W3C home > Mailing lists > Public > xmlschema-dev@w3.org > August 2004

RE: Pattern constraint on string content of complexType having attribute and choice

From: Wyatt Barton <hiddenharvest@yahoo.com>
Date: Wed, 11 Aug 2004 20:24:22 -0700 (PDT)
Message-ID: <20040812032422.90652.qmail@web40601.mail.yahoo.com>
To: Michael Kay <mhk@mhk.me.uk>, 'Jeni Tennison' <jeni@jenitennison.com>
Cc: xmlschema-dev@w3.org

Hi Michael,
Thanks for the feedback. A complicating factor in our
requirements is that (unlike your <note> tag [see
your message below] which, when present, always
comes as the second of a pair of tags), our tag for
<location-specific_Comment> can appear *anywhere* in
the character data. [Basically, we are marking up
transcriptions of medieval manuscripts, where the
manuscript content is character data, and editorial
comments are treated as markup.] The content of our
<target_Type> must, I think, be "mixed".

JT suggested leaving the Pattern out of our Schema
declarations, and enforcing our RegEx via a separate
validation-step. I expect she is thinking this could
be triggered by an <xs:appinfo> tag, or somehow by
pipelining the Schema-validator with another program.
Such a program would need to enforce the RegEx on
character data within <target_Type> and ignore any
<location-specific_Comment> tags.

By our requirements, it is crucial that the character
data beween the <target_Type> tags *must* be RegEx-
valid. I'm unsure how we could implement pipelining
in an automated way, except at the data-entry side.
Our instance docs are intended to be "first-class
objects" on the Web, and I see no way currently to
invoke a proprietary second-step validator during
the course of browser retrieval of an XML doc. Any
suggestions -- from anyone -- would be most welcome.
Cheers,
- Wyatt

--- Michael Kay <mhk@mhk.me.uk> wrote:
> If you do a direct translation of GEDCOM into XML
> you get a similar
> requirement. GEDCOM uses elements that contain a
> value followed by
> qualifiers, for example
> 
> <occupation>baker
>   <date>10 MAR 1566
>    <note>unsure which calendar
>      <author>MHK</author>
>      <date>10 AUG 2004</date>
>    </note>
>   </date>
> </occupation>
> 
> The obvious answer of puttting the value in another
> layer of markup:
> 
> <occupation><value>baker</value>
>   <date><value>10 MAR 1566</value>
>    <note><value>unsure which calendar</value>
>      <author><value>MHK</value></author>
>      <date><value>10 AUG 2004</value></date>
>    </note>
>   </date>
> </occupation>
> 
> Michael Kay
<
Received on Thursday, 12 August 2004 03:29:33 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:15:10 UTC