Re: RDDL as a delivery vehicle for XSLT extensions?

"Clark C. Evans" wrote:

> In conclusion I'd say that, despite the few keystrokes saved,
> (and with GUI XSLT editors, not even a mouse click)
> implements="namespace" is _far_ more clear and less error-prone
> then implements-prefix="xxx".  Seriously, the WG should consider
> this point.  I'm actually very experienced with XSLT, including
> namespaces, and it wasn't at all obvious to me.  IMHO, left as
> implements-prefix this is going to be an entire FAQ section -- the
> keystrokes saved are just not worth it.  If someone wants to
> save keystrokes, they should be using xsl:import...

I don't think the point is saving keystrokes.  The point is reducing
errors and easing maintenance.  Namespace URIs can be long and
unfriendly (like uuid:); having to put the same URI in multiple places
in a single document is error-prone and makes maintenance harder. (Yes,
I know you could use entities, but entities aren't part of everybody's
long term vision for XML.)

There is also a precedent for this in XSLT 1.0: the result-prefix and
stylesheet-prefix attributes on xsl:namespace-alias and the
extension-element-prefixes attribute on xsl:stylesheet are using
prefixes to refer to namespaces in exactly the same way.  In these
cases, as with implements-prefix (and as with element and attribute
QNames), the prefix has no function other than to serve as a place
holder for the namespace URI.

I can see that the "-prefix" suffix in the names of these attributes is
potentially confusing, but I think it's too late to change these, and I
think it is better for implements-prefix to be named consistently.

James


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list