Re: [office-metadata] Suggested Changes on the Metadata proposal

[Svante Schubert <Svante.Schubert@Sun.COM>]

marbux wrote:
>
>
> On 6/27/07, *Svante Schubert* <Svante.Schubert@sun.com 
> <mailto:Svante.Schubert@sun.com>> wrote:
>
>     Greetings!
>
>     here an updated list of issues to be discussed and editorial changes.
>
>     New suggested of changes to be discussed
>     =========================================
>
>     VI) Dropping non modification requirement
>     "Metadata files should not be modified unless the content of the
>     metadata file is changed.
>
>     This sentence describes application behavior. The described
>     behavior is
>     moreover not essential, nor do we have something similar for ODF
>     content, like keeping the XML structure, when the document is not
>     being
>     changed.
>
>
> Virtually every section of the specification describes application 
> behavior. That is an irrelevant factor. Moreover, the described 
> behavior is essential for round-trip intoperability purposes. And the 
> fact that we have no equivalent requirement for maintaining the XML 
> structure counsels that we create such a requirement; it is not a 
> justification for breaking interoperability in another area.
>
> This suggestion should be withdrawn unless you can provide a more 
> compelling justification for it.
>
First of all, be affirmed I was not ordered by anyone to propose the 
change list.
As Patrick is on the road and helped him out to collect the change 
proposals from the TC and post them to the SC list for discussion.

Let us try to discuss the issues on a technical level as I am still 
convinced that we can sort them out.

The proposal to drop the sentence "Metadata files should not be modified 
unless the content of the
metadata file is changed." was given by myself.

In my words the sentence says that RDF/XML files should not be modified 
at all after the ODF document is being loaded and saved, when the 
metadata has not changed.
No modification implies to me no change of syntax nor format like indent.
Although I agree that it is reasonable to keep the RDF/XML file as they 
were - when there is no need for a change - it is quite impossible for a 
RDF/XML serializer/deserializer to keep the file as it was, especially 
when it was earlier processed from a different RDF/XML de/serializer. 
The reason is that RDF/XML can have many "shapes" for the same RDF graph.

Mapped to the interoperability theme - the interoperability of RDF/XML 
is not bound to the used format or syntax, therefore the sentence is in 
my eyes some kind of overspecification.
Does this clarify the case?

regards,
Svante