Re: what "huge problem" with XML Base? [was: red/green XML]

Paul Grosso wrote:

Subject: what "huge problem" with XML Base? [was: red/green XML]


> At 10:03 2000 06 06 -0400, Tim Berners-Lee wrote:
> >From: Jonathan Borden <jborden@mediaone.net>
> >> There also exists a red/green problem with absolutization. It depends
on
> >>whether a parser implements XBase. A document which is parsed using an
XBase
> >>conformant parser might not be well formed (red XML), while the same
> >>document parsed with a current parser will be well formed (green XML).
> >
> >I agree that that is a huge problem with XBase.  (Ha anyone made that
> >comment formally?)
>
> I fail to understand this.
>
> Please provide an example of a document that, when parsed using an
> XML Base conformant parser, is not well-formed, whereas when parsed
> with an XML Base unaware parser is well-formed *under the same
> assumption of how relative namespace names work*.

Assuming relative URI refs are compared literally:  .... no problem...

Assuming absolutization of relative URI references when used as namespace
names:

suppose the file:  example.txt

<example xml:base="file://whatever.txt" >
    <x xmlns:a="foo">
        <y xml:base="file://another.txt">
            <z xmlns:b="foo">
                <here a:a="1" b:a="2" />
            </z>
        </y>
    </x>
</example>

Parser does not implement XBase:

    namespace prefix "a" expands to:

        file://example.txt/foo

    namespace prefix "b" expands to:

        file://example.txt/foo

    <here a:a="1" b:a="2" /> is not well-formed

Parser implements XBase

    namespace prefix "a" expands to:

        file://whatever.txt/foo

    namespace prefix "b" expands to:

        file://another.txt/foo


    <here a:a="1" b:a="2" /> is well-formed


Jonathan Borden

Received on Tuesday, 6 June 2000 12:59:47 UTC