On tis, 2007-02-27 at 13:06 +0100, Thomas Fischer wrote:
> Hello everybody,
>
> as mentioned in my previous letter, I think that identifiers and resources
> are very closely related. And I agree with Alistair that we need "correct
> procedures for merging DCMI metadata".
> For this we need a method to identify resources if they appear in different
> descriptions, like knowing "do these statements refer to the same author?"
> From this perspective I have some problems with using URIs as the standard
> identifier in the DCAM. Now this is not new to the current draft, but
> thinking about some of its aspects I came across this point again. (There is
> another questions about the relationship to IRIs, I leave this aside.)
>
> The point is that URIs are fairly poor identifiers unless strengthened by
> some definition of equivalence, which is not precisely spelled out in RFC
> 3986, the most recent one on URIs that I know.
> You could try to settle for letter-for-letter identity, but this is not the
> way URIs are usually used.
>
> Otherwise, they are not unique, i.e. different URIs can be equivalent.
> The RFC does not give a clear definition of identity or equivalence of
> different URIs:
> "...determination of equivalence or difference of URIs is based on string
> comparison, perhaps augmented by reference to additional rules provided by
> URI scheme definitions. We use the terms "different" and "equivalent" to
> describe the possible outcomes of such comparisons, but there are many
> application-dependent versions of equivalence."
>
> For example:
> http://www.MathGuide.de/ is equal to
> http://www.mathguide.de/,
> http://www.mathguide.de and
> http://www.mathguide.de:80/.
No to all
> Is this equal to
> http://MathGuide.de/,
> [log in to unmask]" target="_blank">http:[log in to unmask] or
> http://www.mathguide.de/index.html?
No to all
> The latter is equal to
> http://www.mathguide.de/X/../index.html
> for any X without a slash.
Not in this context.
> On the other hand, not all URIs are useful as identifiers,
> http://localhost/myFile.html
> is one example.
This is certainly a valid URI. Let's not mix validity with usefulness...
>
> So, to merge DCMI metadata, which identifications are we allowed to use?
While not explicitly stated (maybe it should be?) it's a reasonable
assumption that the DCAM will use the same merging methodology as RDF,
i.e., merging URIs based on
http://www.w3.org/TR/rdf-concepts/#section-Graph-URIref
"Two RDF URI references are equal if and only if they compare as equal,
character by character, as Unicode strings."
subject to a number of encoding rules.
So, of course a certain level of care must be taken when designing/using
URIs.
Also, nothing forbids an application to make certain guesses regarding
equivalence beyond what is required by the RDF model.
/Mikael
> Thomas
>
--
<[log in to unmask]>
Plus ça change, plus c'est la même chose
|