Re: [office-metadata] More RDF/XML encoding...

["Florian Reuter" <freuter@novell.com>]

Hi Bruce,

> So you're observing the first is valid RDF; right?

Indeed. I would like

<style:style ... >
   <ext:paragraph-properties ext:property=”value” ext:intprop=”123”>
</style:style>

and

<style:style ... >
   <ext:paragraph-properties>
     <ext:property>value</ext:property>
     <ext:intprop>123</ext:intprop>
      <ext:property>
</style:style>

to be "equivalent".

> How would you propose to handle this? Would you require all literals  
> include datatypes?

Just the one which come after ODF1.2 --- I guess.

~Florian


>>> Bruce D'Arcus <bruce.darcus@OpenDocument.us> 01/24/07 6:46 PM >>>

On Jan 24, 2007, at 11:53 AM, Florian Reuter wrote:

> More RDF/XML encoding...
>
> 1. Use RDF/XML for generic parts
>
> The metadata SC already agreed to use an RDF/XML syntax for the  
> manifest.xml file.

We actually didn't settle this, and doing so would require some minor  
changes to the manifest schema. But I'd be interested in discussing it  
if it's possible.

> There are at least to other part of an OpenDocument file where the  
> same statement would be usefull: the <office:meta> stream and  
> <style:style> content. Both already use a structure compatible with  
> RDF/XML.
>
> Since the <office:meta> stream encoded meta data of the document use  
> of RDF/XML should be clear.
>
> The reason to restrict content of the <style:style> element to an  
> RDF/XML syntax is to allow ODF1.2 application to easily preserve  
> unknown child elements of the <style:style> element.
>
> Consider the following example:
> <style:style style:name="SampleStyle" style:family="paragraph">
>   <style:paragraph-properties fo:margin-top="0.212cm"  
> fo:margin-bottom="0.212cm"/>
>   <style:text-properties style:font-name="Times New Roman"  
> fo:font-size="12pt" />
>   <ext:paragraph-properties ext:property=”value” ext:intprop=”123”>
> </style:style>
>
> which would be equivalent to
> <style:style style:name="SampleStyle" style:family="paragraph">
>   <style:paragraph-properties fo:margin-top="0.212cm"  
> fo:margin-bottom="0.212cm"/>
>   <style:text-properties style:font-name="Times New Roman"  
> fo:font-size="12pt" />
>   <ext:paragraph-properties>
>     <ext:property>value</ext:property>
>     <ext:intprop>123</ext:intprop>
>   </ext:paragraph-properties>
> </style:style>

So you're observing the first is valid RDF; right?

> 2. Make use of rdf:datatype
>
> It is very important to give ODF processing entities information about  
> the data type of unknown elements. This will allow the application to  
> perform a better validation and provide better APIs.
> The following style fragment for example contains information about  
> the types which allows the application to perform a validation and  
> provide a typed API:
> <style:style style:name="SampleStyle" style:family="paragraph">
>   <style:paragraph-properties fo:margin-top="0.212cm"  
> fo:margin-bottom="0.212cm"/>
>   <style:text-properties style:font-name="Times New Roman"  
> fo:font-size="12pt" />
>   <ext:paragraph-properties>
>     <ext:property  
> rdf:datatype="http://www.w3.org/2001/XMLSchema#string”>value</ext: 
> property>
>     <ext:intprop  
> rdf:datatype="http://www.w3.org/2001/XMLSchema#int";>123</ext:intprop>
>   </ext:paragraph-properties>
> </style:style>
> We should define a list of datatypes to be understood by every ODF1.2  
> compliant application.

How would you propose to handle this? Would you require all literals  
include datatypes?

Bruce