Proposed Recommendation Call for Review: XProc - An XML Pipeline Language

Norman Walsh, Alex Milowski, Henry S. Thompson (eds), W3C PR

The W3C XML Processing Model Working Group has published a ProposedRecommendation for "XProc - An XML Pipeline Language", together with an"Implementation Report for XProc: An XML Pipeline Language." Given thatthe changes to this draft do not affect the validity of that earlierimplementation feedback, except in specific areas also now covered bymore recent implementation feedback, the Working Group is now publishingthis version as a Proposed Recommendation. The review period ends on15-April-2010; members of the public are invited to send comments onthis Proposed Recommendation to the 'public-xml-processing-model-comments'mailing list.

An XML Pipeline specifies a sequence of operations to be performed on acollection of XML input documents. Pipelines take zero or more XMLdocuments as their input and produce zero or more XML documents as theiroutput.

A pipeline consists of steps. Like pipelines, steps take zero or more XMLdocuments as their inputs and produce zero or more XML documents as theiroutputs. The inputs of a step come from the web, from the pipelinedocument, from the inputs to the pipeline itself, or from the outputs ofother steps in the pipeline. The outputs from a step are consumed byother steps, are outputs of the pipeline as a whole, or are discarded.There are three kinds of steps: atomic steps, compound steps, andmulti-container steps. Atomic steps carry out single operations and haveno substructure as far as the pipeline is concerned. Compound steps andmulti-container steps control the execution of other steps, which theyinclude in the form of one or more subpipelines.

The result of evaluating a pipeline (or subpipeline) is the result ofevaluating the steps that it contains, in an order consistent with theconnections between them. A pipeline must behave as if it evaluated eachstep each time it is encountered. Unless otherwise indicated,implementations must not assume that steps are functional (that is, thattheir outputs depend only on their inputs, options, and parameters) orside-effect free. The pattern of connections between steps will notalways completely determine their order of evaluation. The evaluationorder of steps not connected to one another is implementation-dependent...A typical step has zero or more inputs, from which it receives XMLdocuments to process, zero or more outputs, to which it sends XMLdocument results, and can have options and/or parameters. An atomicstep is a step that performs a unit of XML processing, such as XIncludeor transformation, and has no internal subpipeline. ] Atomic steps carryout fundamental XML operations and can perform arbitrary amounts ofcomputation, but they are indivisible. An XSLT step, for example,performs XSLT processing; a Validate with XML Schema step validates oneinput with respect to some set of XML Schemas, etc..."

http://www.w3.org/TR/2010/PR-xproc-20100309/See also the XProc Implementation Report: http://www.w3.org/XML/XProc/2010/02/ir.html