.=20=20
> 
> What fix do you propose? http://xmlns.com/wordnet/2.0#Hoary_Marmot is
> just a complete non-starter given that the representation is on the
> order of 14Mb.
> 
> I suppose http://xmlns.com/wordnet/2.0/Hoary_Marmot#entity is a
> possibility, but it sure is ugly.

I'm fond of SlashRedirection [1] for applications like this.  The
current URLs work fine -- just issue a redirect (preferably 303 See
Other, taking advantage of the actual if not necessarily intended
meaning of its documentation).  Everything at purl.org has a redirect
in there already, and it seems to work fine.

(It's still not perfect -- Patrick will complain about the extra
round-trip, and others about the challenges of setting up redirects,
but it works.)

      -- sandro

[1] http://esw.w3.org/topic/SlashRedirection

Received on Thursday, 9 September 2004 20:01:44 UTC