Re: SVG12: <absoluteIRI> etc

On Monday, April 18, 2005, 12:59:16 AM, Bjoern wrote:

BH> Dear Scalable Vector Graphics Working Group,

BH>   From http://www.w3.org/TR/2005/WD-SVGMobile12-20050413/linking.html
BH> section 14.1.1

BH> [...]
BH>   local IRI references, where the IRI reference does not contain an
BH>   <absoluteIRI> or <relativeIRI> and thus only contains a fragment
BH>   identifier (i.e., #<elementID>)) 
BH> [...]

BH> It is highly unclear what this means. The symbols <absoluteIRI> and
BH> <relativeIRI> are not defined anywhere and if e.g. irelative-ref is
BH> meant to be referred to (as defined in RFC 3987) the grammar would
BH> not work the way it is assumed to work. It is thus unclear whether
BH> http://example.org/ with "http://example.org/#x" is a "local IRI
BH> reference" or not.

We agree that the form of words used here was confusing. We now explain
things as follows:

 <p>SVG supports IRI references, both relative and absolute. However,
    some elements have restrictions on these IRIs, as noted in the table
    below. There are three restrictions:
    <ol>
      <li>Fragment identifiers are disallowed on some elements (they can
      only point to complete files)</li>
      <li>Some elements constrain the IRI to <em>same-document references</em> - those which 
        point into the same document tree as the one containing the link anchor. </li>
      <li>Others are restricted to
      <span class="SVG-Term">local IRI references</span>, which require
      no additional network access. There are two types of local IRI
      reference:
        <ol><li> IRIs that point into the same document tree</li>
        <li>data: IRIs (which, if XML, produce a different tree, but the data is
          already loaded as it is part of the IRI itself)</li>
        </ol></li>
    </ol>
    </p>

Same-document references are one class of constraint, and local IRI
references are another class (one which requires no further access once
the current document is loaded, but does not necessarily point into the
same document tree).

BH> Please change the draft to be clear about these matters. Further,
BH> same-document references (the proper term for "local IRI references",
BH> I suspect), in the context of xml:base processing have been source of
BH> much confusion and there are changes to RFC3986 on these matters which
BH> make this section further unclear. Please change the draft such that it
BH> uses proper terminology and clearly explains processing in context of
BH> xml:base.

We believe that we are now using the correct terminology and that
processing of relative IRIs with xml:base is therefore clear.

Please inform us within two weeks if this change does not satisfy your
concerns. However, if there is some additional subtlety in RFC3986 whose
import we have missed, we trust that you will supply further and more
specific elucidation.


-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG

Received on Thursday, 13 October 2005 18:53:44 UTC