W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > July 2006

Re: Naming steps or naming outputs

From: Alex Milowski <alex@milowski.org>
Date: Wed, 05 Jul 2006 13:56:26 -0700
Message-ID: <44AC277A.8000301@milowski.org>
To: public-xml-processing-model-wg@w3.org

Rui Lopes wrote:
> Alex Milowski wrote:
>> I don't think changing <my:safeguard/> to <p:step type="my:safeguard"/>
>> is going to keep your CPU from melting.  The pipeline processor has to
>> know what my:safeguard means in either case.  As such, the burden is the
>> same.
> 
> With a generic syntax every xproc implementation will be able to 
> identify what a step is. If some non-core component is used (e.g. 
> my:safeguard) and my implementation doesn't support it, it can stop the 
> execution (or even not execute the pipeline at all). By having a direct 
> syntax, a pipeline implementation may not be able to distinguish between 
> a non-core component and, for instance, some annotations in rdf/dc.

Good point, but not necessarily true.

Much like in XSLT, in smallx I added an extension-element-prefixes
attribute on the root pipe element.  This allows the compiler to know
whether it has a step it doesn't understand.  That is, if the namespace
associated with that prefix is the same as the namespace as the
unknown step, it can "halt and catch fire" on that step.

We could do the same here and so, again, I think this is just a choice
we have to make based on preference of syntax.

--Alex Milowski
Received on Wednesday, 5 July 2006 20:56:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:48 GMT