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

Re: The Scope of Step Names

From: Alex Milowski <alex@milowski.org>
Date: Wed, 04 Oct 2006 10:07:16 -0700
Message-ID: <4523EA44.2080704@milowski.org>
To: public-xml-processing-model-wg@w3.org

Jeni Tennison wrote:
> In other words, the explicit dependencies between steps provide a 
> partial ordering; where the explicit dependencies do not specify an 
> ordering between two steps, the result must be as if the implementation 
> performed the steps in the order given in the pipeline specification.

Another bit I just realized here is that by doing this you've
imposed a total ordering over all the steps.  That prevents the
possibility of parallel step execution--which is an important
optimization.

If we want to control ordering of side-effects that can't be
ordered via input/output chains, I'd be in favor of an explicit
action by the user to introduce another order relation (e.g.
this step is after that step).  That allows us to keep
a partial order and allow for parallel optimizations.

We need to ask ourselves whether this is a 1.0 issue.

--Alex Milowski
Received on Wednesday, 4 October 2006 17:07:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:49 GMT