[
Lists Home |
Date Index |
Thread Index
]
On Mon, 2006-07-24 at 09:26 -0500, Peter Hunsberger wrote:
> On 7/24/06, David Lyon <david.lyon@preisshare.net> wrote:
> > On Mon, 2006-07-24 at 21:15 +1000, Rick Jelliffe wrote:
> > > There is nothing stopping you doing this in real XML, just moving the
> > > type tag inside the attribute value. For example (off the top of my
> > > head, details may be wrong):
> > >
> > > <Item Information PLU="A256" Name="Kitchen Veneer" Rate="$420"/>
> > >
> > > Then you can validate with Schematron, for example
> > > <sch:pattern name="typedAttribute" abstract="true">
> > > ...
> > > <sch:rule context="starts-with($node, '$')">
> > > <sch:assert test="number(string-after($node, $))"
> > > >A currency attribute should have a number</sch:assert>
> > > </sch:rule>
> > > ...
> > > </sch:pattern>
> > >
> > > <sch:pattern is-a="typedAttribute">
> > > <sch:param name="node" value="Item/@*"/>
> > > <sch:pattern>
> > >
> > > Cheers
> > > Rick Jelliffe
> >
> > Ok, say I did that.. what do I now get?
>
> Among other things:
>
> 1) the ability to use standardized parsers without having to write and
> maintain your own;
>
> 2) the ability to use an ISO standard for data validation without
> having to write and maintain your own;
>
> 3) compatibility with external systems with no extra transformation
> requirements (reduced processing overhead).
>
No, I just end up with a whole lot of bloatware...
|
