W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > December 2010

RE: Connecting output ports of compound steps - inconsistency in the spec?

From: <vojtech.toman@emc.com>
Date: Fri, 3 Dec 2010 07:08:41 -0500
To: <public-xml-processing-model-wg@w3.org>
Message-ID: <3799D0FD120AD940B731A37E36DAF3FE32B116B84C@MX20A.corp.emc.com>
> > I may be interpreting the text in 2.2 incorrectly (as has happened
> > to me many times before), but it seems to me that the list in 2.2
> > should be updated by allowing connections to the output ports of
> > some other steps (not just contained steps).
> 
> Yes, it does seem to me that something is missing from the 2.2 list,
> but it's something such as
> 
>   * An input port of the step itself

I don't think this is the intention. I don't think that within a compound step, we ever wanted to allow connecting its output ports directly to its input ports. I think we even have a test in the test suite that expects that to be an error.

> 
> I don't see any basis in 5.11 for connecting to "some other steps" --
> what other steps would those be?

By "some other steps" I meant steps that are not contained in the compound step, but whose output ports are among the compound step's readable ports, like in the example below:

<p:identity name="ident"/>

<p:group>
  <p:output port="result">
    <p:pipe step="ident" port="result"/>
  </p:output>

  ...subpipeline... 

</p:group>

The current rules in 2.2 don't seem to allow this, but the text in 5.11 does.

Regards,
Vojtech

--
Vojtech Toman
Consultant Software Engineer
EMC | Information Intelligence Group
vojtech.toman@emc.com
http://developer.emc.com/xmltech
Received on Friday, 3 December 2010 12:10:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 3 December 2010 12:10:07 GMT