RE: [XInclude] Syntactically incorrect IRIs in href attributes

Norm Walsh reports that java.net.URL does not distinguish between
resource errors and syntax errors, which is one existence proof of a
popular component that exhibits this limitation.

Also, URI processors will not always be able to distinguish between
syntactically correct IRIs, and an unknown URI scheme, because unless
you understand the scheme you can't verify the syntax.

Does this explanation satisfy you that some implementation-dependence
cannot be avoided?

> -----Original Message-----
> From: public-xml-core-wg-request@w3.org [mailto:public-xml-core-wg-
> request@w3.org] On Behalf Of Jonathan Marsh
> Sent: Monday, July 12, 2004 3:33 PM
> To: public-xml-core-wg@w3.org
> Cc: Elliotte Harold
> Subject: FW: [XInclude] Syntactically incorrect IRIs in href
attributes
> 
> 
> [Elliote, the XML Core WG has moved to a public mailing list.  We hope
> this will increase the visibility of our process.]
> 
> In reviewing our final issues, I realized I missed Eliotte's objection
> to one of the resolutions we made (xi-2).
> 
> IIRC, it was Daniel that had layering problems with a fatal error.
Can
> we elaborate a bit more on the implementation strategy in question?
> 
> -----Original Message-----
> From: Elliotte Rusty Harold [mailto:elharo@metalab.unc.edu]
> Sent: Wednesday, June 09, 2004 4:39 PM
> To: Jonathan Marsh
> Cc: www-xml-xinclude-comments@w3.org
> Subject: RE: Syntactically incorrect IRIs in href attributes
> 
> At 2:54 PM -0700 6/9/04, Jonathan Marsh wrote:
> >This is a bit tricky, as the useful thing for authors is to flag the
> >error as early as possible, but that precludes some implementation
> >strategies.  So the WG agreed to add a statement to this effect:
> >
> >---
> >A syntactically invalid IRI <termref def="dt-must">should</termref>
be
> >reported as a <termref def="dt-error">fatal error</termref>, but some
> >implementations may find it impractical to distinguish this case from
a
> ><termref def="dt-resource-error">resource error</termref>.
> >---
> >
> 
> I don't really mid it being a fatal error. However, I dislike making
> this an implementation option. I prefer to lock down the behavior as
> much as possible for the sake of interoperability. Could you
> elaborate on the implementation strategies requiring this to be a
> fatal error would preclude? They're not obvious to me.
> --
> 
>    Elliotte Rusty Harold
>    elharo@metalab.unc.edu
>    Effective XML (Addison-Wesley, 2003)
>    http://www.cafeconleche.org/books/effectivexml
> 
>
http://www.amazon.com/exec/obidos/ISBN%3D0321150406/ref%3Dnosim/cafeaula
> itA

Received on Wednesday, 14 July 2004 14:20:39 UTC