Eric Bohlman wrote:
> As I see it, the big problem is that there's no fully-accepted standard for
> resolving public identifiers, whereas system identifiers that take the form of
> URLs have standard resolution mechanisms.
Sure, but what you buy with standard resolution, you surely lose in abstraction. I
consider a public identifier to be something similar to a namespace and although
applications do require different syntax to resolve them, I don't consider it to
be an overbearing burden. It's just one of the hoops that applications make you
> A couple people have implemented John Cowan's proposal
> (<URL:http://www.ccil.org/~cowan/XML/XCatalog.html>) for a stripped-down
> version of SOCAT (along with a parallel XML-based alternative syntax); I've
> done one for Perl (XML::Catalog) and Norman Walsh has done one for, I believe,
I think a mechanism based on XML syntax makes sense for resolving public
identifiers. I would use it along side or instead of SOCAT and OmniMark's library
I still don't understand why an abstracted name is deprecated in favour of a URL.
Surely having both available offers the best hope of a sucessful resolution? I
would think that in ascending order, you might expect the best results with:
a) Public and system,
b) Public only, then
c) System only
Certainly the first two require more setup than the third, but only by a couple of
Marcus Carr email: [log in to unmask]
Allette Systems (Australia) www: http://www.allette.com.au
"Everything should be made as simple as possible, but not simpler."