From: "Andrew E Switala" <[log in to unmask]>
> The info URI scheme touts as a feature that info URIs are not URLs, so
> tools (like browsers) won't try to fetch an info resource as if it were
> a link, and info URIs can be managed independently of network resources.
> On the other hand, there is an advantage to URLs in that the online
> resource it references can describe what the URI is about.
Let's back up for a moment. The discussion is about whether we can safely
abandon enumerated lists in our schemas (Mods, Mads,...) and instead define
the elements/attributes that define them as AnyURI. (And I think we've
covered the advantages.)
Well, yes we can, provided we're comfortable that we have a method to
identify the values by URI. So the question becomes, "do we"?
The info:xv proposal is not a prescription for constructing URIs, it's one
way to do it. It's not intended that a MODS document be invalid if a URI
other than 'info' is used. (It's 'AnyURI' not 'InfoURI'.)
So, the philosophical question of whether it's better to have a uri that
resovles to something (or, the more abstract question of whether an 'info'
uri may indeed resolve to something or not), I think is out-of-scope.
If we agree that it's a good thing to eliminate the enumerated lists, and
that info:xv is *one* way to identify values, and if you want to use an HTTP
uri (either because you want your uri to resolve to a description, and you
don't believe that an info uri can do that; or because you believe the w3c
that the only good uri is an http uri) then you can coin your own http
identifiers.
--Ray
|