- From: Alex Milowski <alex@milowski.org>
- Date: Wed, 18 Apr 2007 10:08:17 -0700
- To: public-xml-processing-model-wg@w3.org
- Message-ID: <28d56ece0704181008r22286914t74d12b4d5725e2cd@mail.gmail.com>
On 4/18/07, Norman Walsh <Norman.Walsh@sun.com> wrote: > > / Alex Milowski <alex@milowski.org> was heard to say: > |> <p:store> > |> <p:option name="method" select="'html'"/> > |> ... > |> </p:store> > | > | That doesn't work for my use case. I want to associate > | serialization parameters to output ports of pipelines. > > Of pipelines, or of steps? It seems to me that the pipeline processor > can do whatever it wants with pipeline outputs. That's explicitly out > of scope. > > | That way > | the environment in which the pipeline is executed can do > | the "right thing" much as is the case when you execute > | an XSLT transform with the 'html' serialization output method. > > I don't follow the analogy. An XSLT processor provides the output method to whatever environment runs the XSLT transformation. The environment is responsible for doing the right thing. For example, if you invoke an XSLT transformation from the JAXP API and store the result in a DOMResult instance, you wouldn't know that the requested output method was actually 'html'. You could get the output properties and understand that when you serialize that DOM you should use 'html' semantics. Similarly, if we don't provide a way for pipelines to have "output methods" similar to XSLT, then we have the problem that when someone transitions from a single transformation to a pipeline the wrong thing will happen at serialization. For example, if someone does something simple like replace a single XSLT transformation with an 'html' output method with a pipeline that sequences XInclude and XSLT, they'll get the *wrong* output. That's because we don't have the "output method" for the pipeline and serialization will assume XML. That user is going to be very unhappy. But in any event, my knee jerk reaction to having outputs carry around > extra metadata about their anticipated serialization is strongly > negative. So, the same feature in XSLT invokes a negative reaction? I'm only talking about endpoints of pipelines. -- --Alex Milowski "The excellence of grammar as a guide is proportional to the paucity of the inflexions, i.e. to the degree of analysis effected by the language considered." Bertrand Russell in a footnote of Principles of Mathematics
Received on Wednesday, 18 April 2007 19:05:30 UTC