Re: Need to add SPC

(Re-Replying ccing the mailing-list)

On 8/13/07, Lloyd Harding <lloydhardingjr@gmail.com> wrote:
> Thank you for responding to my note.
> Maybe confusion in communication.
>
> Yes schemas, dtds, relax ng etc. do validate input and outputs, and
> therefore processing - yes yes yes  absolutely needed.
> However none of these is perfect as 'humans created them'.
> SPC is there for just those situations.  Especially since we are dealing
> with language.  20% of the time we have problems with communication and we
> know what we are doing.  My experience doing natural language programming
> makes me very cautious when we expect the machine specification we created
> to handle every situation.
> SPC is used when there is no error captured or processed by any of those
> listed.
> This is a quality issue not a error found because of a specification. It is
> used to validate the specification at this point in time with this input.
> Not to validate that this input meets the specification.  The tools listed
> work to validate against a specification but nothing is there to provide for
> validation of the specification against the current input.
>
>
> On 8/13/07, mozer <xmlizer@gmail.com> wrote:
> > On 8/12/07, Lloyd Harding <lloydhardingjr@gmail.com> wrote:
> > > SPC
> > > XProc did a good job of providing the 'attachment' and the 'association'
> > > principles for the creation of assembly lines mentioned in 1993.
> > >
> > > Xproc error capabilities identify if the step does not conform to the
> XProc
> > > specification and provides a mechanism for the step internals to report
> > > using its own built-in error capabilities. Basic programming viewpoint.
> > >
> > > However, XProc does not include the use of Statistical Process Control
> (SPC)
> > > to maintain and improve quality of product produced by the step
> internals.
> > > This capability is essential for ....lines to maintain quality and
> should be
> > > part of the line not dependent on the developers of steps to include.
> >
> >
> > Interesting paralel with industrie !
>
>
> Yes check out the SGML conference papers - 1993
>
> > But you should also be aware that Schemas *are* the way to make
> > Continued Process Control (and not only statistical)
>
> I am not talking about continued process control  but statistical process
> control a well  known tool on assemblylines. There is a difference and none
> of those you mentioned, that I can see, provide for embedded SPC in Xproc
> lines - the periodic routing of output to a human for verification of
> quality.
>
> So show me where you can in a pipeline turn on a process in the step to
> route every 10th output irrespective of errors to a quality control dept.
> Please provide example. Please resolve my concern. Thank you

Ok let's try to propose a solution

Let's say that we have a process that generates sequence of documents
And you want for 1/n (n beeing 10 or 20 if you want)

The prefect tools for that is p:split-sequence

You just use

<p:split-sequence name="SPC" test="position() mod 10 = 0">
  <p:input port="source">
    <p:pipe port="the-output-to-check" step="the-step-to-check"/>
  </p:input>
</p:split-sequence>
<p:group name="do-the-check">
 <my:first-step-to-check>
   <p:input port="source">
     <p:pipe port="matched" step="SPC">
   </p:input>
 </my:first-step-to-check>
 :
 :
 :
</p:group>

So you can put any action you want in the p:group or output it if you want


Hope this helps

Xmlizer

Received on Monday, 13 August 2007 18:32:19 UTC