OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Resource Gloss (Human Readable)



Rick Jelliffe wrote:

>
> ...the people who want to retrieve a schema (i.e.
> the people who only want to provide a schema on the net, or the people who
> want the schema to be the first and fastsest related resource to be
> retrieved) will simply choose to ignore it: unless our system supports
> existing uses, ...

I can think of a number of ways that software systems use resource
directories while supporting existing uses and not imposing an efficiency
overhead.

For example: a system which really wants to be fast won't download a schema
over a perhaps unreliable net, rather such systems will prefetch and perhaps
precompile such schemata. Such a cache based solution does work especially
for relatively static documents as we expect RDDL documents to be. As I see
it, this is an implementation issue.

Even if a performance overhead were imposed in some cases, I'd rather trade
a fast system for a reliable system and human readable documentation
enhances reliability. It doesn't get much easier than plopping a directory
in a browser to see what it contains.

> So I think we need to supprot
>    document-(namespace URI)->XML
> Schema-(xl:resource)->ResourceDirectory->other resources
> as well as Tim's preferred
>    doumnet-(namespace RI)->ResourceDirectory->other resources including
XML
> Schema
>  and (to push the hobbyhorse)
>  socument-(namespaceURI)->implied ResourceDirectory-(namespaceURI)->XML
> Schema
>
> Which is not to say that Tim's prefeerred chain is not the best practise
one
> for public namespaces.

    The reason "Namespace" is not in "RDDL" is that resource directories are
themselves independent of namespace URIs. You can insert the XHTML based
RDDL document anywhere you have the term "ResourceDirectory" above and get
exactly the desired functionality.

    It does turn out that this does provide a solution for the otherwise
contensious and nasty problem of what a namespace URI ought resolve to.

    But furthermore, suppose a namespace URI is not resolvable by DNS e.g.
is a URN not a URL? An entity catalog with a NAMESPACE entry would support
resolution of such a URN into a RDDL document just as it supports resolution
of an FPI into a document.

-Jonathan