[
Lists Home |
Date Index |
Thread Index
]
On 7/11/05, bill palmer <palmer@execpc.com> wrote:
> >>Given the above, I'd say your heading toward reinventing XSLT (but without
> using any of those messy namespaces)...<<
>
> I am envisioning that this rearrangement of the information depends on XSLT
> or equivalemnt.
>
> In my example of <record>s, and from the stand point of XSLT, I imagine a
> template with select="e" which copies each <e/> to a <record/> plus all the
> attributes on <g><gx><gi>...
>
> So I do not see it as a coding method like XSLT, for defining how to
> transform arbitrary xml tags to new forms, but it is rather one particular
> obvious transform always from <e/> tags. The xml preparer is allowed to defer
> the technical implementation of the transform to the recipient of the file.
> In my specific example, we might expect the recipient expands to <record>s
> before doing validation etc.
Umm, I think you missed the point: you don't need any of this, you can
already do this (and far more) using XSLT. Eg, you can reference
(include) an XSLT from a document, you can add in-line XSLT to a
document as a (semi-proprietary) macro format, or add a "document" to
an XSLT stylesheet as a variable (or include)....
>
> If it is a compression idea, then it is a persistent form of compression
> residing within the otherwise identical flat xml. The "compressed" file is
> the created file in step one, rather than creating a file and then
> compressing.
>
<snip/>
--
Peter Hunsberger
|
