Re: [ubl-fpsc] <xsd:documentation> --> <xsd:appinfo> & metadata

[Dan Vint <dvint@acord.org>]

At 10:01 AM 7/23/2003 +0800, Chin Chee-Kai wrote:
>There's been some discussion within FP on recognising
>that certain presentational metadata should rightly
>be classified as first-class values as the values are
>mostly used by transformational scripts, programs, etc.
>
>Likewise, in the course of generating the schemas and
>doing some transformational work myself, I find it
>increasingly more appropriate to move all the elements
>within the current <xsd:documentation> to <xsd:appinfo>
>because those token values and even brief definition
>strings are much more applicable inputs to machines
>than mere descriptive prose meant for human beings.


Then maybe we need to improve our "documentation". I would not move all the 
documentation to appinfo only. You can duplicate some information there but 
appinfo should not be the only source.


>Moving to <xsd:appinfo> doesn't mean that human's cannot
>read the same metadata;  it's just to recognise the
>status of such token and string values, and not to
>accidentally discard them as mere harmless discardable
>documentation during processing.


It's all documentation and if anything deleted this information I would 
think the whole annotation structure would be removed, not just one set of 
elements.



>It also does not mean setting a rule saying we now use
>only <xsd:appinfo> and no longer <xsd:documentation>,
>as we're just trying to vacate current resident elements
>to <xsd:appinfo>, leaving real English sentences
>in <xsd:documentation> (in whatever appropriate format,
>such as XHTML, etc).


Some of my organizations members have gone as far to say that markup in the 
documentation section is not appropriate - because that is for the machine 
and not the human. Also tools like display the markup in <documentation> 
rather than stripping or formatting it like HTML. I don't believe Spy shows 
any of the appinfo (I haven't confirmed this) but they do display the 
documentation section.



>At the same time, FP is also beginning to structure
>some elements to be contained within the <xsd:appinfo>
>under the overall FP parent element <ccts:Presentation>.
>
>We haven't discussed yet if it should become necessary
>to start allocating unique namespace values to FP's
><Presentation> and LC's <Component> and any further
>additional metadata groupings.  But my rough thought
>now is that we need not waste too much time on that
>aspect at the moment since the element names are very
>local to FP & LC and we can easily make unique any
>remote need when local names clash within UBL.


I agree, we could probably live with a single UBL namspace rather than 
making any separation. We need to talk a little on <Presentaiton> that 
might not be the "right" impression we want to give to the content.



>I can see contextualization needs that may require
>"parking" some metadata in the <xsd:appinfo> space.
>Those could use other namespace values and prefixes
>and would be out of current scope of discussion.
>
>
>
>So, I'd like to suggest, by way of example, to perform
>the following change in all the schemas:
>
>-----------------------------------------------------
> From existing (0.81 draft 6):
>  <xsd:annotation>
>    <xsd:documentation>
>      <ccts:Component>
>        <ccts:CategoryCode>ABIE</ccts:CategoryCode>
>        ....
>      </ccts:Component>
>    </xsd:documentation>
>  </xsd:annotation>
>
>
>To:
>  <xsd:annotation>
>    <xsd:appinfo>   <!-- Changed to appinfo -->
>      <ccts:Component>
>        <!-- LC Object Metadata Elements -->
>        <ccts:CategoryCode>ABIE</ccts:CategoryCode>
>        ....
>      </ccts:Component>
>      <ccts:Presentation>
>        <!-- FPSC Presentation Metadata Elements -->
>        ....
>      </ccts:Presentation>
>    </xsd:appinfo>
>  </xsd:annotation>
>
>-----------------------------------------------------


In general that looks good, if we have only one namespace I think we need 
something other than ccts as its prefix - not sure what they current URL is 
or what if should be but ccts seems to specific.

..dan



>Perhaps LC & FP could discuss/approve/amend it at
>Montreal's meeting.  Sorry I can't be there to participate,
>so I hope to bring it up here and to summarise as clearly
>as possible for your discussion/approval.
>
>Thanks.
>
>
>
>Best Regards,
>Chin Chee-Kai
>SoftML
>Tel: +65-6820-2979
>Fax: +65-6743-7875
>Email: cheekai@SoftML.Net
>http://SoftML.Net/
>
>
>
>
>You may leave a Technical Committee at any time by visiting 
>http://www.oasis-open.org/apps/org/workgroup/ubl-fpsc/members/leave_workgroup.php