- From: Norman Walsh <ndw@nwalsh.com>
- Date: Tue, 07 Aug 2007 13:37:35 -0400
- To: public-xml-processing-model-wg@w3.org
- Message-ID: <87lkcn9rfk.fsf@nwalsh.com>
/ "Alex Milowski" <alex@milowski.org> was heard to say: | On 8/7/07, Norman Walsh <ndw@nwalsh.com> wrote: |> Are we going to define a MIME type (application/xproc+xml?) and |> a fragment identifier syntax so that it is possible to point |> to steps/ports/etc. from outside a pipeline document? | | Yes, we should define a mime type. | | I'm not sure how the fragment identifier would work with nested | steps. My thinking in the shower this morning was: #step-name/childstep-name^port-name #step-name/step-name$option-name So #step-name finds the first (in document-order) step named 'step-name'. If that's not unique, you can do: #step-name/childstep-name which finds the first step named 'childstep-name' that is a child of the first step named 'step-name'. Repeat as deep as you need. I think that our rules for the scope of step names guarantee that that simple algorithm will always uniquely find a step. I don't think we need to support pointing to p:logs or p:parameters. But we could invent something more complex if we needed to. Be seeing you, norm -- Norman Walsh <ndw@nwalsh.com> | Mankind are always happy for having http://nwalsh.com/ | been happy; so that if you make them | happy now, you make them happy twenty | years hence by the memory of | it.--Sydney Smith
Received on Tuesday, 7 August 2007 17:37:46 UTC